In the context of achieving a goal, one end of the seriousness spectrum is complete unseriousness, where the player doesn’t do anything to further their goal. On the other end is obsessiveness, where the player does the bare minimum in other parts of their life to maximize progress towards their goal—they may lose friends, suffer health issues, or enter financial straits in their pursuit of glory and greatness.

But actual seriousness (AS) and claimed seriousness (CS) can be, and often are, two very different points on the spectrum. Economists call AS a revealed preference and CS a stated preference. I’ll call them AS and CS because it feels like a subset of preference theory. It also makes it sound like I came up with something original.

I suspect CS follows a normal distribution and AS follows a skewed-right distribution, i.e., the bulk of people have below-average AS. Put simply, most people are not as serious towards their goals as they claim. There are a few reasons for this:

• They know they’re not that serious, but claim to be so for status, fitting in, etc.

  • Ex: Lydia doesn’t like music that much, but it’s gotten her respect, friends, and a paycheck, so she continues doing the bare minimum, which is far less than what she could actually be doing if she truly cared about music.

• Their calibration is off, and they think their AS equals their CS. This mismatch can go both ways: they can bring their AS up to match their CS, or bring their CS down to match their AS. In the former, they may not know that they could be more serious. In the latter, the extra CS may be taking up resources, e.g., time spent trying to back up the CS, even when they know they don’t care. The person would be better off just saying “I don’t care as much as I said I did”.

  • Ex: Sharon says she is serious about music, putting her time and money where her mouth is by spending $1000 on a violin and practicing an hour or two a day. In reality, professional violinists easily spend five figures on their violin and up to four hours per day practicing. Sharon thought she was nearing the ceiling, when in reality she’s nowhere close.

  • Ex: Adam says he is serious about volunteering for a cause. He is on many mailing lists and regularly gets contacted to be asked to help or donate in some way. He doesn’t enjoy this and eventually unsubscribes from all contact. He volunteers on his own terms and feels much better.

Both of these are issues. Intentionally lying to others corrupts the soul. How else could that time be spent if the honest path was taken and they chose to spend effort on something that was actually enticing? The miscalibration leaves plenty of progress on the table. What if Sharon learned that there are still returns to be had between the one- and four-hour practice marks? That’s 15+ hours a week of potential gains she’s missing out on!

The rest of this post attempts to help the reader gauge where their CS and AS points lie relative to one another, as well as serve as a reference for me to see how much more I could be doing in my seriousness studies.

Questions to Help Determine Seriousness

Any “no” answers to the following questions probably means you aren’t as serious as you could be.

Are You Doing the Thing?

No, like actually doing it. Getting better at writing requires writing, cycling requires cycling. Fuckarounditis is a serious disease that afflicts countless serious people every year without them realizing it. The SAID principle is real.

Are You Incurring a Cost?

Cost being time, money, social life, opportunity, etc. Revealed preferences strike again, and if someone is serious about something, they should be willing to pay for it in proportion to how serious they are. Are you spending money towards it? Is your free time being taken up by it? Are you sacrificing social and romantic opportunities to pursue it?

Are You Learning and Applying Those Lessons?

Are you reviewing your work and trying to find ways to get better? Are you doing post-mortems, analyzing them for deficiencies, then finding ways to patch those holes quickly so the same mistake isn’t made again? Are you seeking out ways to learn, whether online, in person, through books and film and podcasts and interviews? Are you actively trying to learn? Not just sitting there letting the knowledge go over you, but actively thinking, generating new ideas, testing them, refining them? Are you seeking out failures to learn from?

Are You Seeking Feedback to Get Better?

Are you asking others to give you raw, honest, constructive feedback and really accepting it and making the necessary corrections? Are you even open to feedback? Are you increasing your surface area for feedback to reach you, and then following up when it does to get even more?

Are You Competitive?

Are you trying to become the best? Are you trying to outwork your competitors or your past self? Do you have a way to rank yourself, whether on an absolute scale or against others?

If a Stranger Watched You For a Week With No Context, Would They Think You’re Serious About This?

Or describe your day to a stranger in 15-minute increments and ask them “what do you think I’m serious about?”. If their answer is different from yours, think about why that may be. Does your time and actions actually reflect your claimed seriousness?

Are You Looking at Serious People and Trying to Mimic Them?

Look at people you consider to be serious. How is their life different from yours? Are they really built different, or are they doing things you aren’t willing to? Kobe Bryant practiced basketball for hours every day. Startup founders and their teams work 100 hours a week in pursuit of their mission.

Examples of Serious and Unserious People

Here are some fictitious examples of both seriousness and lack thereof across different fields:

• Krista doesn’t use LLMs for coding because “they’re just stochastic parrots“. She instead chooses to write everything by hand. She refuses to even test a model on principle and scowl at colleagues using it. Meanwhile, her colleagues use a hybrid approach, getting LLMs to do the bulk of the work while they check its accuracy and completeness, ultimately speeding up their deployment time immensely.

• Olga doesn’t have a way to measure cycling power because “it’s too expensive” and “RPE-based training works just fine” and “analyzing power data is too time-consuming”. While she’s busy training on feel and not making much progress, her competitors are using data to guide their training, making leaps and bounds in both fitness and recovery. (RPE 7 will always feel RPE 7, so if you’re not recovering enough you can’t tell because RPE 7 = RPE 7.)

• Francesca is trying to lose weight by eating less. But counting calories and weighing food is “too inconvenient”, so she doesn’t do it. Using a GLP-1 is also “cheating”. Her friends, on the other hand, are more involved: Eliot obsessively counts and tracks calories, taking care to measure ingredients out carefully and asking restaurants for nutrition info if they have it; Andris got prescribed a GLP-1 and follows the prescriptions. Both of them shed pounds by the week.

• Todd wants to spend less time on his phone, so he sets a non-password-protected screen time limit on the most distracting apps. This doesn’t change anything—he blazes right past the gates through self-persuasion and habit. Scott takes a different approach: he purchases a phone lock box, removes the addictive apps, makes it greyscale, and stores his screen time password in an online locker that’s guarded by 60 minutes of inconvenient, annoying work.

• Whitney is a competitive athlete who wants to up her game, and everything points to needing to spend more time on recovery. She starts to eat a bit cleaner, get more sleep, and be less stressed. She makes some progress. Andris is in the same boat, but puts more effort into his recovery: he tracks macros to ensure he’s hitting what he needs to, ensures he fuels within 60 minutes of the end of his workout, follows sleep hygiene practices, regularly foam rolls, prehabs his various joints and muscles, and actively recovers instead of taking an entire day off.

Deciding Where the Limit Is

Most of the examples above have no ceiling; there are more interventions and actions to take than the person has resources of time, energy, and money and can easily lead to burnout or other issues if everything is attempted at once.

While more is generally better, it isn’t always better. Professionals know themselves and when to temper effort for long-term gain. An athlete who went hard for two weeks then didn’t do anything for the following week would never be considered serious. Some proxies of seriousness—actually doing the thing, seeking feedback—are better than other types and should be prioritized accordingly.

Limits exist on a case-by-case basis. Determine them, test them, then raise them.

See Also

• Is that the best you can do?

• Everyone Can Tell if You’re a Serious Person

• Do Ten Times as Much

• 95%-ile isn’t that good

Every person, process, and goal has a bottleneck somewhere in the line because not everything is equal. The pursuit of eliminating bottlenecks just shifts it to a different spot in the line—when capacity in one area outgrows capacity in another, the latter becomes bottlenecked.

Certain things should not be bottlenecks depending on the area; if they are, then measures should be taken to remove it quickly. Time and money are often the bottleneck in growing a startup, as evident by the lack of sleep and (hopefully multiple!) funding rounds the founders go through on their path to glory and riches, both of which are the employees’ measures to relieve the time and money constraints, respectively. If they had access to the mythical infinite time and money glitches, then scaling would be easy. But alas, those don’t exist, so they’re stuck doing things the old-fashioned way.

(This is a blatant message to you, the reader, to take some time to think deeply about your current bottlenecks in life and what you are doing about them, if anything. If you aren’t doing anything, then consider why.)

Energy Bottlenecks Are Especially Bad

Lacking energy to do things despite knowing what to do and wanting to want to do it is a terrible feeling. The problem is identified, the path clear, but the joules needed just aren’t there. And at least for me, there’s also a lingering fear that forcing myself to do the thing while motivation and energy are low will breed resentment towards it because of a potential strong negative association that results.

Time or money feels like a better bottleneck. With time, activities can be selectively reduced or discarded altogether to buy back some minutes. It’s all in the person’s court and up to them. With money, ingenuity and hard work can make up for lack of money. Figure out how to do it cheaper, with what is accessible, ask for a loan, the list goes on. But the options are there.

Methods to Reduce Energy Bottlenecks

Energy bottlenecks can be reduced by allocating time and money towards certain areas. The same can be said of time bottlenecks (spend money and energy towards things) and money bottlenecks (spend time and energy towards things). It’s almost like a triangle of choices... where you can only pick two...

I try to include literature where possible, but have found that literature isn’t everything. If something works, it works! Forget the scientists and “peer-reviewed” journals that claim it doesn’t—this is my lived experience, bro. (Half-joking. Replication is important to an extent, but blindly rejecting an already-tried-and-found-effective treatment because some random study said it doesn’t work isn’t good. They can certainly help to question the efficacy or offer up better solutions, but studies—sometimes even the best of them—can still have gaps.) This post is meant more as a quick reference for me than a holy grail of energy interventions, so some sections may be a bit sparse.

To pre-register my hypothesis, my prediction is that it’s easy to raise the floor (i.e., energy improvements by avoiding things that cause lack of energy, e.g., poor sleep, poor lighting, etc.), but difficult to raise the ceiling, at least for extended periods of time. Plenty of quick bumps exist—stimulants, food, cold showers, bright lighting—but they are rarely long-term solutions. The focus should be placed on protecting the downside and strategically supplementing the upside in situations as needed. This will result in the total energy over time still increasing.

Lastly, I’m confident that a SOTA LLM could do a better job than me at compiling all of this and the corresponding literature, running some cost-benefit analysis, and recommending specific interventions based on personal data and background. I still wrote this because it helps me to remember and internalize what works and what doesn’t, as well as think about what I’m doing and not doing to increase my energy levels.

Sleep

Sleep Hygiene

See Sleep: Effects and Hygiene.

Hydration

Hydration is generally accepted as an important health practice, and even if the evidence is scarce, it’s probably worth doing regardless for reasons unrelated to sleep. Hydration practices are fairly simple: drink water regularly throughout the day. Following this incredibly generic rule of thumb is probably plenty to check the box.

Literature and resources:

• Effects of Fluid Intake on Sleep Duration and Quality Among Healthy Adults:

  Mild dehydration does not appear to affect sleep measures in healthy adults. However, fluid intake correlates positively with REM sleep length, sleep duration, and sleep efficiency, suggesting that the quantity of water consumed may influence sleep quality.

• Mild Dehydration by 24-h Fluid Restriction Led to Difficulty Falling Sleeping and Increased Sleep Duration:

  Our results indicate mild dehydration increased sleep duration because of increased fatigue and decreased perceived ease of falling asleep. This suggests that maintaining euhydration may assist with perceived ease of falling asleep and feelings of fatigue.

Naps

Naps, when done correctly, can boost alertness and such, but do not replace a full night’s sleep.

Avoid naps within 8 hours of bedtime.

Naps should be less than 30 min or around 90 min to avoid entering and waking up within REM.

Find Your Chronotype and Exploit It

Chronotypes are a thing. Determining which applies and leveraging it will ensure waking hours are productive and alert. Substantial time can be lost by trying to be productive when the body just isn’t into it. Motivation also wanes during these times and can negatively reinforce whatever behavior or task was being worked, making future sessions less productive.

Drugs

Newsflash: stimulants are a thing and they work! And are very easy to get! And very cheap!

Caffeine

Gwern’s Caffeine page.

Michael Dickens’ posts:

• Avoiding Caffeine Tolerance: Self-explanatory title. Michael finds that his research suggest his “approach [of taking caffeine MWF] was reasonable and there’s not compelling evidence for taking caffeine more or less often” and “taking caffeine more frequently, there’s a good chance that I will develop a tolerance”.

• Caffeine Cycling Self-Experiment: Michael finds that he developed no tolerance from MWF caffeine usage and it maintained, if not improved, its effectiveness.

• Continuing My Caffeine Self-Experiment: Michael finds that he developed no tolerance from MWThF caffeine usage and vows to continue with four days a week of usage.

On a not-directly-related note, if you have trouble waking up, there are rich-man and poor-man solutions:

• Rich: Take a delayed-release caffeine pill (example here) before bed to have it hit around your wake-up time. I’d be slightly worried about “microdosing” caffeine throughout the night and that leading to worse sleep quality.

• Poor: Wake up at t minus 30 minutes, take a caffeine pill, then go back to bed until actual wake up time. Caffeine effects start around 15 minutes after taking it, so 30 minutes is plenty and will help with waking up alert.

Caffine is insanely cheap: at the time of writing (August 2026), 250 100-mg pills costs a whopping $12, or $0.05 per pill. If you’re drinking coffee for the caffeine, save yourself $5 a coffee and buy a bunch of caffeine pills instead.

I recommend using any caffeine half-life calculator to correctly time doses to avoid negatively affecting sleep.

Modafinil

Gwern’s Modafinil page is the best page on the internet regarding modafinil. He describes its advantages over other common stimulants as:

more powerful and less addictive & tolerating than caffeine or khat; much longer-lasting than nicotine; less likely to alter mood or produce ‘tweaking’ behavior than Adderall or Vyvanse; and much more legal & with almost no side-effects compared to methamphetamine or cocaine. On any specific aspect, there may be a stimulant superior to modafinil, but few stimulants come close to modafinil’s overall package of being a long-lasting, safe, effective, non-mood-altering, quasi-legal stimulant, and that is why it has become so popular.

ADHD Medication

From my friends’ experiences—both diagnosed with ADHD and not—ADHD meds are very effective at increasing their productivity through focus, which I’d argue is a different form of energy.

See also: Lorien Psychiatry notes

Exercise

Fast(er) Recovery

Quickly recovering from exercise, especially very hard days (100s of TSS, for those who know), is important because a) being absolutely smoked post ride/run/lift/climb is just not fun and prevents other things from happening, and b) it may hinder or altogether prevent training sessions in the next few days, resulting in a lower total TSS dose than just going easier in order to exercise the following few days.

Recovering quickly is primarily a function of carbs to replenish glycogen stores, protein to repair muscles, and sleep. None of them can be ignored. Inadequate amounts of any will prolong recovery and decrease overall efficiency.

Training for the New Alpinism claims the 30-to-60-minute post-exercise window is critical for recovery:

• Carbohydrate: A 150-pound climber needs 270–400 calories (70-100 g) of carbohydrate during this post-exercise window. For best recovery repeat this every two hours to re-stimulate glycogen resynthesis for up to six hours.

• Protein: Consuming protein after a workout is useful to help the body repair damaged muscle, but studies are inconclusive about whether or not the timing is as important as it is for carbohydrate consumption.

• Fluid. Drink water and replace electrolytes. Adjust your intake to match the temperature and the amount you’ve sweated.

Blood Flow

Improving circulation/blood flow will remove waste from muscles to facilitate recovery (see above bullet point). It also delivers fresh oxygen to the brain, resulting in a boost in energy.

Activities to promote circulation include massage and easy activities (walking, cycling, swimming).

Diet

Calories

Calories are literally energy. Eat them. But be aware they may not all be equal—see below for information on the glycemic index.

Alcohol

Alcohol is a depressant. It increases drowsiness, negatively impacts sleep, and dehydrates the body. It is extremely counter-productive to increasing or even maintaining energy levels.

Digestion

Poor digestion can cause the body to focus on digestion, sapping the body of important resources, leading to fatigue.

Glyemic Index Notes

High glycemic index (GI) foods will cause blood sugar (glucose) to spike, likely leading to an energy crash later. Eating low GI foods allows a slower, steadier release of glucose via carbs into the bloodstream, preventing a crash and providing lasting energy.

High GI foods can sometimes be substituted for low GI foods:

• White rice → brown rice, basmati rice

• White bread (are you sensing a pattern here??) → wheat or rye bread

• Cornflakes/sugary cereal → steel-cut oats, bran cereal, or muesli

• Mashed potatoes → lentils or chickpeas, mashed

From Claude Sonnet 4.6:

On ripeness: unripe fruit has more starch, ripe fruit has more sugar (mostly as free glucose/fructose), so GI climbs as fruit ripens.

On cooking method: the longer and wetter the cooking, the more the starch gelatinizes and the higher the GI. Al dente pasta has a lower GI than soft-cooked pasta.

On storage/cooling: cooking and then cooling starchy foods (rice, potatoes, pasta) converts some of the starch into resistant starch, which digests more slowly and doesn’t spike blood sugar the same way. [This is why I prefer frozen rice.]

Pairing high GI and other foods together can reduce the blood sugar spike.

Miscellaneous

Stress

Stress releases cortisol. High cortisol levels lead to fatigue. Stop being stressed!

Cold Showers

Cold showers work by triggering a norepinephrine and cortisol response, both of which provide a short-lived boost of energy. I used to shower cold every day and didn’t notice a long-term benefit to them, but also wasn’t tracking any metrics diligently. (If anything, it’s worth dialing down shower temperature because too hot of water damages skin.)

Literature and resources:

• The Effect of Cold Showering on Health and Work: A Randomized Controlled Trial:

  The most commonly reported beneficial effect was an increase in perceived energy levels (including many reported comparisons to the effect of caffeine).

Sunlight and Bright (Indoor) Lighting

Bright lighting—either from the sun or high-power lamps—suppresses melatonin secretion and helps to regulate the circadian rhythm. The former effect staves off drowsiness, the latter improves sleep regularity and quality. This works in my experience and is well worth investing in.

Blog posts (with some literature embedded inside of them):

• Your room can be as bright as the outdoors

• LessWrong lighting posts

• Brighter Lamp: 60,000 lumens in a single lamp. Holy shit!

Literature (with no blog posts included):

• Effects of light on human circadian rhythms, sleep and mood

Physical and Mental Decluttering

Keeping a bunch of information in your working memory requires mental effort.

Literature and resources:

• The Magical Mystery Four: How is Working Memory Capacity Limited, and Why?

• How to deal with working memory exhaustion and mental fatigue?

• Getting Things Done

Sleepy Triggers

“Sleepy triggers” should be avoided. These may manifest themselves as:

• Lying horizontally in a way that signals the body that it’s time to sleep

• Spending time in bed, even if not lying horizontally. The bed is for sleep and sex, nothing else!

• Dark rooms. See Lighting.

• Physical touch (back or head scratches, massage)

• Anything associated with tiredness or sleep

Literature and resources:

• None, but seems very sensible to me

Mindsets

I will kindly redirect readers to Energetic Aliens to learn how passion and energy relate to one another.

Being ambitious (startup founder), excited (nerd who just got an update on his favorite package), or scared shitless (junior investment banker) about something helps to unlock some hidden stores of energy. Find that thing and push for it.

Masturbation

The NoFap cult will have you believe that abstaining from masturbation is the key to unlimited energy (example 1, example 2). Pseudoscience aside, the anecdata seems to suggest that chronic masturbation leads to general fatigue, whereas infrequent masturbation does not.

Literature and resources:

• I couldn’t find any actual studies that weren’t NoFap coded

Commitment Devices

I’m doubtful that commitment devices such as Focusmate and Beeminder will increase physical energy in a true, biological sense, but they serve as a concrete goal, and that drives something.

Anecdotally, I’ve found Focusmate to work when I’m low on energy. Beeminder is much more spread out, so I can’t use it tactically in the same way I can Focusmate, but when coming up on a derail I suddenly feel a boost in energy as a result of my precious money being threatened.

CO2 Levels

Indoor CO2 levels can affect cognitive performance and tiredness levels.

Solutions are centered around improving ventilation by opening windows or doors, adding fans, or improving recirculation capabilities of the HVAC system. While plants do absorb some CO2, an indoor jungle would be required to appreciably decrease levels. Improving ventilation is the cheapest, easiest, and most effective solution.

Literature and resources:

• Is CO2 an Indoor Pollutant? Direct Effects of Low-to-Moderate CO2 Concentrations on Human Decision-Making Performance:

  Relative to 600 ppm, at 1,000 ppm CO2, moderate and statistically significant decrements occurred in six of nine scales of decision-making performance. At 2,500 ppm, large and statistically significant reductions occurred in seven scales of decision-making performance (raw score ratios, 0.06–0.56), but performance on the focused activity scale increased.

• Associations of Cognitive Function Scores with Carbon Dioxide, Ventilation, and Volatile Organic Compound Exposures in Office Workers: A Controlled Exposure Study of Green and Conventional Office Environments:

  On average, cognitive scores were 61% higher on the Green building day and 101% higher on the two Green+ building days than on the Conventional building day (p < 0.0001). VOCs and CO2 were independently associated with cognitive scores. [”Conventional” had high VOCs, “green” had lows VOCs.]

Prioritization and Outsourcing

Sort by Importance and Required Energy

Eisenhower’s important-urgent matrix (or any sorted to-do list, for that matter) could take on a third or replacement of dimension of energy requirement.

Replacing urgent with energy and prioritizing the four quadrants results in the following rank:

1. High-energy, high-importance: Ensures there’s enough in the tank to complete the activity

2. Low-energy, high-importance: Ensures the important things get out the door first

3. High-energy, low-importance and low-energy, low-importance: eliminate or outsource if possible. Shouldn’t all activities be somewhat important?

Outsource Where It Makes Sense

Delegation ensures the correct person is doing the job. Specialization exists for a reason and should be abided by.

Other Interventions I Didn’t Feel Like Researching Because They Felt Like Low Expected Value

• Screen time: Is screen time inversely proportional to energy? Does all the context switching between work and personal and social media take a toll? Does it lead to eye fatigue, which in turn leads to physical fatigue? Becoming bored on the phone isn’t uncommon given how easy it is to doomscroll to the end of the new, exciting stuff, so wouldn’t that lead to lack of energy? See Smartphone dependence and its influence on physical and mental health.

• Associations: Does your physical space actually help in locking in? I’m of the opinion that my apartment office should be done for office-like work and the rest of the place is for not-office-like work. Thus when I enter the office physically I also enter it mentally. Same for clothes: does wearing a certain style of clothing tell your body “hey it’s go time”, whether it’s a suit or joggers?

• Temperature: Both temperature extremes appear to cause fatigue. Cold causes cellular activity to slow in an effort to save energy, while hot leads to an effort to cool off. I assume every person has an optimal temperature. How do we find that? Has it been found?

Ranking and Integrating Methods Into Everyday Life

The primary challenge with using the above methods is, well, actually using them! They take some sort of effort to implement. Thankfully a lot of them can be done with little to no cost and concurrently with one another.

Here’s my vibe-ordered list from most to least effective:

1. Sleep

2. Passion

3. Diet

4. Exercise

5. Drugs (stimulants)

6. Bright lighting

The remaining ones that were listed in their own section but not included in the list all feel of similar value.

See Also

• Energetic Aliens

• Stamina Succeeds:

  Of course stamina isn’t the only thing you need. It also helps to have motive, drive, intelligence, beauty, connections, charm, and many other things. But without stamina, you won’t be able to use those other things as many hours a day, which in close contests can make all the difference.

• Beware The Man of One Study

• Notes on Industriousness

• How To Be Successful:

  You have to figure out how to work hard without burning out. People find their own strategies for this, but one that almost always works is to find work you like doing with people you enjoy spending a lot of time with.

  I think people who pretend you can be super successful professionally without working most of the time (for some period of your life) are doing a disservice. In fact, work stamina seems to be one of the biggest predictors of long-term success.

A common feature of bad conversations is when person1, call them Jesse, is obviously not listening to what person2, call them Celine, is saying and doesn’t even attempt to engage. Sure, Jesse may be nodding his head and mhm-ing along, but he definitely isn’t listening and even if he was, he wouldn’t care; he’s crafting the next few sentences that very well still could’ve been formed had Celine just stayed quiet the entire time.

This type of conversational selfishness creates a terrible environment: imagine if everyone was a Jesse, ignoring their counterparts and speaking at them instead of with them. There would never be any actual conversations, only one-way noise and impatence while waiting for the other person to finish. This phenomena is precisely why self-promotion should be quarantined and strictly moderated in large communities. It can quickly and easily get out of hand since one-off defections pay well in a place where traffic is at minimum due to quality standards and any post can quickly be seen by many. Self-promoters rarely care about bidirectional engagement, thus they shouldn’t be given any at all.

The beauty in conversations is the engagement. The back and forth. The saying something and having it be heard—really heard, not just in the ears, but in the heart and mind, and in rare cases, soul. The pushback. The exposure to new ideas, regardless of their prettiness, ugliness, or intelligence. The discovery that another person out there agrees with you; or, in the case of many internet comments, vehemently disagrees and thinks you’re an absolute idiot for even thinking, much less writing, about the idea. The chance to put an idea out into the arena and see what challengers come and what the result is. The rejection and escape of single-person echo chambers.

While these features apply to both analog and digital conversations, these norms aren’t enforced on the digital side because there’s less awkwardness and more ease in violating them—being a conversational douchebag in person results in not having friends, whereas there aren’t any hard consequences for not replying to an internet comment: forgetting to reply, being overwhelmed with the sheer number of comments, not having something interesting or productive to respond with are all arguably valid reasons to shirk responding. And I’ll admit there’s something nice about not being obligated to reply and just letting comments come in versus taking the time and effort to send a comment out. It’s a lot of work to type out a thoughtful response! (This is one reason why I think using AI for writing in any interpersonal form is so frowned upon—it comes across as lazy and apathetic, as though you just want to check a to-do off a list and be done with it.)

New Rule: Reply to Everyone

Or at least to everyone who’s engaging in good faith. Most people wouldn’t ignore someone talking to them in person, so why do it on the internet? The person took time to comment, and while it may have just been for them to develop some thoughts or gain sweet karma, it almost certainly can’t hurt to see what comes of responding. Become known as a prolific commenter; gain a friend; learn something; get downvoted to hell and be scared to ever comment again; change your mind; strengthen an idea. The possibilities are both endless and enticing.

Exceptions

Exceptions exist. Given the choice of contributing to the discourse and not engaging with others versus not contributing because engagement wouldn’t happen for whatever reason, choose the former. Putting ideas and information out there is generally a good thing. Engagement can still happen elsewhere: commenters can talk amongst themselves, the piece can get reposted elsewhere, and so on. But zero engagement should not become the rule.

Enforcement

Enforcement is (or at least could be) straightforward: top-level post authors must maintain some ratio of responses written to responses received. “Great post!” and other vapid words don’t count. The good responses can be determined by number of upvotes, an LLM that gets asked “is this a high-quality reply?” or something of the sort, or trusted moderators deeming it such.

In an ideal world, the free market of the internet would reign supreme and contributors should be trusted to keep the place nice because it’s in their best interest. Simply setting a rule like “treat others’ comments and posts like you want yours to be treated” gives the soft nudges of “you want people replying to your stuff, so reply to theirs” and “you want people putting effort into their comments, so put effort into yours”. But market failures exist. People get busy and replies get missed.

So while a lot of things in society do come back to the Golden Rule, sometimes it needs a bit more enforcement.

See Also

• Well-Kept Gardens Die By Pacifism

1. There might be more abductions of foreign leaders? More wars? More stuff to bet on? More stuff, generally? More volatility; more unexpected events. The least likely outcome is, now, the most profitable: If you bet on an event at a 1% probability, and then cause it to happen, you will make 99 times your money. My overarching theory of current US policy is that everyone involved in the Trump administration basically loves creating volatility. If you’re in the business of creating forms of volatility that no one has ever imagined before, the simplest (not only!) way to monetize that is on prediction markets.

2. Conversely, foreign leaders whom Donald Trump dislikes should probably be checking their removal probability on Polymarket every few minutes. If it suddenly jumps up, you’ll want to get to the bunker quick. It might just be uninformed speculation, but at this point it’s probably someone on the helicopter getting in one last trade before rappelling down into your compound. Prediction markets are now a way to probabilistically leak military plans, and the targets of those plans are the obvious users of the leaks.

3. If you’re on the helicopter getting in one last trade before rappelling down into a foreign leader’s compound, you might be distracted? You might be less good at your job? Making war and government policy idiosyncratically profitable might reduce people’s intrinsic motivation to do a good job and leave them distracted by, like, constructing multi-leg same-raid parlays.

• Hacking Smartphone ESP Apps: “Illustration of how to think about security and reward-hacking by walking through ways to fake psychic powers even on someone else’s smartphone and ESP application. Supply-chain attacks, sleight of hand, bugs...” I find red-teaming things like this very useful across the board, especially in the age of AI where efforts that were once difficult and time-consuming are now much more accessible. For example, I just saw some guy on Twitter who lost his phone and had Claude suggest and write a program that pinged the Bluetooth for its strength to find it. And it found it. Where else is this applicable? Models are able to figure out exact locations from a single picture a lá Rainbolt; stylometry capabilities are extremely strong and accurate; etc etc etc.

• The Criterion Closet: An app that mimics the Criterion Closet. You feel like you’re in there looking around. Could probably rig up something identical with a personal film or book collection.

• Wikipedia File Explorer: Explore select Wikipedia articles through the feel of a Windows XP (?) interface.

• Emoji Book Synopses: Taylor and Claude summarize books using emojis. Would be fun to make a quiz out of this: given just the emojis, can you guess the book? I did this blind and got Flowers for Algernon and Frankenstein—most of the others I haven’t read. I think it’d be cool to have your favorite books and films printed on a shirt as a conversation piece.

• Mathematics Without Mathematicians: Life for math people after AI solves math, or “a list of ways people will cope about AI taking over mathematics, and how each cope is likely to be refuted by reality.”

• Linda Linsefors on offering advice through anecdotes

• Claude Fable is relentlessly proactive: This proactiveness, combined with all of human knowledge at their fingertips and in their DNA, is what appears to make the models so powerful. They will not stop. They do not know exhaustion or frustration or despair. They will continue until they hit their token limits or their user tells them to stop. This is so beautiful, so dangerous, so exciting, and so scary all at the same time.

• Sam Altman May Control Our Future—Can He Be Trusted?: Ronan Farrow’s look into Sam Altman as a (the?) leader of the AI revolution.

• Review of The Native Tribes of Central Australia (Baldwin Spencer/Francis James Gillen, 1899)

• Elevators: John presents customizable animations for different types of elevator algorithms and their corresponding performance. I think the next step would be adding some learning ability, e.g., 8:00-9:00am a majority of elevators should return to the ground floor and 12:00pm it should be split 50/50.

• Ronald Dale Harris: “a computer programmer who worked for the Nevada Gaming Control Board in the early 1990s and was responsible for finding flaws and gaffes in software that runs computerized casino games. Harris took advantage of his expertise, reputation and access to source code to illegally modify certain slot machines to pay out large sums of money when a specific sequence and number of coins were inserted. From 1993 to 1995, Harris and an accomplice stole thousands of dollars from Las Vegas casinos, accomplishing one of the most successful and undetected scams in casino history.”

• GCC steering committee announces AI policy: The policy, in part, states that the project will decline any “legally significant contributions which include LLM-generated content or are derived from LLM-generated content”. Seems short-sighted and shoot-yourself-in-the-leg-like. LLMs are here to stay and contribute. If the code quality is indistinguishable from human-written code, what’s the difference? Shouldn’t the priority be making better software?

• A Proof of the Cycle Double Cover Conjecture: OpenAI solves yet another high-ish profile math problem.

• Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic’s Claude goes rogue

• Investigating three real-world incidents in our cybersecurity evaluations

• AI Timeline - The Road to AGI: “This timeline attempts to tell the story of the last decade in artificial intelligence, from cultural trends to technical advancements. Each event is a clickable link to source material.”

• The OpenAI Graveyard: All The Deals And Products That Haven’t Happened: This comes across as an insult, and while it seems like they were too product-focused (causing them to arguably lose the lead to Anthropic), there’s probably an optimum between an SSI approach of “one product: superintelligence” and the old-OpenAI approach of “a bunch of products that cause us to lose focus on AGI/ASI”.

• Don’t be a meat proxy: Or just ask Claude to put it into your own words so others won’t know you’re using Claude! /s. There’s so much context that Claude can be missing (although they are quickly catching up), making human understanding and validation important. If you don’t understand Claude’s output, then maybe you are in too deep.

• Millenarianism: “belief held by a religious, social, or political group or movement in a coming fundamental transformation of society, after which “all things will be changed”.” Terrorists groups, such as Boko Haram, often hold this stance.

• Justice Department Requires RealPage to End the Sharing of Competitively Sensitive Information and Alignment of Pricing Among Competitors

• 52-hertz whale: “colloquially referred to as 52 Blue, is an individual whale of unidentified species that calls at the unusual frequency of 52 hertz in the north Pacific Ocean between Aleutian and Kodiak Islands to the California coast. The whale itself has never been sighted: it has only been heard via hydrophones”

• AI Content Is Everywhere on Social Media, Especially LinkedIn: Pangram Labs launched a Chrome extension to help users detect AI slop and collect data on what site-level slop data was like. The results aren’t super surprising? LinkedIn’s slop notoriety has made it to my social circle where not too many people LinkedIners. Also, LinkedIn has now introduced as “Seems Like AI Slop” button.

• Has AI Already Killed How-To Nonfiction? Sales Trends, My Personal Data, and What It Might Mean for the Future: A rare of example of Betteridge’s law of headlines being false! How-to nonfiction dying is arguably a good thing: people need custom solutions to their problems, not one-size-fits-all approach. The current set of frontier models are excellent on this and help get to the root of the problem quickly and effectively.

• Wife acceptance factor: “assessment of design elements that either increase or diminish the likelihood a wife will approve the purchase of expensive consumer electronics products such as high fidelity loudspeakers and home theater systems.”

• The Case for Physical Media Ownership: Arguments and examples of why you should own physical media. The tech companies have proven themselves unreliable when it comes to guaranteeing consumers consistent access to their own media, even if they’ve paid. Profits rise when going to digital-only, ceteris paribus; users can’t share discs as easily, so the sharee is forced to spend money themselves. (Of course, it’s not a 1:1 conversion rate since some people just won’t buy it, but it appears to be profitable since companies are slowly moving that way. I haven’t searched for any literature on the topic.)

• 98% isn’t very much: Context matters. I once saw an apartment advertise 98% internet uptime—in other words, you wouldn’t have internet for 30 minutes a day. That ain’t right! Context matters.

• FelonyBench: How many felonies each AI company has committed.

• OverpAId — Fire Your CEO. Hire The Future.: “OverpAId is an Artificial Intelligence built from the ground up to do your CEO’s entire job — strategy, “vision,” motivational all-hands emails — better, faster, and without ever once asking the board for a bigger jet. Runs on a single desk-sized AI computer. Real hardware, real price, zero mystique.”

• One ant for $220: The new frontier of wildlife trafficking: “A single fertilised queen [giant African harvest ant] is able to create a whole colony and can live for decades – and can be easily posted as scanners do not tend to detect organic material.”

• CERN levels up with new superconducting karts: Wait a second, that guy kind looks like... Mario?

• The Goon Squad, by Daniel Kolitz: “In the case of the gooners, one can hope—and in more cheerful moments, I do think it’s possible—that sustained overexposure to porn will dampen the medium’s effectiveness as a numbing agent. That at a certain point, the gooner will open his eyes, find himself in a room filled with lube but void of love, and decide that the boredom of staying in that room outweighs the fear of whatever lies beyond it.”

• Laws of Software Engineering: “A collection of principles and patterns that shape software systems, teams, and decisions.”

• The Silicon Valley Founder Meat Grinder: “A few make it and get celebrated, most get squished and thrown away. After all, it turns out that steady is indeed, in 99.9% of the cases, fast.”

• How to Earn a Billion Dollars: Understand what is missing in the world, build it, then scale it. Pretty simple!

• Intel Starts Shipping High-NA EUV Silicon: Pretty quiet. You’d think they’d be bragging about getting the first shipment of the world’s most advanced machine to produce wafers!

• Staring at walls to improve focus and productivity: “Don’t use any screens/entertainment when trying to focus on work. When you start to feel mentally drained, sit and stare at a wall for x minutes to recover focus.”

• How I Use Claude: Avital shares a bunch of Claude (or really LLMs in general) workflows, including writing letterss, editing to a certain style, length, or perspective; summarizations; language tutoring; counting calories from generic food desrcriptions; medical diagnoses. I’ve found similar benefits from LLMs in these areas, as well as plenty of others. The more you use the models and understand just how far their tentacles can go, the more you realize is possible and the more ideas come to you. If you aren’t hitting the rate limit, you aren’t using the models enough.

• On Being Bad at Counting: “Driverless cars will make us safer. They will lead to fewer premature funerals. They will allow for less wasted human time commuting (which claims over a year of people’s lives!). They will necessitate fewer parking lots.” Avital shares some numbers on driving-related deaths and why autonomous driving will prevent this. Go Waymo!

• OpenAI and Hugging Face partner to address security incident during model evaluation

• The Whistleblower Who Uncovered the NSA’s ‘Big Brother Machine’

• Most Wanted and Least Wanted Paintings: Filtered by country and size of the painting. I’m not surprised that people like landscapes that much, but was expecting abstract to sweep the field of the least wanted.

• Anthropic and Dario Amodei - Internal Tech Emails

• Resetting XBOX: “History is full of companies that mistake longevity for inevitability. We will not be one of them.” XBOX announces major layoffs for a variety of reasons. Doesn’t seem unreasonable given the numbers stated in the post.

• I prefer “Yankee” over “Usonian” over “American”: Taylor talks about why the term Yankee is preferable to American: “We should make Yankee synonymous with the best of the United States -- dynamism, ingenuity, hospitality, self-sufficiency.”

• How I Built an AI Receptionist for a Luxury Mechanic Shop - Part 1: Her sibling was unable to answer the phone, which cost him potential business, so she built an AI receptionist for scheduling, answering questions, etc.

• God sleeps in the minerals: Arthur Young once said “God sleeps in the minerals, awakens in plants, walks in animals, and thinks in man.” Chamblissian proves this by taking pictures of some beautiful minerals in the Natural History Museum of Los Angeles County’s Unearthed: Raw Beauty exhibition.

• Empty Screenings: “About 10% of AMC movie showings sell zero tickets. This site finds them. Go enjoy your private theater.”

• I made my phone slow on purpose: If you really wanted to use your phone, you’d sit through the slowness.

• Apologia for the Person Who Carved His Initials into the Oldest Living Longleaf Pine in North America

• Jurassic Park computers in excruciating detail: Literally every computer in Jurassic Park. Includes information, trivia, and commentary.

• Black Book (gambling): “a list of people who are unwelcome in casinos. The name is due to the people listed being blacklisted. ... In the case of gaming control boards, people listed are generally suspected of having, or known to have, ties to organized crime. Casinos are obliged by regulations to exclude all such people from entry and can be subject to sanctions for failure to do so.”

• The worst job interview I ever had

• 4chan battlestation images: I remember seeing this in high school and all I could think was “how do they get, or even pay for, internet access?!?!”

Other Link Aggregators

• Daily Links

• Splitting Infinity

• Gwern

• Astral Codex Ten

• Nintil

Semiconductor Fabs I: The Equipment

Semiconductor Fabs II: The Operation

Semiconductor Fabs III: The Data and Automation

Preface

I tried to include as many links as possible to allow the reader to go down rabbit holes as they see fit.

I don’t work in an advanced or new fab, but have some glimpses into them.

I used Claude Sonnet 5 for two (2) text blocks and interactive animations. I disclose Claude’s writing for the text blocks. The animations are disclosed by this paragraph.

Philosophy

Automate Where Possible

Semiconductor Fabs III: The Data and Automation discusses why automation is so important in a fab from a material perspective: it’s faster, safer (for the product), easier, and cheaper. But it also applies physical safety as well. Humans in the loop can be catastrophic—they misinterpret something, make the wrong decision, perform the wrong action, or be really stupid. Having physical connections, electrical signals, and default safe conditions do the work minimizes the chance of taking the wrong action and enforces stupidity, just in the way that computers are stupid and do exactly what they’re told to.

All Intention, No Accidents

Sometimes safety mechanisms must be overridden for legitimate reasons. When this happens, it shouldn’t be an accident or something someone stumbled upon. Instead, it should be the physical equivalent of those annoying pop-up boxes making sure that continuing is the desired action. This almost feels like the equivalent of long-term nuclear waste warning messages: as one gets closer to fully overriding some protective measure, it should be more obvious that they probably shouldn’t be doing that.

(In)Conveniences

Safety should be easy. Doing the right thing should be easy. Doing the wrong thing should be difficult (see section directly above) and inconvenient. Trivial inconveniences exist and apply in the fab, even if the rules try to claim they don’t. People will be unsafe if being safe is too much of a pain in the ass.

Some examples of making safety easy and convenient:

• Ensuring safety gear (hard hats, safety glasses, gloves) are adequately stocked in all necessary areas so employees don’t have to go searching for it.

• Creating custom tools that are simple to find, simple to use, and make the job easier. The tool should have a designated storage location where users store it after use.

• Avoiding pressuring employees to work faster than is safe or they are comfortable. Calling them four times in an hour to ask for “status updates” conveys something that feels more like desperation and pressure than urgency and checking in. Being kind and ensuring that safety is a top priority—and reflecting that in actions—is key to encouraging safe behavior.

Better Safe Than Sorry

When in doubt, take a step back and get a better idea. Some organizations use the PAUSE methodology (pause, assess, understand, share, execute), which is great! It gives someone who is uncomfortable a way to assess the situation at hand before continuing.

Everyone should be better safe than sorry (injured, maimed, dead). This could mean stopping work until someone more experienced can work on it, asking for help, or refusing to do something outright because it is a safety concern.

Redundancy

Shit happens. It can come in the form of a failed part, a forgotten procedure, or plain human mistakes. This is why redundancy is built in to the equipment and other methods. We want to have at least two points of “shit happening” before it can affect the worker!

Redundancy comes in the form of interlocks, safety practices, and proper procedures.

Why Does Safety Matter?

Liability

Reportable incidents are defined by OSHA as:

when an employee is killed on the job or suffers a work-related hospitalization, amputation, or loss of an eye

Since I’m not an expert on OSHA violations, I asked Claude Sonnet 5 to summarize what happens:

When a semiconductor fab has a fatality, in-patient hospitalization, amputation, or loss of an eye, the clock starts immediately: employers must report a death to OSHA within 8 hours, and a hospitalization, amputation, or eye loss within 24 hours. Failing to report on time is itself a separate, citable violation under 29 CFR 1904.39.

Fatalities almost always trigger a full OSHA inspection, and severe injuries frequently do — especially in a hazard-dense environment like a fab, with toxic gases, high-voltage equipment, and cleanroom chemical exposure. Employers are also required to preserve the incident scene until OSHA releases it, which can mean idling the affected tool or line for days or weeks while the investigation runs.

If OSHA finds violations, the fines add up fast. As of 2026, serious or other-than-serious violations carry penalties of up to $16,550 each, while willful or repeat violations can reach $165,514 per violation — and a single inspection often produces multiple citations, so totals can climb into six or seven figures. Willful violations tied to a worker’s death can also trigger criminal referral.

The costs that outlast the headline fine are often the bigger operational hit: a spike in the company’s workers’ compensation experience modification rate (raising premiums for years), disqualification from safety-prequalified client contracts, and — for repeat or egregious cases — placement in OSHA’s Severe Violator Enforcement Program, which brings mandatory follow-up inspections. For a fab, where a single tool going down can stall an entire process flow, the downtime from a scene-preservation hold or a corrective-action order is frequently more costly than the citation itself.

I’m unable to find papers rigorously quantifying downtime impact in terms of $/hr, although it’s certainly a lot. Engineers and technicians may have to physically fix the equipment, parts may break and require replacement (money and time), equipment has to get requalified, and line speed drops substantially as a result of all the delays associated with each of those. Time and money are the name of the game!

Reputation

Being known as an unsafe company speaks volumes about the culture, repelling potential talent and pushing them towards places where they know they’ll get to see their family at the end of the day.

The People

Mandatory Clothing

Let’s first dispel the myth that the cleanroom suits fab personnel wear are for the people’s safety. They are not. They are for the chip’s safety. They serve to prevent the wearer’s skin and hair from getting onto the chips. We lose a fair amount of skin and hair every day and any of that falling on a chip can be catastrophic, even existential. Why risk this when you can simply doom your wage slaves to a day of sweat, discomfort, and modesty?

Next, we’ll move onto a fit check. Safety glasses? Check. But those better be ANSI Z87.1 or OSHA will want a word. Z87.1 is focused on minimizing and/or preventing injuries such as “impact, non-ionizing radiation and liquid splash exposures in occupational and educational environments such as machinery operations, material welding and cutting, chemical handling, and assembly operations”. Fabs are a place where liquid can splash, parts can fall through the floor into the subfab below, and both very high and very low pressure environments can fail (also known as explode and implode, respectively) if the right sequence of events happens. I’d want to protect my eyes, too!

Shoe-wise, some folks like rocking the Nike Fab 11s, which is my overused joke for steel-toed boots. (Nike or Jordan really should start supplying cleanroom boots for the collab potential—just imagine a TSMC x Air Jordan.) Fabs are basically just filled with super fancy, super expensive machinery. And machinery comes with heavy parts that people may need to carry around. I’d want to protect my toes from a big block of steel from falling on them, too! (To be clear, steel-toed boots aren’t required at my fab, but can be purchased.)

That’s about it for mandatory clothing. Wear some safety glasses, wear something on your feet, and wear a bunny suit. And gloves! But those are also for the wafer’s protection. What’s between the bunny suit and skin is pretty much up to the person—I’ve seen some questionable shirts while walking through the office!

Sometimes-Needed Clothing

Sometimes special garments must be donned for safety purposes, generally when someone is entering an area where special dangers exist.

Acid Clothing

There are a lot of acids in a fab: hydrofluoric, sulfuric, phosphoric, just to name a few. All of them are harmful to the human body, so we protect it by wearing:

• Tychem suits: This godforsaken suit has zero breathability whatsoever. It feels like it is constantly exchanging dry air for humid air to make the inside of the suit even hotter. Their unofficial motto is “start sweating within 30 seconds guaranteed or your money back!”. I’ve never gotten my money back. In all seriousness, the lack of breathability is by design. I don’t want some horrible liquid leaking through the fabric if it spills on me!

• Face shield: Covers the entire face, not just the eyes. I like my face without any burns or holes, thank you very much.

• Acid gloves: More protection in case you need to handle something around acid.

Electrical Clothing

Arc flashes are no joke and will fry your skin to a crisp if you aren’t wearing equipment. If you are wearing equipment, then at least it’ll be an open casket funeral. Electrical work in certain areas, especially when “hot” (or electricity is live), requires arc flash equipment that is not dissimilar to a bomb suit (which it’s colloquially called by electricians), minus the protection against shrapnel.

Head Protection

Hard hats are required in spaces where you can easily hit your head on something or parts can fall down from above. Some people wear bump caps in confined spaces to protect against bumps, but they are no substitute for hard hats when falling objects are a risk.

SCBA

Robotic dexterity hasn’t caught up to humans yet, so we still occasionally have to do some nasty work with gases and chemical that requires self-contained breathing apparatuses (SCBA; note the lack of a ‘U’(nderwater) since fabs aren’t underwater). SCBA users get fit checks before using it to ensure the mask seals securely against their face. Not coincidentally, it’s also the only time some people shave their beards—this helps the mask better seal against the face, else the hair would cause leaks.

Practices

The practice that keeps you safe 99% of the time—where the 1% is the unpredictable, unexplainable, unstoppable shit that very rarely happens, and much less often affects someone when it does happen—is pretty simple: don’t be an idiot. I stop and ask myself “if someone else did this and got hurt because of it, would I think they’re an idiot?” before doing something potentially dangerous and if the answer is yes, I reassess my strategy and refine my plan until the answer is no. There are other, more official strategies like PAUSE.

Not on the via negativa side of things, being a clear, loud communicator is important when working with a partner. There are many loud noises and dangerous situations that require messages to be received correctly lest the receiver act based on an incorrect message.

My modus operandi is a command-acknowledgement structure of:

1. The person in the dangerous position (p_d) commands the person in the safe position (p_s) what to do. This prevents p_s from doing something that endangers p_d.

   • Ex: p_d says “move robot1 blade1 10 counts forward”

2. p_s acknowledges the command by verbally repeating it, waiting a second, then performing the command. p_d should be in a safe position since they know what’s about to happen.

   • Ex: p_s replies by saying “moving robot1 blade1 10 counts forward”, then waits a moment, then issues a robot command to move 10 counts forward

3. Repeat as needed for the rest of the work.

   • Other commands-responses may be hold-holding, lift-lifting, etc.

Lockout/Tagout

The Lockout/tagout (LOTO) Wikipedia page does a good job of explaining it:

a safety procedure that ensures dangerous equipment is properly shut off and not able to be restarted prior to the completion of maintenance or repair work. It requires that hazardous energy sources be “isolated and rendered inoperative” before work is started on the equipment in question. The isolated power sources are then locked and a tag is placed on the lock identifying the worker and reason the LOTO is placed on it. The worker then holds the key for the lock, ensuring that only that worker can remove the lock and start the equipment. This prevents accidental startup of equipment while it is in a hazardous state or while a worker is in direct contact with it.

This system is incredibly effective provided the worker actually follows it. Equipment is designed so that at least the main power switch can be locked out. Other individual components (such as RF power generators) may have LOTO capabilities on their associated electrical breakers so that the equipment as a whole may operate as normal except for said component; this is to allow troubleshooting or maintenance that requires general power, but doesn’t require that specific component. (To connect this with the Interlocks section, interlocks should prevent any of these bad components from turning on if someone is working on the equipment, but there may be some freak accident, so LOTO is a foolproof way of eliminating the risk.)

LOTO locks can be removed by someone other than the owner under certain circumstances that must be run through the site’s safety lead. If it’s removed without consulting said person and someone finds out, bye bye!

Click here to test your LOTO skills!

Procedures

Procedures are important because they standardize tasks into best known methods to ensure consistency, quality, and safety when performed. By documenting them into an easy-to-read, step-by-step format, workers won’t miss any safety-related subtasks.

The Equipment

Interlocks

From Wikipedia:

An interlock is a feature or device that is commonly used in engineering and safety systems to keep machines, devices, and processes from operating until the guards are in place or the required circumstances are met. When being utilized, interlocks are used to prevent or reduce the chances of injury to the operator, damage to the equipment, and actions being completed in the wrong order or in an unsafe way.

Fab equipment is filled with interlocks of various types:

• Hardware: Enabled or disabled by jumper configuration, these can be physically changed at the tool, but require specialized knowledge of what to move and where.

• Software: Written into the equipment’s programs, these are effectively absolute. There exist ways to trick the interlocks into being satisfied to allow, but these are a very quick ticket to the unemployment line if done improperly.

What are some examples of interlocks? Let’s look at this random interlock card I found on the internet!

Going down the list in a non-exhaustive manner:

• GP OK: Is the gas panel door closed and are any other issues present?

• SLT VLV CLSD: This is the “door” to the buffer chamber where the robot that transfers wafer in and out of the chamber exists. The SV being open could expose the buffer to nasty gases, which could in turn contaminate other chamber if said gases lingered for too long. Overridden by taking a magnet and tricking the sensor that detects the SV being open or closed.

• CUST VAC: Not exactly sure, but this means customer vacuum, so likely verifying the facility is pulling vacuum somewhere.

• RGH PUMP FAIL: Checks if the pump is in a good state?

• CHAM LID CVR: There’s a cover that sits over the chamber lid itself, likely to protect against electricity or other hazards. This needs to be installed for the interlock to be satisfied.

• CHAM LID: The chamber lid needs to be closed because we don’t want to flow nasty gases into the fab air.

• HTR HUB OT: The heater hub temperatures needs to be low enough for the interlock to be satisfied.

Why would someone want to override these interlocks? After all, they’re there for our safety! This is true, but sometimes overriding interlocks is needed to troubleshoot. For example, watching problematic robots move while troubleshooting is very helpful in order to hear noises, feel vibrations, watch for jerky movements, etc. But often times the only way to get a clear view is to override interlocks. The equipment manufacturers understand this and make it easy enough to override interlocks while ensuring it’s intentional.

Click here to play an interlock game!

Alarms

Tools will output alarms (or warnings, which are less severe and disruptive) that will result from abnormal conditions, which aren’t necessarily always a safety hazard, but can be. This may stop the tool from processing to prevent something small from turning catastrophic. For example, a low water flow fault may be the result of a large leak. If the tool kept processing and applying gross amounts of power that eventually connected with the water, then neither the tool nor the equipment engineer would have a very fun day.

Ergonomics

Equipment parts can be heavy and often cumbersome to lift because the equipment’s small footprint doesn’t allow for personnel to get easy leverage on said parts. It can be a jungle inside, underneath, or on top of equipment!

Equipment manufacturers will often include various ergonomic fixtures to help personnel with maintenance. Examples may include motorized cranes that can assist with lifting parts straight up and down instead of forcing personnel to come from the side; fixtures to compress high-pressure gas springs; and specialized tools that fit in certain areas to minimize potential pinch points or awkward angles for employees.

If custom solutions are required, then fabs can outsource to companies who will build tools to their specification. For example, take this lift from Alum-a-Lift:

(Not super-related to safety: fabs often have their own machine shops on-site that engineers can send drawings to for quick prototyping. These have most of the standard shop equipment—mills, lathes, bandsaws, drill presses—so most things can get done. If they don’t have the capability or capacity, the job may be sent to a larger shop that the company works with.)

The Emergency Off Circuit, or The Big Red Button

All equipment comes equipped with emergency off (EMO) buttons at strategic locations to maximize accessibility. When pressed, the entire tool and anything external connected to the EMO circuit shuts off immediately because the physical connection is severed. Uninterruptible power supplies (UPS) will generally be connected to equipment computers to maintain data and allow for proper shut down if needed.

Why would someone want to press the EMO button? Here are some example situations:

• Pyrophoric gas, or gas that ignites into flames when exposed to air is leaking from the tool. Pressing the EMO button will cause all valves related to that gas to close immediately, stopping the leak (this is presuming a weld or something didn’t break, of which there are other solutions since a valve may not be able to stop that depending on where it is).

• A robot is actively pressing against a person who wasn’t aware it was still able to move. Pressing the EMO button will turn off the robots motor and allow the person to move the robot out of the way.

• A water line is leaking water all over sensitive electronics. Pressing the EMO button will turn off valves and stop water flow.

Because power is removed from everything, parts will default to the power-off state, which is generally synonymous with “safest state”. For example, valves that control the flow of gases can be either normally-open (N/O; power off means gas can flow) or normally-closed (N/C; power off means gas cannot flow). Valve states are strategically chosen based on the gas: purge gases that help dilute toxic gases, like nitrogen, have N/O valves so that if power shuts off, the equipment gets purged properly; toxic gases have N/C valves so that they stop flowing immediately upon power off. The orientations are physically manufactured, so the only way that someone can mess up is by installing the wrong valve, which is mitigated by labeling and different colors.

Click here to test your valve knowledge!

I’ve pressed an EMO button once in my life, but that was at my university fab and because of a major operator error. EMO presses at the industrial level are rare because of how well the equipment is designed and the people are trained. You do not want to press the EMO! If you are pressing the EMO, your day is probably really bad or about to get really bad! (I’ll admit that pressing the EMO is kind of fun because of how forbidden it is and its nice tactile feel, so if the equipment is ever off do yourself a favor and get a nice lil’ press in.)

The Building

The building itself—the foundation, the trusses, the waffle floor—isn’t anything special, but rather the features and systems that interact with the equipment and facilities.

Fire Suppression Systems

NPFA 318: Standard for the Protection of Semiconductor Fabrication Facilities “presents requirements to safeguard facilities containing cleanrooms from fire and related hazards to protect against injury, loss of life, and property damage”. (Note the linked document is from 2006 because I couldn’t find an updated one on the internet.) Here are some nice features:

Optical flame detectors that will respond to the flame signature of silane shall be provided at silane gas cylinders in the open dispensing systems described in Section 8.4. Activation of detectors shall result in the closing of the cylinder automatic shutoff valves described in 8.1.2.

A local visual and audible alarm shall be provided to indicate activation of any interlock. [Intentionally silencing the alarm is much better than not hearing it at all.]

Tools utilizing hazardous chemicals shall be designed to accept inputs from monitoring equipment. [This allows the monitoring equipment to tell the tool to shut off because something is wrong.]

They seem to really care about silane given it has two subsections and one main section devoted entirely to the gas!

Gases

Some gases that fabs use have very low median lethal doses (arsine to phosphine), low flammability limits, or are asphyxiants. I’ve heard the high-(LD/LC)₅₀ gases referred to as “two-step gases” because once you inhale them and take two steps, you’re dead. Here are the most toxic gases in a fab (and an explanation on LC₅₀ and LD₅₀):

Gas detectors are placed at strategic locations to maximize detection capabilities so the faster it’s detected, the sooner it can be stopped and fixed.

Some places that gas detectors are located:

• Exhaust: Vacuum chambers and places that have one barrier of protection between hazardous gases and humans (e.g., gas panel gas lines) have exhaust pipes that are connected to vacuum pumps, creating slight negative pressure and pulling in any gases in the volume it’s connected to. Gas detectors are connected to these pipes and constantly monitoring for any nasty gases. If it detects any, there is definitely something wrong, so it sends a signal to stop the tool and/or stop the gas flow on the facility side.

• Gas Cabinets: Similar to the exhaust pip detectors, some gases are housed in standalone cabints to improve detection capabilities (smaller volume means a larger concentration per a constant leak) and improve safety.

Some gases are housed in double-walled pipes, where the pipe that carries the main gas is enclosed within another pipe filled with an inert gas, such as nitrogen. This creates an added layer of safety and immediately dilutes the dangerous gas.

General Mortality and Disease Rates for Fab Employees

I’ve heard some crazy health-related rumors throughout my fab life: exposure to photoresist will cause men to become sterile; photoresist will cause women to only have female children; working in a fab will make you go clinically insane because of the long hours and insane pressure and mind-boggling physics that make magic seem real.

That said, fabs are incredibly safe thanks to the equipment manufacturers and fab companies working together to minimize their liability take care of their precious employees. Here’s Claude’s response after researching semiconductor fab deaths and major safety incidents:

Looking at recent history, the semiconductor industry’s fatal incidents fall into three distinct categories, and conflating them overstates the risk of actually working in an operating fab.

Acute accidents inside a running cleanroom or gas system are rare but do occur. In January 2013, a contractor died and four others were injured after a hydrofluoric acid gas leak at Samsung’s Hwaseong plant. In Japan, older incidents include a 1989 CVD gas cabinet explosion and a 1996 pipeline-cleaning explosion, both fatal to a single worker.

The more common source of fatalities tied to “fabs” in the news is actually construction of new facilities, not fab operations. TSMC’s Arizona buildout has seen at least one confirmed death — a worker killed by an over-pressurized waste tank in May 2024 — plus disputed reports of additional construction-site deaths from labor representatives. Samsung’s Taylor, Texas project has also seen a fatal fall involving a subcontractor and a separate contractor death, both under investigation. Wolfspeed’s Durham, NC site saw a contractor die the day after a workplace accident in 2026.

A third, distinct category is long-latency illness from chemical exposure, which shows up in the data as elevated disease risk rather than a single incident. A cohort study of South Korean semiconductor workers employed between 1998 and 2012 found elevated leukemia mortality specifically among female wafer-fab line operators, though not across the workforce as a whole.

Taken together, the pattern suggests that acute fatalities on an active fab floor are genuinely uncommon; most of the fatal incidents making headlines happen during the construction of new fabs, which carries risks more typical of large industrial construction generally rather than semiconductor manufacturing itself.

See Also

• Slowing Moore’s Law: How It Could Happen

In Dino Buzzati’s The Tartar Steppe, young Giovanni Drogo is sent to Fort Bastiani, a remote fortress on the edge of the Tartar Steppe. Like every other young officer in Italy, this is the last place he wants to be. There are no women, no peers, no fun. Other cities are considerably more attractive in terms of amenities and people. But like every other young officer in Italy, he accepts his assignment, albeit begrudgingly, having no other choice.

Arriving at Fort Bastiani, Drogo is met by the tenured soldiers on post and is quick to express his intent to leave this place as soon as allowed. This is met with a hearty chuckle and warning that it won’t be that easy. Drogo soon learns about the reason for the fort: the Tartars. An ancient people that are “legend more than anything else”, they pose a threat to Italian sovereignty and all it stands for, requiring soldiers to guard the border they would invade from.

Eventually the tailor, Prosdocimo, is introduced, and explains how he plans to leave within only a few short days:

I am here on a merely temporary basis. I expect to leave any day. But that the colonel won’t let me go.

The tailor’s assistants then explain the situation to Drogo:

Did you hear him? Do you know, sir, how long he has been in the Fort? Fifteen years, sir, fifteen accursed years, and still he goes on repeating the same story—I am here on a temporary basis, I expect to go any day... But he will never move from here. He and the commanding officer and lots of others will stay here till they’re done—it’s a kind of illness. You’re new, sir, watch out—you’re newly arrived; watch out while there is time.

Drogo naively replies:

I am here for only four months. I haven’t the slightest intention of staying.

Drogo eventually changes his mind without realizing it, rejecting a chance to be sent back to the city and sentencing himself to a life at the Fort.

Drogo has decided to stay; what keeps him there is a longing, but more than that alone—for perhaps the heroic cast of his thoughts itself would not have sufficed. For the time being he thinks he has done something noble, and is genuinely surprised to find himself a better man than he had thought. Only many months later, looking around him, will he recognise the paltry ties which bind him to the Fort.

Four months passing with the monotonous rhythm of routine duties had been enough to entrammel him:

He had got used to the trips every so often with Morel to the nearest village...to the wild races up and down the level ground behind the Fort...to his room, to reading quietly at night, to the crack in the ceiling above his head...to the creaking of the door when it rained...All these things had now become part of himself and it would have hurt him to leave them. But Drogo did not know, he did not suspect, that his departure would have been an effort nor that life in th Fort would swallow up the days one after another, one exactly like the other, at a giddy speed. Yesterday and the day before it were the same; he could no longer have distinguished one from the other. Something which happened three days befoore or three weeks before seemed equally distant. Thus unknown to him time fled on its way.

Time continues to seep by in this manner, days and weeks and years blending in with one another, the quotidian routines lubricating the flow and blinding Drogo to its passing. His desire to stay there is fueled by the thought—calling it a chance is a long shot, historically speaking—of the Tartars invading.

Meanwhile time was slipping past, beating life out silently and with ever increasing speed; there is no time to halt even for a second, not even for a glance behind. “Stop, stop,” one feels like crying, but then one sees it is useless. Everything goes by—men, the seasons, the clouds, and there is no use clinging to the stones, no use fighting it out on some rock in mid-stream, the tired fingers open, the arms fall back inertly and you are still dragged into the river, the river which seems to flow so slowly yet never stops.

Talks of war with the Northerners make their way around the Fort, followed by predictions of when it will happen. Six months some say, 10 others, and two years yet others. But time continued to move, turning Drogo’s a “sad yellow colour” and a regular “disheartening feeling of fatigue”. The invasion is the final straw, causing him to faint with excitement after witnessing battalions marching in.

The penultimate chapter involves Drogo being taken from the Fort as his fellow soldiers prepare the battle of a lifetime: their lifetime. The one most have been waiting for since arriving 10, 15, 30 years ago. The one many had foregone a family and memories for. The one they’d believed was on the horizon, but never came to fruition. Drogo helplessly watched as he was carted away for his own good, away from his sad life’s purpose.

Allegories

Drogo’s story is an allegory for two concepts: waiting and complacency.

Waiting

Many people wait for major events to just happen, blinding themselves to life and other less significant, but arguably still important, events passing them by. The anticipation of what could be, yet isn’t, grips their attention and energy, pulling a majority of each towards the could-be event. Responsibilities are neglected, one-time opportunities forgone, and ambitions and desires from before forgotten.

The opportunity cost in these situations is immense, and a bird in the hand is often better than two in the bush. Much is to be lost and little to a lot is expected to be gained. People often overestimate the impact—both positive and negative—an event will have on their life. Those waiting for the big event expect a drastic increase in their quality of life, happiness, and so on, but the lucky few who do get to experience their event are left still wanting more, unsatisfied with what was supposed to be a massive impact on their life. While Drogo doesn’t necessarily fit the bill since he didn’t truly experience his life’s big event, his opportunity costs can still be examined. He never experiences having a love life, much less a family. He never forms and develops a true career—sure, he attains the rank of major, becoming second-in-command of Fort Bastiani, but what does that really mean? He never ventures outside of the corridor between Fort Bastiani and his home city. Meanwhile his old friends have avoided the waiting disease:

his old friends have made careers for themselves, occupy important positions...his brothers are constantly away from home; one has married and lives in another city, another still travels; there are no more signs of family life in the living rooms

All of this in anticipation of an invasion by people no one truly knows exist (read: in anticipation of what is probably nothing). In the end Drogo’s “prediction” (if it can be called that) paid off in the sense of it occurred, but he never experienced it and incurred a rather extreme cost in waiting for it: his life. The entirety of life was wasted waiting for that moment, and in the end he had nothing to show for it.

Complacency

A friend once mentioned that ambition tends to decrease with age. This is true as a generalized statement. Ambition also (generally) decreases as a function of time spent in a condition. This can be reworded to say complacency (lack of ambition) increases as a function of time spent in a condition. These effects can be cumulative and self-perpetuating: as one ages, ambition decreases, making it less likely to leave the current condition, increasing complacency. The hurtful cycle continues on, leaving the person trapped in a sub-optimal, less-than-desired life.

The only way out is a burst of intrinsic motivation or an impactful, external event that forces a reevaluation of the situation, values, and goals. The motivation spike is difficult to come by. It can happen at random with little to no stimuli, or can be initiated if the trigger(s) are known and understood. Reliability of this is hit-or-miss, with spikes varying in size despite identical stimuli, random or coordinated.

The “impactful, external event” is required to be random, or at least unknown up to the time of its happening. Near-death experiences often result in this: workaholics have a heart attack, barely survive, then reevaluate their decision to work 80 hours a week for marginally more pay; soldiers lose a limb (or multiple, or their life), then reevaluate their decision to serve while supporting a family. These events can be experienced vicariously by others, albeit the effect may not be the same. A parent dies from a stress-induced heart attack, causing the children to reevaluate and reconcile career and life goals.

Drogo quickly becomes complacent after arriving at Fort Bastiani. He makes excuses for why he shouldn’t leave, and ultimately rationalizes his decision to stay at the decrepit, long-forgotten Fort. He occasionally recollects the ambitions from his former life, but quickly dismisses them. His increasing rank is not due to hard work and diligence, but (seemingly) earned simply through time in the service.

Suggestions

Note that the following suggestions are entirely subjective, and based on personal goals and tolerances.

Waiting

How long one waits for “the big event” should be determined based on probability, risk tolerance, and expected value earned from said event occurring.

Probability is obvious: higher chances of the event happening justify both waiting itself and waiting for longer durations.

Risk tolerance is related to probability in that if the event doesn’t happen, some negative cost is incurred, whether it be real or in opportunistic terms. Regret is a powerful force, causing more of an effect when inaction is the driver. From Gilovich and Medvec’s The Experience of Regret: What, When, and Why:

When people are asked about their biggest regrets in life, it seems that they tend to focus on things they failed to do in their lives: “I wish I had been more serious in college”; “I should have told my father I loved him before he died”; “I regret that I never went to Europe.” As troubling as regrettable actions might be initially, when people look back on their lives, it seems to be their regrettable failures to act that stand out and cause greater grief.

If the event isn’t realized, how much regret will be felt at the fact that the time was “wasted” (note that just because something didn’t work out doesn’t mean it wasn’t a valuable experience—lessons can be learned, skills honed) and could have been better spent elsewhere (again, the experience during that time may be more valuable in the long-term). The intensity of the regret is likely to increase based on a few factors.

Longer waiting periods generally result in more opportunities lost along the way, assuming they aren’t being pursued concurrently. Waiting a week loses little, but a year? A decade? Much can be lost in time periods on that scale, and can very easily be things that can’t be recovered.

As mentioned, other tasks can be done while waiting. These may be supplementary to the main event or completely unrelated. If no progress is made towards any project while waiting, regret will increase.

Expected value is another obvious aspect: how much utility will the event provide? More utility justifies both waiting itself and waiting for longer durations. Interim or alternate activities’ expected values should be considered before waiting for too long. Intentionally underestimating the expected value also provides value. Unless there is legitimate data or similar available, expected values are likely to be overestimated, causing some degree of disappointment if realized. If the the expected value was underestimated initially, then a happy surprise is in store.

In sum, if waiting is an option that’s being considered, look at the probability the event occurs, personal risk tolerance and past experiences with inaction regrets, and the expected value the event is likely to return.

Complacency

Complacency correctly has a negative connotation associated with it. It’s a form of satisfaction coupled with laziness and lack of ambition towards greater goals. Drogo’s mistake and problem was a potent combination of complacency and ignorance, the former of which requires some type of conscious recognition of the situation, the latter of which requires no conscious recognition of the situation. If the ignorance is in effect for too long, the likelihood of complacency occurring is greater. Excuses take effect: “I’m too old”, “It’s too late”, “It’s not worth it”.

Catching this early and critically examining the situation is the best mitigation and/or prevention strategy, and the only way to do this is to regularly remind and review one’s ambitions and status.

Reviews are best done on a time-based schedule, with some immediate reviews being done on an as-needed basis. Depth, and by extension intensity, should increase by time period, i.e., daily reviews are short and simple, weekly or monthly longer and more critical, and bi-annual or annual the longest and most critical.

An example review schedule could be as follows, adding in other frequencies (quarterly, bi-annual) as seen fit:

• Daily:

  • Are there any current pressing issues? Write it down to review at the next monthly.

  • Work on installing trigger-action plans.

  • Visually review all goals to keep them in mind. This can be done by printing out the goals, keeping them as a wallpaper/screensaver, etc.

• Monthly:

  • Review pressing issues from daily review and outline plan to resolve issue.

  • Review current status of all short-term goals, the more important long-term goals, and methods of achieving both. Can the methods be improved?

• Annually:

  • Review pressing items over last year and rate resolution (is it completely resolved, improved, not improved, worse?) and method of resolution. What works can be noted to be used for future issues.

  • Review current status of all long-term goals and methods of achieving them. Can the methods be improved?

I rarely visit the testimonial section of websites. Too often they vaguely boast about how they got a [insert disadvantage here] person to do X in Y time. It’s phrased in such a way that nobody else could have done such a thing, that only they have the shortcut to get the magic pill to get a free lunch. They tout their program’s effectiveness on getting clients [insert goal here], despite it being based on something that almost every other program is: progressive overload, basic periodization, focused nutrition, and recovery schemes. Choosing the principles of a program based on the goals is not revolutionary nor innovative.

My problem is not with beginners making progress. My problem is with coaches being glamorized and rewarded for such easy-to-progress clients. Getting an overweight client to lose 10 kg is easy; getting a lean client down to 8% bodyfat while maintaining strength is difficult. Getting an untrained client to a bodyweight-added squat is easy; getting a hard-working client to a 140 kg front squat is difficult. Getting a new runner to a sub 8:00-minute mile is easy; getting a Jack Daniels follower to a sub 5:00-minute mile is difficult.

No one bats an eye at the former part of those three examples, while the latter part turns heads and demands respect, both to the client and coach. Yet we incorrectly say someone is a good coach when they get anything up to or past the first achievements. Virtually anyone can lose 10 kg, squat their bodyweight on a barbell, and run faster than an 8:00 min mile. The formula is incredibly simple: find a well-reviewed program, fervently adhere to it for at least one year, and boom! the goal has been achieved. This logic cannot be extrapolated for loftier goals because progress is logarithmic with both time and effort: intelligent programming must be implemented alongside hard work, patience, and consistency.

It’s here that the worthwhile coaches make themselves known. They have to know their client’s training history and lifestyle and how they react to certain stimulus; they have to motivate the client in times of despair (this should mostly be the client’s burden, but some still falls on the coach); they have to instill trust in the coach-client relationship when things don’t make sense or progress has stalled; they have to hold the client accountable (again, this is mostly on the client). These are not easy skills to attain or maintain. They require vast experience across a breadth of clients, empathy and sympathy, the ability to connect on a personal level, and a drive to see the client through to their goal despite an obstacle-ridden road. Coaches that have these skills are doing something right.

The ability to change behavior for the long-term is an incredibly underrated skill that any coach—exercise-related or otherwise—can and should have. Most people who are deep in some hole have unproductive behaviors they’ve instilled over many years, behaviors that are ingrained in their life and habits and conscious and unconscious choices, behaviors that will require years of hard work to unlearn, behaviors that may be permanently installed and need to be fought against.

A good coach can help with all of that. There is nothing new under the sun—no revolutionary training methodology has been invented or discovered in the recent past (besides GLP-1s)—but crafting and feeding ideas and change in a palatable, personalized manner to a client may be the difference between them continuing to claw at the hole’s walls and discovering the ladder that had been behind them the entire time. Books and courses are often one-size-fits-all for the good reason that most of the population will receive and respond to it well. The problem lies in neglecting the fact that some respond significantly differently to suggestions: what works for persons 1-9 won’t work for person 10. Person 10 keeps trying to fit that square peg through the round hole that Dr. Coach Guru ® swears is foolproof and works for everyone, when the square hole is just a perspective and schedule change away. Epiphanies happen more often than one would think. Concepts clicking because of a small tweak in an explanation happen. The wheels of progress can continue to turn because of a minor adjustment or a small detail being noticed.

Identifying the Liars, the Cheats, and the Fakes

I propose standards for celebrating coaching success. They aren’t rigid nor well-defined, but it will be obvious when it should be applauded. Anything that falls short is ignored and doesn’t count towards testimonials—come back again when you’re tall enough to ride. Anything that surpasses the bare minimum is given a nod of approval—good job, but keep working. Anything that greatly exceeds expectations is celebrated and rewarded with attention—teach your methods to the world. Clients deserve to know when they’re getting the real deal and not a charlatan.

Standards raise the bar and force wannabe coaches to improve their craft or get out. I bet that exit line’s gonna be long.

You can find people gathered around in an effort to experience novel qualia: holographic chocolate, the thermal grill illusion, feeling like their arms are sinking through the floor. Some attendees walk around in bird jackets, kindly explaining they represent the great grey shrike, a bird that impales its dead prey on thorns or barbed wire to store or tear apart. Bouncing from conversation to conversation is expected, even encouraged—there is only so much time available (interpret that in both ways) that one shouldn’t waste it on bad conversations or spend too much time because diminishing returns exist. Sessions are hosted by enthusiastic speakers wanting to share their thoughts or experience with the group. Rooms are often packed, some sold out, some more desolate due to the...specificity. There’s something for everyone. Sometimes too much. Slots in both time and location are limited and force attendees to choose and consider opportunity cost, weighing session this versus session that versus conversation this versus conversation that. They decide at some point, rarely disappointed in the result, but always happy in the moment, or so the smiles and shine in their eyes say.

Spending time in The Bay—and Lighthaven particularly—imbues a sense of “I’m not doing enough” or, for certain people, “I’m doing exactly what I should be”. The crowd at Lighthaven never comes across as humbly bragging like described in the Bay Area House Party series, but instead passionate and confident in their choices of risky career decisions and quirky hobbies. For the former set of people who are receptive, it serves as a jumpstart, wake up call, reminder, and kick in the gut combined and delivered in one convenient weekend package: it jumpstarts motivation; wakes those up who didn’t know that things are happening; reminds those that did know that things are happening that things are continuing to happen; and kicks everyone in the gut that some people are just built different when it comes to energy, ambition, risk tolerance, and sheer intelligence. One would think that envy is a natural emotion to feel because of all of this, and for some it may be! But envy feels zero-sum, where the enviable are on the positive side and the envying on the negative. It’s not like that. Status exists only in the minds of the beholder; misplaced fear is the only thing stopping a discussion with any of the microfamous celebrities (lack of availability may also be an issue!).

Icebreakers were easy thanks to the abundance of topics available to talk about and kindness and patience of the attendees. A go-to was “what’s been your favorite session or conversation”. A few people were caught off guard and at a loss due to sheer volume of answers; others promptly took out their soapbox, stepped on, and delivered a monologue worthy of an award. The vibe-coded social media app and its integrated LLM facilitated finding “your people”, maximizing efficiency for those who were there with a purpose while still allowing high variance for those who are a bit more daring and lax.

And finally, an abbreviated list of conversations I had for posterity and showcasing:

• Why China isn’t super AGI-pilled and is instead focusing on integrating AI across their economy for better automation. Do the American labs have the right approach in going straight to AGI and hopefully picking up intelligence gains along the way, or should they slow down a bit on the way?

• Why certain firms aren’t as good as they’re hyped up to be. Are the mythical firms of Jane Street and RenTech really as awesome as people make them out to be, or is there a strategy around some extra hype to attract more talent than they otherwise would?

• What I think of Terafab. Will the strategy work? What are the bottlenecks in the fab space and being successful? How much does the supply chain matter? (To answer it explicitly, I think Terafab is unlikely to succeed (<10% sounds about right given the stated goals), but Musk is by far the person who maximizes its likelihood of success.)

• What my job is like at a more granular level. What I do day-to-day, equipment manufacturers, etc.

• What makes a good hobby (the answer is progressable, social, healthy, and competitive, or at least some combination); how fortunate we are to have good hobbies; what we can do to

• Cryonics as related to term and whole life insurance.

• Culture and atmosphere in a frontier AI lab. How intense it is, the weight they feel on their shoulders, the long hours.

• Chinese EVs and their small, tasteful features and how they’re purchased. Apparently customers purchase EVs and then they’re built, rather than a company producing a batch and praying they’re all sold while it sits on the books.

• Mormonism, their love of sugar, and why trampolines are so prominent in SLC.

• Lucid dreaming and why it’s such a good idea to start.

• One-on-ones and my strategy around them, their efficacy, improvements, and why people should do them if they’re not already.

• Hugs and patting on the back. Does a pat on the back feel more “bro-ey” and no pat more personal and warm? I say yes, others no.

• National security apparatus and their awareness of AI systems and rate of improvement.

• Why Alex Bores is so important to AI going well and what can be done to support. I donated $500 to his campaign thanks to the session’s convincing ideas.

• Film recommendations. I quickly realized just how poorly watched I am as my two conversation partners rattled off the names of films and directors I had never heard of. So much screen time, so little actual time!

• Why is sex so stigmatized in certain communities and what can be done to improve it?

• “Flooding” as a method of reducing anxiety.

• Community building by way of convincing friends to move into nearby neighborhoods (preferably walking distance) and then throwing awesome parties to keep it going. Why don’t people do more of this? We have to live somewhere, so why not right next to our friends?

• Normalizing athletic achievements against technological and methodological improvements (shoes, training, nutrition, etc). Does the 2026 sub-2-hour marathon really count given the shoes and everything else?

• Doing good for AI in the world, even the labs and technical work are out of reach intelligence- or experience-wise. There are plenty of low-hanging fruit for altruistic efforts in areas that others won’t touch.

• The horrors of pork farming (amongst others). Tech bros attend conferences dedicated to finding ways to make pork eating more prevalent and attractive, like trialing communications that try to convince younger generations that authentic insert-ethnicity-here cuisine is made with pork.

The 1999 cult classic Boondock Saints has an unforgettable scene about halfway through the movie where the FBI agent (played by Willem Dafoe) hot on the Saints’ trail comes across a bullet-riddled crime scene and begins to give a play-by-play reenactment of what happened. (I’ll spare the intermediate details, but it involves a stun gun, a whole lotta bodies dropping, and some badass slo-mo shooting.) The real magic is in Dafoe’s final monologue where he (incorrectly) relives what the Saints were met with when leaving the crime scene:

They exited out the front door. They had no idea what they were in for. Now they’re staring at six men with guns drawn. It was a fucking ambush. This was a fucking bomb dropping on Beaver Cleaverville. For a few seconds, this place was armageddon. There was a firefight! [There then proceeds to be a massive shootout between the Saints and their assailant.]

Dafoe was accurate except for one thing: it wasn’t six men, it was one. One guy did what an FBI agent thought six could. One guy almost took out the deadliest vigilantes in Boston and still lived to tell about it.

In the Boondock Explorers, the Boondock-Saints-knockoff film starring Ernest Shackleton as an Antarctic explorer and Willem Dafoe as an FBI agent hot on his trail for inexplicable reasons, Dafoe has a similar explanation after finding the remnants of the Endurance‘s crew’s belongings scattered throughout the ice and the ship itself sunk to the bottom of the Weddell Sea:

They exited out the front gangway. They had no idea what they were in for. Now they’re staring at six animals, teeth drawn: a yeti, a leopard seal, a polar bear (how did that get here?), an emperor penguin, a killer whale, and an albatross. It was a fucking ambush. This was a fucking bomb dropping on Shacky Shackleville. For a few seconds, this place was snowmageddon. [There then proceeds to be a massive battle between the Explorers with harpoons and guns and the army of animals.]

Dafoe was still accurate except for one thing: it wasn’t animals, it was the ice and wind that together destroyed a well-built ship and drove 28 hardened men to the brink of starvation and death, all without batting an eye.

The ice was unbreakable, insurmountable, defiant to their attacks and entreaties, mocking them the same way a club bouncer ignores the midget at his waist trying to move him aside to get into the exclusive venue:

A full head of steam was raised, all sails set, and the engines put full speed ahead in an attempt to break through to the crack. For three hours the ship leaned against the ice with all her might—and never moved a foot.

It was endless and expansive, as Lansing put into perspective:

The Endurance was one microcosmic speck, 144 feet long and 25 feet wide, embedded in nearly one million square miles of ice that was slowly being rotated by the irresistible clockwise sweep of the winds and currents of the Weddell Sea.

The “speckiness” is unfathomable. The Endurance’s size compared to the ice was one to around 8 billion (3600 ft²:1 million mi²), or one pixel out of 1000 4K monitors, one grain of sand out of a cubic meter of beach, one second out of 250 years. Good luck getting rescued if your ship sinks and you have to hike out. Oh wait...

But seriously, imagine it. The desolation. A vast wasteland of white and cold that does not respond to anything except sun and wind and ocean currents, three things humans have no control over. It doesn’t care what happens, it just does. This kind of environment is what changes an agentic man into a believer of God and the heavens, the goodness of which depends on the direction of the ice. Its scale is magnificent, towering over them and asserting dominance with ease. Lansing makes this—the fact that the crew had zero control over their environment and thus fate—abundantly clear in his descriptions and recounting of the events.

Other times the ice was kind and forgiving. It acted as a barrier between the men and certain death in the frigid water. It served as a vast, easy, free-going, safe transport where they could relax and play games, getting them closer to civilization without any work when the wind and current were in their favor.

The Crew

Understanding the crew gives a contradictory explanation on how they were able to survive. On one hand, some of the men were veterans of the polar variety, being handpicked and experienced by Shackleton himself. Others were, well, less vetted than they could have and should have been, which is shocking given a) the trip is one of the first of its kind, b) Shackleton cared a lot about the leaders, and c) going to Antarctica is a crazy endeavor, especially in 1914, that requires certain skills and personality to do well at:

In the matter of selecting newcomers, Shackleton’s methods would appear to have been almost capricious. If he liked the look of a man, he was accepted. If he didn’t, the matter was closed. These decisions were made with lightning speed. There is no record of any interview that Shackleton conducted with a prospective expedition member lasting much more than five minutes.

The sheer volume of applications was no surprise if the urban—or some may say, antarctic—legend is true of Shackleton’s advertisement in looking for men to join:

Men wanted for hazardous journey. Low wages, bitter cold, long hours of complete darkness. Safe return doubtful. Honour and recognition in event of success.

While this is almost certainly fictitious given the fact the original ad was never found, it might as well have been true given the men (and three girls) who applied. It also goes super hard, is very recognizable, and can be mixed and matched for any job. Lansing further explains that:

almost without exception, these volunteers were motivated solely by the spirit of adventure, for the salaries offered were little more than token payments for the services expected

I’m all for judging books by their covers, but Shackleton’s haste is perhaps a step too far! Is the spirit of adventure really the best litmus test for who would fare well? Almost certainly not. Where does the adventure start and begin for these scallywags, some of whom had never left the country of England before? Does it start at the southern tip of Argentina, where if they take one more step, they’ll be the farthest away from home they’ve ever been. That mindset would send them home at the first sign of trouble, ticking their adventure box while making their pants brown. Or is it once they make it to the South Pole and meet Santa Claus? Putting my Shackleton hat on, I’m looking for crew members who aren’t there just for adventure, but also the idea of pushing human progress forward, of charting the unknown and being the one of the first to do so. Mindsets like this are what push people and expeditions forward, while almost guaranteeing a similar intensity and drive if things go wrong and they need to travel hundreds of miles and kill hundreds of cute, cuddly penguins.

And yet despite this haphazard interview process, (spoiler alert!) they made it home. This kinda-unqualified group of men were able to muster up what was needed to survive: massacring innocent penguins and seals, pulling heavy-ass boats across miles of ice (and not the smooth, slippery kind!), and eating some pretty nasty shit because calories are calories when you’re trying to live.

This gives me hope that I, too, can survive if the time came. I’m not an outdoorsman. I’m not the handiest of people when it comes to tools, home repairs, or anything of the sort. I’m not really an inventor of truly new things. But zooming out, I’ve probably consumed more knowledge—accurate, well-researched knowledge—through my perusings of Wikipedia and the internet in general in the past few years than any of these crew members had throughout their entire lives. Does this give me an edge? Maybe. I sure hope so. It’s been said that necessity is the mother of invention, but people forget about the father: survival.

The Decisions

One of the times I’ve been closest to a really bad day (outside of driving on one-lane highways at high speeds with no median) was on a single-day hike to King’s Peak. My hiking partner got altitude sickness while some nasty thunderclouds were forming above us, putting a rather uncomfortable image in my mind of me having to help a soaking wet, hypoxic, weak person through a slippery, full-exposure rock field while temperatures were in the 50s. Not fun and not safe! I assumed command of the two-person democracy with Gaddafian force and made the decision to get the fuck out of there ASAP—no stopping, no second-guessing, nothing. There was some power to this that made me feel better and safer; we just committed to what we decided was the best course of action and went for it without a second thought.

Shackleton’s orders were followed both because he was the leader and because he was a great leader. Being the leader gives followers a natural person to look to in times of peril or indecision. Being a great leader gives people comfort in the decisions being made. These two combined make for a wonderful combo when stranded in the middle of some of the world’s most inhospitable land with no communication and limited food.

That’s not to say the decisions made were easy.

Dogs and even puppies had to be killed:

Tom Crean, tough and practical as ever, took the younger puppies and Mrs. Chippy some distance from camp and shot them without a qualm, but it was Macklin’s duty to destroy Sirius, and he could hardly face the task. Reluctantly he got a 12-gauge shotgun from Wild’s tent; then he led Sirius off toward a distant pressure ridge. When he found a suitable spot he stopped and stood over the little dog. Sirius was an eager, friendly puppy, and he kept jumping up, wagging his tail and trying to lick Macklin’s hand. Macklin kept pushing him away until finally he got up nerve enough to put the shotgun to Sirius’ neck. He pulled the trigger, but his hand was shaking so he had to reload and fire a second time to finish the puppy off.

A foot had to be amputated. Food had to be rationed to the point of extreme hunger and made from some not-so-appetizing parts:

For supper that night, Shackleton ordered Green to put some lumps of blubber into the seal meat stew so that the party might get accustomed to eating it. Some of the men, when they saw the rubbery, cod-liver-oil–flavored chunks floating around in their “hoosh,” meticulously removed every trace.

But later they weren’t so picky:

But the majority were so hungry they were delighted to gobble down every mouthful, blubber included.

These decisions had to be made because it meant death otherwise. The dogs would require food. The foot would become a hindrance to the rest of the body. Food was scarce. So in some sense, these weren’t necessarily decisions in the do-it-or-don’t way, they were requirements to live.

The Mindset

Sheer fucking endurance—the lowercase e version—was the name of the game. The Endurance is what took them to Antarctica and endurance is what brought them back home. Unfortunate event after unfortunate event left them in ever-worsening situations: getting the ship stuck in the pressure ridges and being unable to budge; the ship getting crushed and eventually sinking; the ice floes separating without warning; drifting in the wrong direction; making it to islands that were uninhabitable; landing on the wrong side of the final island and having to hike and climb and bushwhack on a path no one had ever before traveled with minimal food and drink available. The hits just kept coming and they just kept taking it on the chin like the badasses they were.

The hits were only one part of it. The uncertainty was another. While they had their sextant handy for finding their latitude and longitude, they didn’t have a genie (Yeti?) in a bottle that would grant their wishes of the wind and current to be in a certain direction at a certain speed. Some days they found they had traveled at a good clip in the right direction, others gave bad news. The hits were variable in both frequency and intensity, forcing them to take advantage when the opportunity presented itself or forced their hand.

With the bravery and resilience came optimism and a sense of happiness, even after their ship had sunk:

Nevertheless, there was a remarkable absence of discouragement. All the men were in a state of dazed fatigue, and nobody paused to reflect on the terrible consequences of losing their ship. Nor were they upset by the fact that they were now camped on a piece of ice perhaps 6 feet thick. It was a haven compared with the nightmare of labor and uncertainty of the last few days on the Endurance. It was quite enough to be alive—and they were merely doing what they had to do to stay that way. There was even a trace of mild exhilaration in their attitude. At least, they had a clear-cut task ahead of them. The nine months of indecision, of speculation about what might happen, of aimless drifting with the pack were over.

When life gives you an ice floe, make snowballs and have snowball fights with your lads. Understandably, a positive mindset was Shackleton’s overaching goal during their effort to survive:

Shackleton was concerned. Of all their enemies—the cold, the ice, the sea—he feared none more than demoralization.

Positivity is what laid the foundation for the endurance that ultimately saved their lives. Dissent and dissatisfaction were quickly squashed whenever it reared its ugly head, else it would spread and infect others quickly. The crew—and most importantly Shackleton—recognized the perniciousness of demoralization and banded together when needed to keep spirits high.

Shackleton, however, was not without his own faults and mistakes, often caused by his excessive hubris and optimism:

Like most of the others, he [Greenstreet, a crew member] considered the laying in of all possible meat the prudent thing to do, as any ordinary individual might. But Shackleton was not an ordinary individual. He was a man who believed completely in his own invincibility, and to whom defeat was a reflection of personal inadequacy. What might have been an act of reasonable caution to the average person was to Shackleton a detestable admission that failure was a possibility.

He tacitly expected those around him to reflect his own extreme optimism, and he could be almost petulant if they failed to do so. Such an attitude, he felt, cast doubt on him and his ability to lead them to safety.

The balance of optimism and realism is a difficult one to strike in a situation like this. Being too hopeful, too confident can be blinding to the reality of the situation and lead to doing some dumb shit or not doing some smart shit. Being too realistic would’ve caused them to accept that they were going to die and lead them to dying. Err on the side of being too optimistic, I say!

The way Lansing presents all of this elicits a sense of inspiration, awe, and pride in the ability of humans to overcome the harshest of environments in an effort to survive. Similar feelings come about when I watch the wonderful Landsailor music video or read about people doing massive feats of endurance in austere environments with limited support. Here these men are, hundreds to thousands of miles away from any semblance of civilization, shipwrecked, limited food, clothes in tatters, a harsh winter quickly approaching, and they manage to thrive with positive attitudes, games, and, most of all, hope.

The Review of the Review

Lansing’s main purpose of the writing this was to tell the world of this feat in the best manner possible.

He did a superb job in collating all of the sources he used and evidently keeping them true with little embellishment (at least the internet doesn’t seem to think he was untruthful, so good enough for me!). The writing was to the point and factual and still kept the reader on the edge and excited. I was able to avoid spoilers for the entire time.

What he does best is capture the vibe of everything: the excitement when they first set off; the apprehension that turns into despair as their boat doesn’t move an inch in the thick ice; the rapid cycling between fear and boredom and pleasure while out on the ice waiting for something to happen. He also describes the barren, uncharted landscape these men are experiencing; the food and life on the ship; the rationale behind the decisions being made.

The Reenactment

Not discussed in the book was the 2013 crossing of the Southern Ocean in a replica of Shackleton’s boat that took him from Elephant Island to South Georgia. Impressive. but not the same because the Bannister effect is real. The new schoolers knew the crossing was possible, the old schoolers didn’t. The new schoolers had some idea of what to expect, the old schoolers didn’t. Seemingly impossible tasks become much more palatable when someone has done it before. This is another reason the entire voyage is so impressive: no one had done all of this before. The ice, the shipwreck, everything was new, everything was uncertain. They had to react accordingly to each new development; there was no flipping through the Antarctic Troubleshooting Guide for Inexperienced Explorers to see what to do when ice is splitting open between your feet revealing an icy, wet, cold death mere feet below.

Reenactments like this also aren’t necessarily worth the effort: success shows that more than one person can do it (so it’s not that cool), failure shows it was as difficult as it sounds or a fluke and the OGs got lucky. (Admittedly, I may be a bit harsh and inaccurate here, as people continue to call this story one of the greatest survival efforts of all time. And in true critic fashion, I have nothing to offer up in either the I’m-the-only-one-to-do-it or number-X-to-do-it departments, so I’m basically just a keyboard warrior shitting on guys going super hard and risking their lives in a dangerous, unpredictable place that is passively trying to kill them by the minute.) Of course, there’s a balance to be had between reenacting difficult feats and pushing to find the next big thing. Content creators nowadays feel immense pressure to keep pushing the envelope lest they be considered dull or washed up by their fans and when compared to peers who are doing bigger and better things. This obviously creates an unhealthy environment at some point, but where is the line drawn? Mark Twight, Steve House, and Scott Backes climbed the Slovak (Czech) Direct non-stop in 60 hours. Dangerous, sure, but it sure as hell pushed the limits and showed other what was possible. But I digress.

See Also

• Interactive Map of Shackleton’s Trans-Antarctic Expedition (1914–1917)

• Endurance Designs: “Faithful to the original images, we are dedicated to reproducing Hurley’s spectacular photographs in the highest quality possible and strive to reflect in our products the spirit and sense of adventure of the original “Endurance” expedition members.” My personal favorite is Hauling James Caird, but unfortunately it’s sold out with no reprints possible. I instead bought a shirt.

I

New fitness practitioners wonder why exercise is soooooo boring and when it will get fun. They are soooooo bored of doing the same boring movements over and over and decide to switch up. They have a blast at the expense of the progress they made and wonder why they’re still in the same place as when they started. They give up, thinking exercise doesn’t work and it’s just not for them.

II

Programmers long for the allure towards code they experienced during their honeymoon phase as they slowly slip into a boring schedule filled with meetings and not-so-fun projects. The coding gurus on the screen promised fancy machine learning and a six-figure job at a FAANG in only a few short months, but all the programmer can see is basic web development and all they know is their cubicle and all they feel is angst at this being their role for the next 40 years before retirement. They give up thinking there’s little room for progression and go back to their former, dead-end job before the bootcamp.

III

Language learners hit a wall around the A2 level and come to the realization a language can’t be learnt in 30 days—it’s going to take much, much longer. They stare down the abyss of grammar, nuances of pronunciation, and seemingly infinite vocabulary, thinking about how boring it will be to learn and practice all of that. Is it really worth it all anyway? After all, English is the lingua franca of the world. They give up in the face of years of practice and continue being Americ...monolingual.

IV

Investors question where the excitement of making money is and why all those nerds on the Bogleheads forum advocate for index funds. They withdraw all funds, gains included, and begin day trading, taking advice from other online forums that obviously know better than the Boglehead chumps and forming their own theories on why this company is going to blow up and that company is going to tank. They finally have something to talk about at the water cooler, even if money is being lost at an alarming rate. When the principal hits zero and all is lost, all trust in the stock market is lost. They give up, never putting another penny into the market and living paycheck-to-paycheck.

V

Overweight people find counting calories and measuring food to lose weight to be inconvenient and boring, nothing like the smiling people on the television made it seem like. After a few weeks of counting and measuring, the scale shows a slightly lower number than it had in the past few years. They begin to question the efficacy and accuracy of calories in vs. calories out, and after reading some pop health articles on some pop health websites, decide to forget about CICO and only eat based on feel and how much food there seems to be on the plate. “I’ve been counting for a few weeks. My calorie estimates are calibrated by now.” The scale quickly jumps back up to the same weight. They give up, claiming their bodies just aren’t made for losing weight and they’re stuck that way forever.

Themes

These examples all have common themes.

First, the expectations on the experience and progress are unrealistic. Expecting an exciting experience throughout the entirety, or even majority, of a process that requires significant effort is naive and borderline asinine. Motivation ebbs and flows on a daily, monthly, even yearly basis, providing energetic spurts in some periods and depressive valleys in others. What’s important in starting an effortful process is not learning how to work off of those spurts—everyone can do that—but rather recognizing, understanding, and ultimately accepting that there will be periods of boredom, of listlessness, of rationalizing why it isn’t worth it, of wanting to quit followed by the opposite feelings of excitements, of energy, of questioning why this wasn’t started sooner, of getting angry that quitting ever came to mind and learning to work through the down times. Once this is accepted and firmly ingrained in the skull, the depressive valleys can be dealt with more easily. They are now expected and known to go away after some time. How long is unknown, but the finiteness is what matters. In the meantime, progress will continue to be had.

One of my favorite quotes is Mark Twight’s “it doesn’t have to be fun to be fun”. In even more basic terms, an experience doesn’t have to be exciting or dangerous (read: fun) in order to enjoy it. I find enjoyment in hard training in the early morning, but most people I know, athletes included, question my sanity when hearing that. I find enjoyment in watching my boring ol’ index funds slowly but surely gain in value over the course of months and years. There is something in these processes that makes them inherently fun. Knowing that I am making progress towards my goals is the most plausible explanation for why I find satisfaction in these “boring” practices.

Expectation of progress must also be adjusted to realistic and personal levels, all dependent on the process, amount of effort supplied, and environmental factors. Some processes, like learning Japanese as a native English speaker, take significantly longer than others, like losing a few pounds of extra weight. Knowing typical timelines for the specific process can help gauge progress more objectively than arbitrarily expecting to be at point 5/10 four months in when most people are only at point 2/10.

Effort supplied can expedite or impede (or even ruin) progress to the point of exponential growth (e.g., fitness) or logarithmic growth (e.g., language learning and fitness (yes, again)). Daily hard efforts in training will stunt progress significantly, while hard efforts in weight loss (e.g., high volume of exercise and significant caloric deficit) will cause faster weight loss. Learning to regulate effort based on outcomes is a valuable, uncommon skill. Most people tend to have only a few effort settings: off, easy but unfocused, easy and focused, hard but unfocused, hard and focused. But effort isn’t discrete, it’s analog. Effort needs to be adjusted in real-time based on outcomes. Some days need to be easier, some harder, some off. It takes time to understand personal responses, requiring this skill should be trained and reviewed often. Note progress and recovery after a hard week of training versus an easy week; note recall capability after a month of focused Japanese-learning versus a month of haphazard studies.

The reviews will likely show that progress is not necessarily linear, nor does it correlate directly with effort. Progress will stagnate on some days and thrive on others. Ability may regress slightly or jump immensely, both without apparent cause. Attempt to identify patterns in this via review.

Second, patience is lacking. Get-rich-quick schemes are prevalent nowadays, where “rich” is referencing progress in some domain and “quick” is referencing either time, effort, or both. Quadruple your initial investment in just 3 days! Learn how to code artificial intelligence like a pro in 30 days! Get the fit body you’ve always wanted in 30 days! These dishonest claims are plastered over advertisements, creating much more false hope than the honest version of “quadruple your initial investment in just four years!” or “get the fit body you’ve always wanted in two years with uncomfortable exercise, small portions, and going to bed early!” will. And yet these schemes persist because people continue to buy in to the lies.

Patience is a virtue. It takes practice to not get antsy when it feels like time is stretching. It takes practice not to push the pace when it feels like progress is stagnating. Like adjusting effort based on outcome, patience takes regular practice to develop and become comfortable with, ultimately paying off due to reserved approaches to the process.

Third, progress was made in the process, but was either not noticed or not appreciated due to lack of excitement. Note that all examples made some sort of progress: fitness person got stronger/faster, programmer got a programming job, language learner made it to A2, investor made money, overweight person lost weight. Some noticed it (overweight person), while others didn’t appreciate it (investor). And yet, despite lack of excitement, the process worked! This brings another key quote to the playing field: if it ain’t broke, don’t fix it

Before switching to something different, determine how much progress was made on the current program and if so, if progress is continuing to be made. If the former question is greater than 0 and the latter a “yes”, then there is absolutely no need to switch. Keep on keeping on. If the answer is greater than 0 and a “no”, then first the program and adherence to it should be assessed for any flaws or oversights. If issues are found, adjust accordingly. If there are no issues, discontinue the program and switch to something different.

Reality

Boredom implies monotony, which implies consistency, which produces results (along with other required factors, of course). Consistency is quite often the key to progress in most realms. Training? Check, strength and conditioning progress will be lost if there’s no training being performed. (I’m leaving programming out of this paragraph, as the principles are not necessarily learned in the same way a foreign language’s vocabulary is.) Language learning? Check, rules and vocabulary begin to be forgotten. Investing? Check, not putting money in regularly leads to low gains or purchasing investments at higher values. Weight loss? Check, a 500 calorie deficit one day doesn’t make up for a 600 calorie surplus the next—the caloric deficit needs to be maintained on a consistent basis, as an entire week’s worth of effort can be destroyed in one meal.

The misconception that in order for something to be effective it must be exciting—or at least not boring—is endemic in society. Companies are aggressive in claiming how their product is not boring, which makes obvious sense from a business standpoint, but not from a progress standpoint. Next time either boredom or excitement is felt, consider the role it’s playing and the benefit it’s providing. From there adjustments can be made.

Boredom educates. Boredom moderates. Boredom progresses.

See Also

• Play in Hard Mode

The trick is not to play zero sum games. This is what I have been saying the whole time. Go create value for others and don’t worry about the returns. If you create more value than you consume, you are welcome in any well operating community. Not infinite, not always needs more, just more than you consume. That’s enough, and avoid people or comparison traps that tell you otherwise. The world is not a Red Queen’s race.

• LEGO’s 0.002mm Specification and Its Implications for Manufacturing: Details on LEGO’s manufacturing tolerances and processes.

• Should You Marry Her?

• RFC 454545 — Human Em Dash Standard: “This document proposes the Human Em Dash (HED), a Unicode character visually indistinguishable from the traditional em dash (—) but encoded separately for the purpose of indicating probable human authorship. Recent proliferation of automated text generation systems has produced a measurable increase in the frequency and enthusiasm of em dash usage. This trend has created ambiguity for human writers who have historically relied upon the em dash as a stylistic device.”

• Kalshi’s Favorite Lie: Kalshi strategically lies to make themselves seem like a “better person” than the typical gambling house, when in fact (well, opinion), they’re not. Kalshi makes a fee off of each trade.

• Griffin Pinney: ULCA math PhD student who likes puzzles and is consistently one of the first to solve Jane Street’s monthly puzzles.

• You can’t get There from Here

• The Jevons Paradox and Insatiable Humans: A few (potentially naive and/or uninformed) jumbled thoughts/responses in no particular order:

  • My rebuttal to the “creates new and innovative jobs” argument is that we still needed human labor for those jobs, whereas now the models are getting close to, if not surpassing, human-level performance, eliminating the need for human labor because, well, AI can just do that too! People seem to miss that productivity tools are not the same as outright labor. Spreadsheets can 4x the accounting profession headcount, but it doesn’t matter if more robots are right there waiting to snatch them up!

  • I’m curious how the lower intelligence --> more use of it works out. Does this mean companies will just keep employment constant and scale productivity via LLMs? Or reduce employment to keep productivity the same via LLMs? I’d unconfidently predict the latter given there is only so big of a market for them to capture and they’d rather have better margins than risk the diminishing returns of getting more market share. (Again, definitely out of my element here!)

  • Will going to college become an even stronger status signal than it already is? John Doe going to college and studying X exhibits some pretty strong confidence: “I’m studying CS despite the annihilation of the X profession because I know I’m better than LLMs”. There’s also a revealed preference angle of people saying “go the trades route, it’s better!” and still encouraging their kids to get a four-year degree (I haven’t seen this yet, but would bet it exists).

  • I appreciate and/or like his examples of where AI will be a boon, what specific fields should prepare for, and him making it clear that we don’t know what will pop up.

  • Potentially-funny side note: I plugged in the first section of the Jevon’s Paradox paper into Pangram because it sounded a bit fishy, and sure enough, 100% AI generated. The rest of the paper was mixed around 50/50, so maybe he was just doing the first part to make a point. But on second thought, isn’t he almost making the point that his job could be obsoleted..???

• A day in the life of a quant researcher at Citadel Securities in Miami: Pretty self-explanatory. Super smart math guy who works a lot and exercises a bit in his free time.

• Dating Net Worth: “A calculator that estimates dating market value from age, attractiveness, height, income, and personality — with coefficients informed by published research on dating preferences. Half science, half art. For entertainment.”

• The gold standard of optimization: A look under the hood of RollerCoaster Tycoon: RCT stayed GOATed in the video game community. Incredibly fun, super optimized with no wasted lines of code. I’m curious of SOTA LLMs will be able to match the performance? Maybe that would be a good benchmark.

• Map of Shark Attacks in the US: A Des Moines zoo employee was bitten by a shark, hence the random dot in the middle of the midwest.

• U.S. vs. Backpage indictment: I’m surprised that Backpage was openly advertising prostitutes in this day and age, but then again, some people have different risk tolerances or intelligence levels than others. I like how specific the details in the indictment are.

• Kat Abughazaleh: American journalist, social media influencer, and politician.

• A Meta employee gets real about the horror of working there right now: I saw a funny tweet that said something like:

Someone once said there’s no such thing as a tall, high-earning incel. I retorted and said of course there is! Being a product manager at Meta is a thing!

• A Dot a Day Keeps the Clutter Away: Visually weighting the boxes he uses the most by putting dots on them after he uses them.

• Treasure hunter freed from jail after refusing to turn over shipwreck gold: “A US deep-sea treasure hunter who refused to disclose the location of a famed shipwreck’s gold coins has been released from prison after a decade, with 500 coins still unaccounted for.” While scummy, this may actually be the financially-correct thing to do? Assuming 1 oz per coin, that would be >$2MM, which is probably worth sitting in prison for for two years. That said, it’s difficult to put a price on the paranoia he’s likely to feel for the rest of his life!

• Chris Donahue (general): Commanding general of United States Army Europe and Africa and commander of Allied Land Command since 2024.

Both multiple-choice (MCE) and free-response exams (FRE) don’t necessarily show how confident the student is in their answers—they just show the final answer or the thought process, respectively. But that means the student can just guess and get it right, even if they don’t know for sure that the answer or thought process is correct!

Take MCEs. For a four-choice question, the student could improve their odds of guessing from 25% to 33-50% chance just by eliminating one or two answer choices. While this demonstrates some knowledge about what the answer isn’t, isn’t necessarily proof that they know what it is, even if they choose correctly.

Or FREs. A student may have some idea of what the question is asking, but could take a guess at which formula/process to apply and still get it correct. (This is much less likely than on a MCE, but still possible based on personal experience.)

This is unfair for the students who truly know the answer, but get scored the same as the guesser. What would happen if students had to note down how confidence they are that they have the correct answer?

Brier Scores for Exams

Introducing Brier scores (BS). The formula for a BS (as related to exams) is:

where:

• N is the number of questions on the exam

• t is the question number

• p_t is the student’s prediction that they got question t correct, ranging from 0 (”I definitely got it wrong”) to 1 (”I definitely got it right”)

• o_t is the result of their answer; values are either 0 (incorrect) or 1 (correct)

A perfect Brier score is 0, since 1-1=0 (“I think I got it right and I did get it right”) and and 0-0 = 0 ("I think I got it wrong and I did get it wrong"). One issue here is that if the Brier score was all that mattered, is that exams could be easily gamed by just intentionally getting the wrong answer and being confident about it, which is generally easier to do than getting the right answer.

The question becomes how are correct, confident answers rewarded?

Implementation

Predictions must be greater than 0.5 (50%) because the goal is to get the answer correct. If the student is less than 50% confident they can get the answer right, then they should change their answer to be more confident or be penalized accordingly. This will make each question have two sections of answers: section A, for the actual answer choices, and section B, for the student’s confidence level in the respective answer choice (50%, 60%, 70%, 80%, 90%, or 100%).

Scantron forms or tests can be modified to have two sets of answers per question to calculate the BS for that question. The total BS is then calculated to give the student their score with the lowest score being the best (alternatively, the BS can be inverted and the highest score wins).

Benefits

Exams that sort people by skill should rarely have perfect scores because the signal gets lost, at least to some extent. One person getting a perfect LSAT or perfect SAT signals that person is incredibly intelligent; 1% of the test-taking population getting perfect scores indicate they’re pretty smart and the test is too easy.

Adding confidence scores lets further sorting happen. If Ernest, Jameson, Douglas, Alistair, Tannatt, and Raymond all get question X correct, but their confidences are 50-60-70-80-90-100%, respectively, then their scores will be 0.25, 0.16, 0.09, 0.04, 0.01, and 0.00, indicating Raymond is top dog amongst his peers. This is done without needing to make the test more difficult, but simply by figuring out how correct each person really is.

Combine this with a more challenging test and the stars will really shine.

Literature

After writing all of the above, I found that literature already exists under the name “confidence-based marking”. There truly is very little that’s new left under the sun.

A.R. Gardner-Medwin discusses it in their Confidence-Based Marking - towards deeper learning and better exams, whose abstract states:

... A critical point is that they [students] benefit either by finding reasons to place greater reliance on an answer or by seeing reasons for reservation. This places a premium on careful thinking, and on checks and the tying together of different facets of knowledge, thereby encouraging deeper learning. In exams it generates higher quality data than conventional scores, with greater statistical reliability and validity as a measure of knowledge, and less contamination from chance factors associated with weak and uncertain knowledge. The puzzle remains, why this seemingly sensible strategy for objectively marked tests is so readily embraced by students and yet so little used by teachers.

I’d guess that high performers are fans because it helps distinguish them from the low performers and low performers aren’t fans because it they are more likely to get worse scores than in the binary correct-incorrect scoring system.

I won’t go through every study I can find, but Gardner-Medwin’s paper has some good references.

Semiconductor Fabs I: The Equipment

Semiconductor Fabs II: The Operation

Preface

I tried to include as many links as possible to allow the reader to go down rabbit holes as they see fit.

I try to include analogies in case the explanation is poor or the topic esoteric.

I don’t work in an advanced fab, but have some glimpses into them, so I rely on my ideas, conjecture, and the literature for a more accurate representation of what state-of-the-art fabs look like (although I’m sure there are features I could never dream of).

I first go over the data side of things, since a lot of that is fundamental to how the automation operates.

Why All the Data?

Fabs are incredibly hungry for data. Insatiably hungry. Data helps to connect patterns, solve problems, troubleshoot issues, and just plain understand what the heck is going on at the atomic level where the transistors and interconnects are made. I call it the fab data monster and Nano Banana 2 thinks it looks something like this:

More data is almost always a good thing because it allows you to fit your conclusion to the data. Just kidding. But not really. If an engineer has no freaking clue what’s causing a problem, blindly scouring the data may help uncover some anomaly that can clue them in on root cause. (But if you’re having to blindly scour the data, you probably don’t have enough data or your FDC systems aren’t developed enough.)

That said, false positives and false negatives are legitimate concerns that have to be considered. False positives may result in troubleshooting efforts in the wrong area and wasted time, money, and effort. False negatives may result in the ultimate root cause being overlooked and wasted time, money, and effort pursuing the wrong area.

What is the Data?

How much data could one fab possibly need? And what could they possibly be measuring that culminates in petabytes (1000 TB, or 1,000,000 GB) of data?

Tool Signals

Here’s a “short” list of potential equipment signals fabs can keep track of. For example, engineers may want to keep track of the temperature, pressure, and power of component1, while voltage, current, and resistance are relevant to component2.

• Temperature

• Pressure

• Angle

• Distance

• Position

• Voltage

• Current

• Resistance

• Power

• Quantity

• Status

Now those are just single characteristics that don’t add up to much storage space on their own. But what happens when we take measurements across multiple components across multiple tools across the entire fab?

Let’s assume the following for the NMP fab:

• 1000 tools

• 250 signals per tool

• 500 kB data per signal per hour (empirical data at a collection frequency of 1 data point per second; this includes the context needed to interpret the data (note that 1 Hz frequency is fairly slow and newer tools and fabs can accept data frqeuencies of up to and even exceeding 100 Hz, or 100 data points per second)

The math is then pretty easy:

total = 1000 tools × 250 signals/tool × 500 kB/signal/hour = 125 GB per hour ≈ 1000 TB per year of data

But wait! Those are just raw signals that the tool is reporting to the fab data monster. Statistics can be performed to get some extra info for each signal: mean, median, standard deviation, minimum, maximum, etc. Just those alone results in five times more data per signal than before (assuming you are constantly updating across the same time period, but generally a time period is defined and a single data point calculated for said period).

That’s a lot of data that is just passively created and recorded, some of which will be looked at, a lot of which won’t be. Regardless, it’s nice to have in case you need it.

Test Results

In-line Measurements

Wafers regularly get measured—either randomly, as determined by some algorithm, or intentionally due to the criticality of the process it just went through—throughout the line to ensure quality control at all processes. The measurements may be thickness after some film was deposited or etched, critical dimensions, number of particles on the wafer, or more specific measurements that are left up to the reader to determine.

These results are generally boring and not looked at because, well, they rarely fail, at least in more mature fabs where the technologies’ manufacturing processes have been optimized for years. Regardless, the tests are required for various reasons and results must sit in storage for some time.

Electrical Measurements

Some fabs (all fabs? Not exactly sure here.) will test their wafers in-house towards the end of the line to shorten the feedback loop if an issue is identified or get test results quickly so they can make changes permanent; it would be weeks if they chose to wait until the wafer got what is called its “final test” results, which measure the chip’s performance at its intended purpose.

In most, but preferably all, cases, the electrical test results are the fab’s gold standard for the quality of the wafer: if it’s passing and within the historical distribution of that parameter for similar devices, great! If it’s not, then something appears to have changed either within the line or with that specific lot. If the next lot of wafers that gets tested has similar out-of-the-distribution results... get investigating!

A few not-related-to-the-main-topic notes here for the more technically curious:

• Matching electrical data is pretty much how all changes ultimately get approved. Say I want to make change X to process Y. I will propose the change, get approved to run some experimental wafers, run them, review the test results, and if they’re good, request approval to fully implement change X for every wafer that runs through process Y. Changes may included small tweaks to processes or entire process flow changes.

• Some parameters are strongly related to certain processes within the line, which helps to narrow down what went wrong instead of having to spot check every tool. For example, if the threshold voltage is all out of whack, that points to a problem in the gate oxide area, which is a particular machine or two in the fab. The engineers reviewing the test results can then notify the gate oxide engineers and have them look into it by digging through the tool signals or maintenance history.

Histories

Histories allow engineers to look back and see what happened on a certain date or to a certain lot. Tool signals are a form of history (what was component X doing at Y time?), but other histories are also important.

Maintenance

Fabs want a way to easily document events in a machine’s life, whether automated or input by a person. These leave digital bread crumbs of-sorts that can be checked to see what happened, why it happened, etc.

Here are some examples of helpful automated comments:

SPC chart X failed with value of Y1 and Y2. Limit is X. Last Z points have been in control. [This is an event that I can anchor to and look around at: what maintenance, if any, was done before the chart fail? What was the response to the chart fail? Was something repaired or replaced?]

Machine alarm X with description “Y” occurred. It has happened Z times over the past 30 days. Review recommended troubleshooting and solutions here: [link]. [This is an event that I can anchor to just like the last one, plus I get to see how bad the issue has gotten.]

And here are some examples of helpful typed-by-a-person comments:

Removed parts A1/2/3, B, and C to better diagnose lift issue. Found that part A1 is sticking at its end range of motion, which corresponds to the side of the lift that’s having the problem. Parts A2, A3, B, and C are all good and have no obvious issues when testing. Regardless, replaced all three part As since everything is open. Original B and C will go back in. The new A1/2/3 serial numbers are 123, 456, and 789, respectively. Next steps are [list of next steps]. [This tells me exactly what the issue was and what was replaced, so future me can just reference back.]

Machine alarmed for X. Found that part D was completely powered off. Verified that all relevant circuit breakers were on and there is no power discontinuity up to the part, so appears that D has failed. Wafer 13 was 10 seconds into step E of recipe F when the failure happened. All wafers placed back into the FOUP. Replaced part D and verified it has power and functions normally. Machine was vented to atmosphere and opened for replacement. No other issues noted on machine. [Same as the above.]

This is all data! It may not be numbers, but it paints a clear picture of the what happened on a machine during a certain time period.

Lot Processing

Lots will get data and information automatically “attached” to them throughout their life. Examples include what machine the lot processed on for a certain process, what time it started and stopped, were there any abnormalities while processing, what associated data (as in in-line measurement results) is there. The list goes on. Like the in-line measurements, this data is really only looked at when there’s an issue.

Automation (of Everything)

Automation is a beautiful thing. It helped get us cheap and abundant everything, including semiconductors.

Automation here is referred to as, well, anything automated, and no, I’m not being a smartass. A majority of fab automation has to do with the actual wafer processing and making it less error-prone and more efficient, but there are plenty of other uses.

The Wafer Life Cycle Without Humans

The life cycle of a wafer—from its start as just a bare silicon wafer to its end when it’s full of chips—can be looked at to get a better understand of what fab automation is and isn’t. I’ve provided significant detail both for nerd-sniping and to get a good picture of how many decisions are actually automated in the fab on a second-by-second basis. While reading, think about how time-consuming and error-filled a fab would be if humans had to make all the decisions and perform all the calculations.

Here’s the basic flow for how a lot runs through the line, along with some non-ideal situations arising throughout to show what automation can and can’t do:

1. Go! (But no collecting $200—the fab needs that money!) The 25-wafer lot is assigned a lot number—call it 123—and device—call it ABC. The manufacturing execution system (MES) knows every single process that 123 has to go to and what it needs to do at each.

2. Lot 123 reaches its first process: thermal oxidation in a vertical furnace. 123 is ready to run and there’s an available furnace, so the MES dispatches 123 to the furnace, which initiates automated pre-flight checks that makes sure everything is copacetic:

   1. Is the machine actually available to accept this job request? The MES will query the equipment management system and tool itself to verify it can accept the proposed job.

   2. Are all of the necessary SPC charts in control? Common SPC charts for qual wafers (wafers that run a test to ensure the machine is performing properly) are:

      • Thickness: The qual is targeted to and representative of the process that the job will run, e.g., if the job’s process grows 10 nm of oxide, then the qual will also grow 10 nm or can be easily and reliably extrapolated from a different thickness.

      • Particles: The qual ensures no particles are being generated by the machine.

      • Contamination: The qual ensures there’s no metallic contamination coming from the tool, often in the form of particles.

   3. Is 123 allowed to run on this machine? Some fabs will automatically qualify like machines if everything matches, while others require individual machines to be qualified for each process. The former is much faster, but risker; the latter slower, but safer. See Intel’s (in)famous Copy EXACTLY! method.

   4. Checks current tool settings against a pre-defined list. Do all of the settings match? Incorrect settings can cause misprocessing or failure to interdict on a process that’s going poorly.

      • Analogy: Assume a vehicle has customizable air-fuel ratio alarm settings, where 14:1 air:fuel is the ideal, 13:1 is the lower limit, and 15:1 is the upper limit. If it goes above or below the upper or lower limit, the car alarms and notifies the driver. Before driving, it would be wise to ensure these limits haven’t been temporarily adjusted by a pesky, speed-happy teenager who forgot to change it back in their excitement of going fast. If either limit were changed to something much larger, the next drive could be in the harmful-for-engine range without the driver knowing.

3. 123 makes it to the furnace and runs the correct recipe that automation told it to run. There is no need to make sure that recipe A1B2C3D4E5 is selected vs. recipe AIB2C3D4E5 (see what the difference is?) because automation commands the correct one to be chosen.

4. 123 gets put on hold after the furnace because the post-thickness measurement was too high. Savvy automation systems would automatically remeasure the lot to ensure the measurement was legit (2 hours of the lot not moving, no manpower required), while stone-age automation would require a person to manually remeasure themselves (4 hours of the lot not moving, plus valuable manpower required). The measurement turns out to be bogus and the lot continues on.

5. 123 continues on and reaches a certain process (call it process 27) in the flow that requires feedforward data. The general feedforward flow goes like this:

   1. Data from process 26 for lot 123 is “sent” to process 27. Process 27 knows how to adjust itself based on the process 26 data. Adjustments can include a variety of parameters depending on the process, such as time, power, pressure, etc.

   2. A handful of wafers from 123 (call them 123-1, 123-2, and 123-3) are sent ahead of the rest of its brothers and sisters in case a sacrifice to the fab gods is in order, i.e., they will be the test wafers for the rest of the lot to ensure everything is fine. 123-1/2/3 all complete process 27 with the process 26 data and measure to check the results, resulting in process 27 data for wafers 123-1/2/3.

   3. Wafers 123-4 to 123-25 are then sent on to process 27 with combined data from both process 26 (for the entirety of lot 123) and process 27 (wafers 123-1/2/3 only).

   Pretty simple, right? Right?! The granularity of customizing this flow is practically infinite (depending on the sophistication of the fab automation, of course). Custom adjustment values can be set based on machine (from both process 26 and process 27), device, measurement tool, etc. Full moon? Adjust accordingly! Good or bad vibes in the air? Adjust accordingly!

   

6. 123 makes it to process 53, where an engineer has defined an experiment that will test a few different conditions. Individual wafers will be separated into distinct lots, process, then recombine into a big lot. The general process goes like this:

   1. Wafers are separated into “sublots”, e.g., wafers 123-1/2/3 become lot 123-A, wafers 123-4/5/6 become lot 123-B, etc.

   2. Lot 123-A goes through process 53 with any experimental conditions that the engineer defined. Repeat for the remaining lots. There is almost always a baseline lot to compare to.

   123-A/B/.../Z continue on until they reach the recombine point, which could be the next process or multiple processes down the line. Now imagine if the split occurred at a feedforward process and how complicated that would be!

7. Wafer 123-13 breaks randomly and other wafers are exposed to the particles that come with the break. Good automation would recognize that there was a wafer break and automatically inspect the lot for particles on other wafers, then make a decision to clean the lot if particles were found. Bad automation would put the lot on hold and wait for an engineer to review the incident, request inspection, review inspection, then make a decision. (In case I’m not being clear enough, I think pretty much anything in fab automation land is possible with sufficiently talented software engineers. It’s just a matter of assigning the necessary resources.)

This flow isn’t an exhaustive list of all of the potential pain points, but illustrates a good chunk of them. Now imagine having to do all of the following by hand:

• Assign a lot number and device. Was everything entered correctly?

• Verify that the machine can accept the job. Are all the settings—the tens of thousands that exist—correct? Are all of the SPC charts in control? Is the machine qualified? Was the correct recipe selected?

• Review the thickness measurement fail. Is it real or fake? Does it need to be remeasured? Do you need to tag the lot for later inspection or testing?

• Make feedforward process adjustments. Was the math done correctly? Was the correct parameter adjusted? Was the full moon accounted for?

• Separate the lots correctly. Were the correct wafers chosen and settings applied to each?

Now rinse and repeat some of these multiple times for the hundreds, if not thousands, of lots running the fab at any given point. It’s unsustainable and overwhelming, hence the need for automated systems and processes to do most of the work. QED.

Other Automation Functionalities

Fault Detection and Classification

Averroes does an excellent job of explaining FDC systems and there’s no need to reinvent (or re-explain) the wheel.

Part Management

I’ve never seen or heard of anything like this, but if I could vibe-code up a part management system, it would look something like this:

• Detect what maintenance is coming due and what parts are needed for said maintenance. If parts are on-hand, excellent; if not, order them while taking into account current stock, lead times, historical delays.

• Learn how often a part fails and extrapolate out to ensure there is always a spare available. For example, if machines 1-12’s partX fails at a rate of once per month, then 12 fail per year on average, so there should be at least one available at all times, but preferably two in case failures occur around the same time. Remember: two is one and one is none—equipment downtime is often more costly than paying for and keeping an extra spare on hand!

• Original equipment manufacturers will often obsolete parts because of obsolete components, upgrades they’ve made, or issues they’ve found. OEMs could alert the company, which would trigger the following actions:

  • Flag what tools are at risk, if there is a risk, and notify the engineers.

  • Update the part number in the company system to reflect the new OEM part number. Companies have their part number, which is separate from the OEM part number. For example, NMP LLC’s part number for partX is 1111-222, while an OEM PN is 1234-12345.

  • Change all existing references to the old part number to the new part number (like a pointer’s memory getting updated)

Praise Be to the Automation Engineers

Automation here is similar to automation elsewhere: it makes people’s lives and the manufacturing process safer and more efficient. And it’s freaking awesome. I can sit at my desk and do a good chunk of my job without moving because some automation guru coded up a wonderful script that helps me out.

And while the fab requires everyone to operate smoothly, the automation engineers are the heroes working in the background that nobody really thinks about. Here’s to them.

See Also

• Equipment Anomaly Detection for Semiconductor Manufacturing by Exploiting Unsupervised Learning from Sensory Data

Three types of lines exist in the policy and behavior sense:

• Soft: These are okay to cross, but not preferable. There may or may not be a tangible action taken afterwards, but the person whose line was crossed should take note.

• Firm: These are somewhere between soft and hard lines and should result in some tangible action being taken that is less drastic than the hard line.

• Hard: These are not okay to cross and (should) result in some tangible action being taken that is more drastic than the firm line.

Most lines are rarely set and rarely thought about in detail. Most line setters use the good ol’ “I know it when I see it” test, waiting for something to happen before they decide what to do. This is a poor practice because of the pernicious force known as normalization of deviance.

When lines aren’t set before they’re crossed, it forces a decision to be made at the time of crossing (if it can even be recognized that something was crossed!), during which many things can happen:

• The line setter convinces themselves that the line wasn’t really crossed and everything is fine. This will land the setter in not-so-nice territory if this occurs enough times because the line effectively moves back each time.

  • Ex: Ruben, Lou’s boyfriend, playfully pinches her, then playfully punches her, then seriously pinches her, then seriously punches her, and so on. Each time she convinces herself that her domestic abuse line wasn’t crossed, ultimately leading to her getting full-on abused.

• The line setter acknowledges the line was crossed, but because taking action is uncomfortable at the time of crossing, vows to wait until it happens a second time because the first time may have been a one-off. This increases the likelihood they give a third chance to the offense when/if they apply the same thought process to the second.

  • Ex: Diane blatantly lies and talks about Joe both behind his back and to his face. Joe explains away the behavior as Diane having a stressful time and continues being “friends”. Diane continues the behavior while Joe accepts and normalizes it as Diane’s personality. Joe’s self-esteem decreases as he continues to spend time with Diane.

• The line setter acknowledges the line was crossed, but convinces themselves that the line really should’ve been just a teeny bit further when they originally set it.

  • Ex: Harlan’s original salary threshold for taking the Giving What We Can pledge was $100k/year, but now that he’s reached it, it feels a bit low. After all, he deserves to treat himself a bit more for all the hard work he put himself through to get to the coveted six-figure salary. Plus, he may have a baby in the next few years! And everyone knows how expensive babies are! Harlan resets his salary goal at $120k, which will be plenty when the time comes.

By setting a line and its corresponding action early, the action becomes the default until proven otherwise. This is similar to trigger-action plans.

How to Set a Line

Here’s the general process of setting a line:

1. Figure out the general line. Whether it’s domestic abuse, talking smack, donating money, or rights being restricted or outright revoked, it must be defined.

2. Define the criteria for both the soft, firm, and/or hard versions, but especially the hard. The soft line being crossed serves as a forewarning to the hard line being crossed, giving ample preparation time for if the hard line is eventually crossed. The criteria must be well-defined with little room for interpretation.

3. Decide how many times each can be crossed before the action is taken. It’s fine to give someone a stern reminder that they crossed the line in case they forgot, weren’t aware of the line, weren’t aware that it was soft/firm/hard, etc. It’s not fine for it to happen more than the set number allows, especially if previous actions were taken.

4. Define the actions for each line. This can also be done in conjunction with deciding the number of times it can be violated, since more drastic actions should have fewer subsequent violations and thus a lower number of allowable violations.

5. Define what circumstances would have to be present for the action not to be taken. What evidence would it take to show that the hard line was crossed, but the action shouldn’t be taken? (This is a bit contradictory to how hard line is defined above, but the hard line action is simply the default, not a blind requirement that must be executed. Setters should double-check they didn’t miss something before taking the default action.)

6. Communicate the lines and actions to people who either may be at risk of crossing them or will help with maintaining accountability of executing said actions.

7. Prepare for taking the action when/if the time comes. Preparation may be mental, physical, or environmental.

Ensuring Accountability

The line means nothing—and in reality, is likely a large cost—if the action is never performed when it should be. Assuming the fourth and fifth steps are done honestly and comprehensively, it should be clear what decision needs to be made when the line is crossed.

Thus, an accountability method must be put in place to enforce the action being taken.

A few ideas that all rely on the honor system to some extent:

• Require a cost greater than that of said action be paid. If the action costs $10, make the cost of not doing the action $20.

• Publicly or privately announce the lines and ask a trusted person to be your accountability partner. They know your lines and make sure you follow through on the actions, else a cost will be incurred (see previous idea).

• Automate the action. For example, write a script that looks to see if the friend who borrowed money ever paid it back by a certain date. If current_date > deadline_date & money_repaid = false, then send an automated email unfriending them.

Line Examples

Here are some hard line ideas and associated actions (in no particular order; assume the case is straightforward with no nuance):

Government Overreach

• Soft line: A government violates a law with the expectation that the lengthy legal process will allow them to reap the benefits before a ruling is made

• Soft line action(s): Protest

• Hard line: A government blatantly violates a constitutional amendment or refuses to comply with a court order

• Hard line action(s): Apply for a visa or similar in another country

Relationships

• Hard line: Romantic partner commits adultery

• Hard line action(s): Break up/divorce

• Hard line: A friend doesn’t repay an $X loan

• Hard line action(s): Stop being friends with said person

• Soft line: A friend makes disparaging comments about you, but claims it’s “just a joke”

• Soft line action(s): Tell them to not do that again, but continue being friends with them

Workplace

• Firm line: Boss makes an immoral or illegal request, but doesn’t retaliate when it’s refused

• Firm line action(s): Submit a whistleblower complaint; submit an ethics violation with the company; begin a new job search; resign

• Hard line: Annual raise is 0%

• Hard line action(s): Begin a new job search

• Hard line: $Xk/year annual liquid compensation

• Hard line action(s): Donate X% to charity

• Soft/firm/hard line: Achieve $X net worth

• Action(s):

  • Retire

  • Say fuck you to your terrible boss

  • Start looking for a new job that pays less, but is better otherwise (stress, hours, culture)

Health

• Soft line: Weight above X

• Soft line action(s): Begin weight loss actions (eating less, exercising more)

• Hard line: Heart attack or stroke

• Action(s):

  • Change health-related habits (diet, exercise, stress)

  • Start a medication

  • Retire

  • Start looking for a new job that is less stressful