Seek Fair Expectations of Others’ Models

Epistemic Status: Especially about the future.

Response To (Eliezer Yudkowsky): There’s No Fire Alarm for Artificial General Intelligence

It’s long, but read the whole thing.  Eliezer makes classic Eliezer points in classic Eliezer style. Even if you mostly know this already, there’s new points and it’s worth a refresher. I fully endorse his central point, and most of his supporting arguments.

What Eliezer has rarely been, is fair. That’s part of what makes The Sequences work. I want to dive in where he says he’s going to be blunt – as if he’s ever not been – so you know it’s gonna be good:

Okay, let’s be blunt here. I don’t think most of the discourse about AGI being far away (or that it’s near) is being generated by models of future progress in machine learning. I don’t think we’re looking at wrong models; I think we’re looking at no models.

I was once at a conference where there was a panel full of famous AI luminaries, and most of the luminaries were nodding and agreeing with each other that of course AGI was very far off, except for two famous AI luminaries who stayed quiet and let others take the microphone.

I got up in Q&A and said, “Okay, you’ve all told us that progress won’t be all that fast. But let’s be more concrete and specific. I’d like to know what’s the least impressive accomplishment that you are very confident cannot be done in the next two years.”

There was a silence.

Eventually, two people on the panel ventured replies, spoken in a rather more tentative tone than they’d been using to pronounce that AGI was decades out. They named “A robot puts away the dishes from a dishwasher without breaking them”, and Winograd schemas. Specifically, “I feel quite confident that the Winograd schemas—where we recently had a result that was in the 50, 60% range—in the next two years, we will not get 80, 90% on that regardless of the techniques people use.”

A few months after that panel, there was unexpectedly a big breakthrough on Winograd schemas. The breakthrough didn’t crack 80%, so three cheers for wide credibility intervals with error margin, but I expect the predictor might be feeling slightly more nervous now with one year left to go. (I don’t think it was the breakthrough I remember reading about, but Rob turned up this paper as an example of one that could have been submitted at most 44 days after the above conference and gets up to 70%.)

But that’s not the point. The point is the silence that fell after my question, and that eventually I only got two replies, spoken in tentative tones. When I asked for concrete feats that were impossible in the next two years, I think that that’s when the luminaries on that panel switched to trying to build a mental model of future progress in machine learning, asking themselves what they could or couldn’t predict, what they knew or didn’t know. And to their credit, most of them did know their profession well enough to realize that forecasting future boundaries around a rapidly moving field is actually really hard, that nobody knows what will appear on arXiv next month, and that they needed to put wide credibility intervals with very generous upper bounds on how much progress might take place twenty-four months’ worth of arXiv papers later.

(Also, Demis Hassabis was present, so they all knew that if they named something insufficiently impossible, Demis would have DeepMind go and do it.)

The question I asked was in a completely different genre from the panel discussion, requiring a mental context switch: the assembled luminaries actually had to try to consult their rough, scarce-formed intuitive models of progress in machine learning and figure out what future experiences, if any, their model of the field definitely prohibited within a two-year time horizon. Instead of, well, emitting socially desirable verbal behavior meant to kill that darned hype about AGI and get some predictable applause from the audience.

I’ll be blunt: I don’t think the confident long-termism has been thought out at all. If your model has the extraordinary power to say what will be impossible in ten years after another one hundred and twenty months of arXiv papers, then you ought to be able to say much weaker things that are impossible in two years, and you should have those predictions queued up and ready to go rather than falling into nervous silence after being asked.

In reality, the two-year problem is hard and the ten-year problem is laughably hard. The future is hard to predict in general, our predictive grasp on a rapidly changing and advancing field of science and engineering is very weak indeed, and it doesn’t permit narrow credible intervals on what can’t be done.

I agree that most discourse around AGI is not based around models of machine learning. I agree the AI luminaries seem to not have given good reasons for their belief in AGI being far away.

I also think Eliezer’s take on their response is entirely unfair. Eliezer asks an excellent question, but the response is quite reasonable.


It is entirely unfair to expect a queued up answer.

Suppose I have a perfectly detailed mental model for future AI developments. If you ask, “What’s the chance ML can put away the dishes within two years?” I’ll need to do math, but: 3.74%.

Eliezer asks me his question.

Have I recently worked through that question? There are tons of questions. Questions about least impressive things in any reference class are rare. Let alone this particular class, confidence level and length of time.

So, no. Not queued up. The only reason to have this answer queued up is if someone is going to ask. 

I did not anticipate that. I certainly did not in the context of a listening Dennis Hassabis. This is quite the isolated demand for rigorI’ll need to think.


Assume a mental model of AI development.

I am asked for the least impressive thing. To answer well, I must maximize.

What must be considered?

I need to decide what Eliezer meant by very confident, and what other people will think it means, and what they think Eliezer meant. Three different values. Very confident as actually used varies wildly. Sometimes it means 90% or less. Sometimes it means 99% or more. Eliezer later claims I should know what my model definitely prohibits but asked about very confident. There is danger of misinterpretation.

I need to decide what impressiveness means in context. Impressiveness in terms of currently perceived difficulty? In terms of the public or other researchers going ‘oh, cool’? Impressive for a child? Some mix? Presumably Eliezer means perceived difficulty but there is danger of willful misinterpretation.

I need to query my model slash brainstorm for unimpressive things I am very confident cannot be done in two years. I adjust for the Hassabis effect that tasks I name will be accomplished faster.

I find the least impressive thing.

Finally I choose whether to answer.

This process isn’t fast even with a full model of future AI progress.


I have my answer: “A robot puts away the dishes from a dishwasher without breaking them.”

Should I say it?

My upside is limited.

It won’t be the least impressive thing not done within two years. Plenty of less impressive things might be done within two years. Some will and some won’t. My answer will seem lousy. The Hassabis effect compounds this, since some things that did not happen in two years might have if I’d named them. 

Did Eliezer’s essay accelerate work done on unloading a dishwasher? On the Winograd schemas?

If I say something that doesn’t happen but comes close, such as getting 80% on the Winograd schemas if we get to 78%, I look wrong and lucky. If it doesn’t come close, I look foolish.

Also, humans are terrible at calibration.

A true 98% confident answer looks hopelessly conservative to most people, and my off-the-cuff 98% confident answer likely isn’t 98% reliable.

Whatever I name might happen. How embarrassing! People will laugh, distrust and panic. My reputation suffers.

The answer Eliezer gets might be important. If I don’t want laughter, distrust or panic, it might be bad if even one answer given happens within two years.

In exchange, Eliezer sees a greater willingness to answer, and I transfer intuition. Does that seem worth it?


Eliezer asked his question. What happened?

The room fell silent. Multiple luminaries stopped to thinkThat seems excellent. Positive reinforcement!

Two gave tentative answers. Those answers seemed honest, reasonable and interesting. The question was hard. They were on the spot. Tentativeness was the opposite of a missing mood. It properly expresses low confidence. Positive reinforcement!

Others chose not to answer. Under the circumstances, I sympathize.

These actions do not seem like strong evidence of a lack of models, or of bad faith. This seems like what you hope to see.


I endorse Eliezer’s central points. There will be no fire alarm. We won’t have a clear sign AGI is coming soon until AGI arrives. We need to act now. It’s an emergency now. Public discussion is mostly not based on models of AI progress or concrete short term predictions.

Most discussions of the future are not built around concrete models of the future. It is unsurprising that AI discussions follow this pattern.

One can still challenge that one needs short-term predictions about AI progress to make long-term predictions. It is not obvious long-term prediction is harder, or that it depends upon short-term predictions. AGI might come purely from incremental machine learning progress. It might require major insights. It might not come from machine learning.

There are many ways to then conclude that AGI is far away where far away means decades out. Not that decades out is all that far away. Eliezer conflating the two should freak you out. AGI reliably forty years away would be quite the fire alarm.

You could think there isn’t much machine learning progress, or that progress is nearing its limits. You could think that progress will slow dramatically, perhaps because problems will get exponentially harder.

You might think problems will get exponentially harder and resources spent will get exponentially larger too, so estimates of future progress move mostly insofar as they move the expected growth rate of future invested resources.

You could think incentive gradients from building more profitable or higher scoring AIs won’t lead to AGIs, even if other machine learning paths might work. Dario Amodei says OpenAI is “following the gradient.”

You could believe our civilization incapable of effort that does not follow incentive gradients.

You might think that our civilization will collapse or cease to do such research before it gets to AGI.

You could think building an AGI would require doing a thing, and our civilization is no longer capable of doing things.

You could think that there is a lot of machine learning progress to be made between here and AGI, such that even upper bounds on current progress leave decades to go.

You could think that even a lot of the right machine learning progress won’t lead to AGI at all. Perhaps it is an entirely different type of thought. Perhaps it does not qualify as thought at all. We find more and more practical tasks that AIs can do with machine learning, but one can think both ‘there are a lot of tasks machine learning will learn to do’ and ‘machine learning  in anything like its current form cannot, even fully developed, do all tasks needed for AGI.’

And so on.

Most of those don’t predict much about the next two years, other than a non-binding upper bound. With these models, when machine learning does a new thing, that teaches us more about that problem’s difficulty than about how fast machine learning is advancing.

Under these models, Go and Heads Up No-Limit Hold ‘Em Poker are easier problems than we expected. We should update in favor of well-defined adversarial problems with compact state expressions but large branch trees being easier to solve. That doesn’t mean we shouldn’t update our progress estimates at all, but perhaps we shouldn’t update much.

This goes with everything AI learns to do ceasing to be AI.

Thus, one can reasonably have a model where impressiveness of short-term advances does not much move our AGI timelines.

I saw an excellent double crux on AI timelines, good enough to update me dramatically on the value of double crux and greatly enrich my model of AI timelines. Two smart, highly invested people had given the problem a lot of thought, and were doing their best to build models and assign probabilities and seek truth. Many questions came up. Short-term concrete predictions did not come up. At all.


That does not mean any of that is what is happening.

I think mostly what Eliezer thinks is happening, is happening. People’s incentive gradients on short term questions say not to answer. People’s incentive gradients on long term questions say to have AGI be decades out. That’s mostly what they answer. Models might exist, but why let them change your answer? If you answer AGI is near and it doesn’t happen you look foolish. If you answer AGI is near and it happens, who cares what you said? 

When asked a question, good thinkers generate as much model as they need. Less good thinkers, or the otherwise motivated, instead model of what it is in their interest to say.

Most people who say productive AI safety work cannot currently be done have not spent two hours thinking about what could currently be done. Again, that’s true of all problems. Most people never spend two hours thinking about what could be done about anything. Ever. See Eliezer entire essential sequence (sequence Y).

That is how someone got so frustrated with getting people to actually think about AI safety that he decided it would be easier to get them to actually think in general. 

To do that, it’s important to be totally unfair to not thinking. Following incentive gradients and social queues and going around with inconsistent models and not trying things for even five minutes before declaring them impossible won’t cut it and that is totally not OK. 

He emphasizes nature not grading on a curve, and fails everyone. Hard. The Way isn’t just A Thing, it’s a necessary thing.

Then we realize that no, it’s way worse than that. People are not only not following The Way. No one does the thing they are supposedly doing. The world is mad on a different level than inaccurate models without proper Bayesian updating and not stopping to think or try for five minutes once in their life let alone two hours. There are no models anywhere.

Fairness can’t always be a thing. Trying to make it a thing where it isn’t a thing tends to go quite badly.

Sometimes, though, you still need fairness. Without it groups can’t get along. Without it you can’t cooperate. Without it we treat thinking about a new and interesting question as evidence of a lack of thinking.

Holding everyone to heroic responsibility wins you few friends, influences few people and drives you insane.


Where does that leave us? Besides the original takeaway that There Is No Fire Alarm For Artificial General Intelligence and we need to work on the problem now? And your periodic reminder that people are crazy and the world is mad?

Microfoundations are great, but some useful models don’t have them. It would be great if everyone had probabilistic time distributions for every possible event, but this is totally not reasonable, and totally not required to have a valid opinion. Some approaches answer some questions but not others.

We must hold onto our high standards for ourselves and those who opt into them. For others, we must think about circumstance and incentive, and stop at ‘tough, but fair.’

Predictions are valuable. They are hard to do well and socially expensive to do honestly. A culture of stating your probabilities upon request is good. Betting on your beliefs is better. Part of that is understanding not everyone has thought through everything. And understanding adverse selection and bad social odds. And realizing sometimes best guesses would get taken too seriously, or commit people to things. Sometimes people need to speak tentatively. Or say “I don’t know.” Or say nothing.

Allies won’t always ponder what you’re pondering. They aren’t perfectly rigorous thinkers. They don’t think hard for two hours about your problem. They don’t often make extraordinary efforts.

Most of what they want will involve social reality and incentive gradients and muddled thinking. They’re doing it for the wrong reasons. They will often be unreliable and untrustworthy. They’re defecting constantly.

You go to war with the army you have.

We can’t afford to hold everyone to impossible standards. Even holding ourselves to impossible standards requires psychologically safe ways to do that.

When someone genuinely thinks, and offers real answers, cheer that. Especially answers against interest. They do the best they can. From another perspective they could obviously do so much more, but one thing at a time.

Giving them the right social incentive gradient, even in a small way, matters a lot.

Someone is doing their best to break through the incentive gradients of social reality.

We can work with that.




Posted in Uncategorized | Leave a comment

Welcome to the World

Epistemic Status: Good news, everyone!

On October 5, 2017, my second son was born. Everyone say hello to the wonderful Gideon Mowshowitz:

As you can imagine, it’s been quite a week. We are all doing well, and slowly recovering aside from lack of sleep.

I encourage you to send your congratulations, but as an experiment I’m adding a twist: You need to say congratulations differently than all previous congratulations. No duplicates!

As you can also imagine, my available time will be limited in the coming weeks and months, so posting likely will be lighter. This is also a natural stopping point. I’ve said the community-related things I most needed to say. I have more thoughts, and will doubtless share some, but they seem less pressing. Sequence complete.

If you wish to join us for Friday night Sabbath dinner, please email me to let us know so we can prepare. If you can found this and can figure out my email, you’re probably invited. We’d love to meet some new people, and to see some old friends. We start at 6:30. Also let me know if you’d like to swing by on a Saturday to chat or watch some football.

My offer of strategic consultation for all causes I don’t actively oppose remains open. Only two people have taken me up on it, so I am reminding everyone that it exists. Response times might be not so fast right about now, and if I get too many requests I’ll have to limit this, but for now the line is open.

In conclusion, celebrate. Good times. Everyone around the world, come on!



Posted in Uncategorized | 9 Comments

Beginners’ Meditation

I was about to meditate. There was a beginners’ meditation class three blocks away at the 14th Street Y. Seemed wrong to not check it out. Low risk, potential high reward. Non-zero story value. Might even learn something.

It was held in Room 403. Not in the gym. In the preschool. Not great.

I looked inside. I saw a circle of about twenty folding chairs. Sitting in most of the chairs were very old women, chatting. Whatever filter was operating was not subtle.

It said, you are not the target.

Welcome to Meditators Anonymous. My name is Zvi. Hi, Zvi.

I asked if I was in the right room, hoping I wasn’t. I was. Despite lowered expectations, I sat down.

They passed around a petition to keep the class going after its one month engagement. This was class two. Impressive customer loyalty. Was a there there after all?

Several minutes late, the circle was complete. Me. One instructor. One Asian man. One Amy Poehler lookalike. Twenty-two very senior female citizens.

We were told meditation was us getting to know ourselves. Accepting our own friend request on Facebook. I tried not to make too much of a face. I tried not to make too little of a face. Would have been a lie. People deserve honest feedback.

We went around, said our names and mentioned something that warmed our hearts this past week. To get to know each other. I noticed I was confused. What did that have to do with meditation?

Some mentioned small niceties. Most mentioned babies. They especially loved fathers with babies, doing heroic things like pushing strollers or carrying the baby.

Here’s to you, baby carrying father. Here’s to you.

My son Gideon had been born the previous Thursday. So I won.

I like victory. Big fan. Passed the test just like all the rest. But never really understood the reasons why I took it in the first place.

We were told to sit comfortably in our folding chairs. Not a beginner task.

We were told to sit up fully straight, in a relaxed position. Some people can do that.

We were told to keep our eyes open, aiming six feet ahead on the floor, to take our practice into the world. We were told to relax our shoulders, and put our hands on our thighs. Can do.

Focus on the breath, either at the belly, chest or nose. If we have thoughts, label that thinking and come back to the breath. She mentioned things we might be distracted by. They were distracting. I got briefly hungry. I came back to the breath.

That was it. The whole instruction. The meditation itself lasted maybe ten minutes. There were people fidgeting. The instructor eyed her watch. I had thoughts. Many were meta. Is that common? Perhaps common but not discussed. I came back to the breath. We finished.

The instructor went over  ‘the practice’ again, asking us to name the steps. Like this was a classroom. It was. So, fair? Still infantilizing. Disrespectful of my time. This wasn’t complicated. Sit up straight, hands on thighs, look down six feet, focus on breath, bring focus back.

That was “the practice.” It had an exotic-sounding name.

Teacher said we keep our eyes open and sit in chairs so we don’t fall asleep. Must. Be. Nice. I envy those who fall asleep that easily. I am not enlightened.

I also wonder if that means that if you don’t fall asleep, you should lie down. Tempting!

Teacher took questions. Students pointed out focusing is hard. She agreed. They asked if it stops being hard. She said it doesn’t.

Teacher concluded with a call to ‘daily practice.’ If you had two minutes, that was all right. Two minutes will change your life!

How, exactly?

I know why I’m meditating. I’ve talked, theorized, modeled, gotten recommendations, read the review post, read the comments, read the other post, done better theorizing and modeling, and read the book up to where I couldn’t tell if it was talking nonsense anymore. Stage Four. Reading further would only distract. Beginner mind. So far so legit. Exceeds expectations. Tentatively recommended, could tell you why. Some day.

This class made no such attempt. No goals. No explanations. You make friends with yourself, on two minutes a day? No there there. Epic fail.

So why were these old ladies meditating? What kept them coming back, even signing a petition? Wasn’t point you only need to come once? Was this an excuse to share warm father and baby anecdotes? Why else would they want the class to continue? Why else were so many of them there so early?

Nobody Does the Thing They Are Supposedly Doing.

I hope that’s it. I wish them well. They seemed nice.



Posted in Uncategorized | 12 Comments

The Fish-Head Monk

Epistemic Status: Parable

By way of the comments on SlateStarCodex, from Essential Sufism:

‘There was a poor fisherman who was a Sufi teacher. He went fishing every day, and each day he would distribute his catch to the poor of his village, except for a fish head or two that he used to make soup for himself His students dearly loved and admired their ‘fish-head sheikh.”

One of the students was a merchant. Before traveling to Cordoba, the teacher asked him to convey his greetings to his own teacher, the great sage Ibn ‘Arabi, and to ask the sage for some advice to help him in his own spiritual work, which he felt was going very slowly.

When the merchant arrived at Ibn ‘Arabi’s house, he found, much to his surprise, a veritable palace surrounded by elaborate gardens. He saw many servants going back and forth and was served a sumptuous meal on gold plates by beautiful young women and handsome young men. Finally he was brought to Ibn ‘Arabi, who was wearing clothing fit for a sultan. He conveyed his teacher’s greetings and repeated his teacher’s request for spiritual guidance. Ibn ‘Arabi said simply, “Tell my student that he is too worldly!” The merchant was shocked and offended by this advice coming from someone living in such worldly opulence.

When he returned, his teacher immediately asked about his meeting with Ibn ‘Arabi. The merchant repeated Ibn ‘Arabi’s words and added that this sounded totally absurd coming from such a wealthy, worldly man.

His teacher replied, “You should know that each of us can have as much material wealth as his soul can handle without losing sight of God. What you saw in him was not merely material wealth but great spiritual attainment.” Then the teacher added, with tears in his eye, “Besides, he is right. Often at night as I make my simple fish-head soup, I wish it were an entire fish!” ‘

The comment finishes, and others agree:

Plausible, but one might think “pretty convenient for those with temporal power, especially holders of hereditary positions, as a lot of Sufi leadership is.”

Responders were also not thrilled by Ibn ‘Arabi’s behavior:

Truly, The Donald is the greatest Sufi master.


Yyyyeah… that sounds a lot like the kind of abuse that gets used as an example when they warn you about joining a cult.

The fish-head monk understands.

Often at night as he makes his simple fish-head soup, he wishes it were an entire fish.

This desire is not compatible with his spiritual path. He journeys hungry and jealous. This gross distraction halts his progression on his path. When one is constantly hungry, it is impossible to focus on other things. His mind turns to fish. He is too worldly!

He also seeks to be seen as the fish-head monk, concerned with what others see rather than his own path. This also is not compatible. Again he is too worldly.

Ibn ‘Arabi admonishes the monk not for giving up too little, but for giving up too much. He says, eat the whole fish! We each must eat our fill. Then turn to your quest.

Ibn ‘Arabi faces the opposite challenge. Wealth frees us from worldly desire by meeting our needs. Wealth enslaves us by adding new needs. Our needs remain unmet.

Perhaps Ibn ‘Arabi has all he needs, and turns his mind fully to the spiritual. What a great man! Truly he has the wealth that he can handle.

Perhaps Ibn ‘Arabi needs all he has, and turns his mind fully to the spiritual. Perhaps not a great man. But a good man. A man trying. Truly he has the wealth that he can handle.

Perhaps Ibn ‘Arabi has all he needs, and turns his mind to needing more. Perhaps he spends his day separating marks from their wealth, ordering servants and indulging in hedonism. What a terrible man!

Which is he? Impossible to know for sure. Listen for the likelihood ratios. Update your priors.


Posted in Uncategorized | 1 Comment

Sabbath Commentary

Epistemic Status: Several months of experimentation, then talking from the hip

Commentary On: Bring Back the Sabbath

Required: Slack

I have a lot of thoughts on the topic that don’t belong in the main presentation. I’m going to put them here in disorganized form for the curious. These claims are believed, but I may not have good explicit evidence to defend them with. I’m fine with that.

Continue reading

Posted in Uncategorized | 6 Comments

Bring Back the Sabbath

Epistemic Status: Several months of experimentation

Previously: Choices are BadChoices Are Really BadComplexity Is Bad, Play in Easy Mode, Play in Hard ModeOut to Get You, Slack

For More Thoughts After: Sabbath Commentary

Alternate Take (Endorsed): Sabbath Hard and Go Home.

Slack is life. It is under attack. We must fight for it.

Choices Are Bad. Really Bad. We need a break.

Complexity Is Bad. We need a break.

Work is exhausting. We need a break.

Relaxation is hard. Our attempts fail or backfire.

The modern world is Out to Get You. We need a break.

We need time for ourselves. Time that is truly our own.

Without setting aside such time, that won’t happen. Even when you take time, you’ll be continuously choosing to take time, and… well, whoops.

Modern life made the problem worse, but the problem is ancient. The ancients had an answer.

We need rules. We need ritual.

We need the Sabbath.

Cabin in the Woods

The parallels of my and Ben Hoffman’s Sabbath realizations are striking.

A few months ago, like Ben, I needed a break. My job puts me under constant pressure. My weekends weren’t refreshing me. Like Ben, I experimented with camping. Like Ben, I had no spare battery, and left my phone off.  I read The Great Transformation. I had meant to do that for weeks. I loved the world leaving me alone. Like Ben, I could relax, slow down, think.

I wasn’t worried about things I could be doing – I couldn’t do them.

Could I get this without the trip? Friends had started hosting Friday night dinners. What about the whole thing? What if we brought back the Sabbath?

Tradition makes rules easier to justify and explain, to others and yourself. These rules were time tested. I could take them and make them my own.

I thought about the components. Which made sense? What rules would let me cut the enemy, and relax?

Return of the Ritual

Rituals need clear beginnings and endings.

Sabbath begins with candles. One lights two candles, and recites a blessing.

For the evening meal, one says additional words and blessings, drinks wine, eats bread from one of two whole loafs and sits down to a proper meal with friends and family.

The candles are a signpost and deadline. Your week is complete and your work is done. There will be guests, so the apartment is ready. The ritual objects, and your needs for tomorrow, are secured. The meal is prepared. Time to feast and relax!

Slack is thus preserved in five ways.

This creates a time and place to see friends and family. Most want more social events, but coordination is hard and events are work. Now there’s always Friday night.

They increase the value of improving your home. Every week you notice the little things that enrich meal, visit and home. They’re Worth It, but easy to forget. Enhancing the little things enhances your life.

They prevent accumulation of personal-and-home-related work debt. A chaotic house is not restful. Postponed chores weigh on youThe deadline forces handling them in advance. Payoff is immediate.

By moving work before the deadline you are forced to make time during the week. You don’t eat into Slack. If you can’t find the time, this alerts you. Emergency!

They create visible failure as you approach hard bounds. When emergency arises, you sacrifice from the ritual. This signals emergency before life falls apart. You still have reserves. The ritual is Slack.

Sabbath ends with another candle. This prevents doing work until you go through non-trivial motions. You must do it on purpose.

Four Freedoms

We need restrictions that free us from the world. We need a new four freedoms.

We need freedom from work. Decide what counts as work to you. Don’t do that. Anything done for money is automatically work. During the week, time is money. Today, do what you value.

We need freedom from interruption. Space to think. Cut off the outside world. Especially cut off anything continuously updating and all periodic rewards. There lie Skinner boxes. Much of the world is out to get you. Today it can wait. Friendly visitors are welcome, but ideally arranged in advance.

We need freedom from choice. Full freedom from choice requires a step beyond the traditional rules. In my version, even among permitted activities, only those explicitly selected in advance are available – particular books, radio stations and so forth – plus things you feel intrinsic motivation to do. No lists. No browsing.

We need freedom from stress. Stressful conversations are not allowed. Doing work is not allowed. Making decisions is not allowed. Outside information is not allowed. If something was still going to stress you out and it was fixable, fix it before the Sabbath. Things can’t change on their own, and you can’t make them change. Why stress?

Sabbath Easy, Sabbath Easy, Sabbath Hard


Tension exists between that which is most restful right now, and that which would be a stable set of rules. There are two Easy Modes, representing each extreme.

One extreme is Orthodox Sabbath. This uses a strict, fixed set of rules. Pure deontology. You can’t carry objects without special preparations. Many objects you can’t even touch. This interferes a lot with relaxation, and forces realignment of life to prevent that. That can be good. There are even rules about violating the spirit of the rules – if you violate the spirit without breaking even those rules, that’s almost encouraged. Restrictions allow maximization.

Another extreme is Reform Sabbath. This asks, what would be most restful today? This is utilitarian and uses causal decision theory. Sabbath is for rest, so if driving a car or making a call would be more restful, do that. You could break the rules. This destroys freedom from choice. Who respects such boundaries? You won’t have urgency before the Sabbath. You can handle things later. Wouldn’t that be more restful? 

The Hard Mode approach asks, what sustainable rule set best preserves long run Slack? Taking stock and encouraging Slack-preserving outside the Sabbath are explicit goals. It uses logical decision theory. It creates personalized rules you can follow that work for you, but understands each divergence you select is expensive.

It asks what would be in the spirit of the rules, and modifying the rules to reflect that spirit. It views breaking current rules during the Sabbath with extreme skepticism, to reinforce following the rules. It modifies rules on Sunday.

In choice-related ways my current system is more restrictive than the Orthodox version. Mostly it is less restrictive, but becoming more restrictive over time. I currently allow Level 4 but everything there is on the chopping block. On Friday night I restrict to Level 2.

Hierarchy of the Shabbistic

There exists a hierarchy of shabbisticness. At one end are activities aligned with the goals of relaxing, recharging and unplugging. Sleep certainly qualifies. At the other are activities perfectly in conflict with those goals. Work done for money.

The hierarchy’s details are different for different people. If you see something as work, it drains you. Move it down towards the unshabbistic. If you see something as invigorating, and have the spontaneous urge to do it for intrinsic reasons, move it up towards the shabbistic.

Then draw a hard line. Deciding whether to allow something is an impactful choice (itself banned) and a slippery slope. The golden rule of Sabbath is not breaking the  rules. When in doubt, don’t do the thing, then refine your rule on Sunday.

I encourage stricter rules for Friday night than Saturday. This enriches without being stifling.

This is my current hierarchy. Levels 1-2 I consider purely good, Levels 3 good, Level 4  questionable. Level 5 is bad, Level 6 very bad. Level 7 is banned all week.

  1. Pure rest. Sleep. Rest. Walking. Intellectual discussion. Friendly discussion. Reading physical books and other physical objects. Meditation. Museums. Taking a bath.
  2. Active rest. Sex. Flirting. Running. Swimming. Playing sports. Arguments for low stakes. Board and card games with no stakes. Puzzles. Building models. Taking a shower. Eating. Watching sports in person. Light switches.
  3. Consumptive rest. Riding elevators. Radio with one station. Listening to music. Food preparation without lighting a fire. Window shopping. Kindle and other e-books.
  4. Potentially toxic actions. Writing for yourself. Taking notes. Practicing and training personal skills that are not work or work related. Working out. Computer games. Pre-selected television. Phone calls and texts for physical coordination purposes. Riding in cars and trains (without payment).
  5. Violations of compactness. Phone calls and texts not for same-day logistical coordination. All other use of smartphones. Making impactful decisions. Planning. Flipping stations on television or radio. Browsing the internet. Browsing a giant music collection. All long lists, especially lists of choices. Checking anything that continuously updates. Lighting a fire. Stressful topics of conversation.
  6. Work and outside demands. Exchange of money. Doing business. Anything that earns money or creates commercial value. Negotiations. All continuous updates. Email.
  7. Considered harmful. All timed and daily rewards. Micro-transactions. Social media.

The Rules Simplified

Start here. Adjust as needed.

Light candles before sundown Friday to begin.

No outside inputs except in person.

No choices impacting post-Sabbath.

Light and extinguish no fires. Do no work or business. Spend no money.

Only preselected and spontaneously motivated actions are allowed. No browsing. No lists.

Light another candle after sundown Saturday to end.

State of Emergency

I brought back Sabbath for Slack and relaxation. Ben brought it back as an alarm system, for when life was out of control. Sabbath shows when you are not okay, and provides method and incentive to get back to okay.

This Saturday I did full Orthodox Sabbath (minus prayer), and also fasted, as an experiment. I won’t do this every week or even month, but it had important alarm value.

Ben’s post is excellent. Read the whole thing. I’ll finish with two key passages from it.

Key motivation:

You would not want to do this sort of thing all the time. But it might make sense to do periodically – perhaps once a week – as a stopgap measure to combat attention drift. If powerful and pervasive cultural forces are out to get you, you ought to check in from time to time with yourself, and other people with whom you have local, high-quality relationships, to give yourself a chance to notice whether you have gotten got for too much.

His conclusion is important and worth quoting in full:

One more useful attribute of the Jewish Sabbath is the extent to which its rigid rules generate friction in emergency situations. If your community center is not within walking distance, if there is not enough slack in your schedule to prep things a day in advance, or you are too poor to go a day without work, or too locally isolated to last a day without broadcast entertainment, then things are not okay.

In our commercialized society, there will be many opportunities to purchase palliatives, and these palliatives are often worth purchasing. If living close to your place of employment would be ruinously expensive, you drive or take public transit. If you don’t have time to feed yourself, you can buy some fast food. If you’re not up for talking with a friend in person, or don’t have the time, there’s Facebook. But this is palliative care for a chronic problem.

In Jewish law, it is permissible to break the Sabbath in an emergency situation, when lives are at stake. If something like the Orthodox Sabbath seems impossibly hard, or if you try to keep it but end up breaking it every week – as my Reform Jewish family did – then you should consider that perhaps, despite the propaganda of the palliatives, you are in a permanent state of emergency. This is not okay. You are not doing okay.

So, how are you?






Posted in Uncategorized | 13 Comments


Epistemic Status: Reference post. Strong beliefs strongly held after much thought, but hard to explain well. Intentionally abstract.

Disambiguation: This does not refer to any physical good, app or piece of software.

Further Research (book, recommended but not at all required, take seriously but not literally): The Book of the Subgenius

Related (from sam[ ]zdat, recommended but not required, take seriously and also literally, entire very long series also recommended): The Uruk Machine

Further Reading (book): Scarcity: Why Having Too Little Means So Much

Previously here (not required): Play in Hard ModePlay in Easy ModeOut to Get You

Leads to (I’ve been scooped! Somewhat…): Sabbath Hard and Go Home

An illustrative little game: Carpe Diem: The Problem of Scarcity and Abundance


Slack is hard to precisely define, but I think this comes close:

Definition: Slack. The absence of binding constraints on behavior.

Poor is the person without Slack. Lack of Slack compounds and traps.

Slack means margin for error. You can relax. 

Slack allows pursuing opportunities. You can explore. You can trade.

Slack prevents desperation. You can avoid bad trades and wait for better spots. You can be efficient.

Slack permits planning for the long term. You can invest.

Slack enables doing things for your own amusement. You can play games. You can have fun. 

Slack enables doing the right thing. Stand by your friends. Reward the worthy. Punish the wicked. You can have a code. 

Slack presents things as they are without concern for how things look or what others think. You can be honest.

You can do some of these things, and choose not to do others. Because you don’t have to.

Only with slack can one be a righteous dude.

Slack is life.

Related Slackness

Slack in project management is the time a task can be delayed without causing a delay to either subsequent tasks or project completion time. The amount of time before a constraint binds.

Slack the app was likely named in reference to a promise of Slack in the project sense.

Slacks as trousers are pants that are actual pants, but do not bind or constrain.

Slackness refers to vulgarity in West Indian culture, behavior and music. It also refers to a subgenre of dancehall music with straightforward sexual lyrics. Again, slackness refers to the absence of a binding constraint. In this case, common decency or politeness.

A slacker is one who has a lazy work ethic or otherwise does not exert maximum effort. They slack off. They refuse to be bound by what others view as hard constraints.

Out to Get You and the Attack on Slack

Many things in this world are Out to Get You. Often they are Out to Get You for a lot, usually but not always your time, attention and money.

If you Get Got for compact amounts too often, it will add up and the constraints will bind.

If you Get Got even once for a non-compact amount, the cost expands until you have no Slack left. The constraints bind you.

You might spend every spare minute and/or dollar on politics, advocacy or charity. You might think of every dollar as a fraction of a third-world life saved. Racing to find a cure for your daughter’s cancer, you already work around the clock. You could have an all-consuming job or be a soldier marching off to war. It could be a quest for revenge, for glory, for love. Or you might spend every spare minute mindlessly checking Facebook or obsessed with your fantasy football league.

You cannot relax. Your life is not your own.

Our society even does it to our children. They stay on college and career tracks, fearing one bad grade or mark on their record destroying their whole lives. All is negative selection. Every spare minute is spent padding resumes. In some circles this starts before kindergarden.

It might even be the right choice! Especially for brief periods. When about to be run over by a truck or evicted from your house, Slack is a luxury you cannot afford. Extraordinary times call for extraordinary effort.

Most times are ordinary. Make an ordinary effort.

“You Can Afford It”

People like to tell you, “You can afford it.”

No, you can’t. This is the most famous attack on Slack. Few words make me angrier.

The person who says “You Can Afford It” is saying to ignore constraints that do not bind you. If you do, all constraints soon bind you.

Those who do not value Slack soon lose it. Slack matters. Fight to keep yours!

Ask not whether you can afford it. Ask if it is Worth It.

Unless you can’t afford it. Affordability is invaluable negative selection. Never positive selection.

The You Can Afford It tax on Slack quickly approaches 100% if unchecked.

If those with extra resources are asked to share the whole surplus, all are poor or hide their wealth. Wealth is a burden and makes you a target. Those visibly flush rush to spend their bounty.

Where those with free time are given extra work, all are busy or look busy. Those with copious free time seek out relatively painless time sinks they can point to.

When looking happy means you deal with everything unpleasant, no one looks happy for long.

The Slackless Like of Maya Millennial

Things are bad enough when those with Slack are expected to sacrifice for others. Things are much worse when the presence of Slack is viewed as a defection.

An example of this effect is Maya Millennial (of The Premium Mediocre Life of Maya Millennial). She has no Slack.

Constraints bind her every action. Her job in life is putting up a front of the person she wants to show people that she wants to be. If her constraints noticeably failed to bind the illusion would fail.

Every action is being watched. If no one is around to watch her, the job falls to her. She must post all to Facebook, to Snapchat, to Instagram. Each action and choice signals who she is and her loyalty to the system. Not doing that this time could mean missing her one chance to make it big.

Maya never has free time. There is signaling to do! At a minimum, she must spend such time on alert and on her phone lest she miss something.

Maya never has spare cash. All must be spent to advance and fit her profile.

Maya lacks free speech, free association, free taste and free thought. All must serve.

Maya is in a world where she must signal she has no Slack. Slack means insufficient dedication and loyalty. Slack cannot be trusted. Slack now means slack later, which means failure. Future failure means no opportunity.

This is more common than one might think.

“Give Me Slack or Kill Me” – J.R. “Bob” Dobbs

The aim of this post was to introduce Slack and give an intuitive picture of its importance.

The short-term practical takeaways are:

Make sure that under normal conditions you have Slack. Value it. Guard it. Spend it only when Worth It. If you lose it, fight to get it back. This provides motivation for fighting things Out To Get You, lest you let them eat your Slack.

Make sure to run a diagnostic test every so often to make sure you’re not running dangerously low, and to engineer your situation to force yourself to have Slack. I recommend Sabbath Hard and Go Home with my take to follow soon.

Also respect the Slack of others. Help them value and guard it. Do not spend it lightly.

A Final Note

I kept this short rather than add detailed justifications. Hopefully the logic is intuitive and builds on what came before. I hope to expand on the details and models later. For a very good book-length explanation of why lacking Slack is awful, see Scarcity: Why Having Too Little Means So Much.




Posted in Reference | Tagged , | 13 Comments