I just tried asking GPT-4 to respond to me, placing a question near the end of its response, and to roll a die at the end. If the die has its maximal value, I need to spend a minute thinking before I respond, else I can just respond normally. And it worked. GPT-4 asked good enough questions, which lead to interesting enough places, that I'd recommend other people try this. Though please ask for, like, a three sided or even two-sided dice as a ten-sided dice is too much.
I chose 1-minute to make the feedback loops faster, and also because I think I understand why you recommended 5 minutes first and knew I didn't need to start at that value. For people whose minds are resistant to thinking at all on the spot, going blank or wandering off topic, 5 minutes gives them enough time to have a minor flash of inisght relevant to whatever they're meant to think of. Then they can start improving that over time until they can think for several minutes if forced to. I've already got that skill, I just haven't made it habitual. So that's why I went with 1 minute.
But you can also make feedback loops tighter by asking GPT-4 to give short responses, and sticking to short questions yourself so you can simulate answering questions under pressure and resisting that pressure when needed.
Another way to improve things would be to notice what questions are important yourself, but being able to notice that you can just think about a question at all, and practicing that motion of thinking, is a useful subskill of its own. Then you can practice noticing when a question is worth seriously thiking about. Also, GPT-4 just asked a lot of questions worth thinking about, so I don't think you need to worry about it never asking you questions worth thinking of.
In fact, you could ask it to just throw you some softball questions and that may well work.
Huh! I would not have previously guessed chatGPT had enough of the social norm pressure to get people to respond too fast. Thinking about it, that's irrelevant, because I can just practice thinking by the clock with a textbook. Upvote, that seems like a potentially useful way to practice and you're practicing!
I do suggest trying at least one pause on a 5 minute timer to see if anything interesting happens with the extra minutes, not because I particularly doubt you, but it is a cheap test to try. (By definition it's not going to take that long!)
I'm definitely going to vary the pause duration.
But I'm curious how you practiced this skill. Are you one of those incomprehensible beings that can just set a trigger action plan for something like that? I've always struggled to create TAPs as sophisticated as that, so I'd be curious if you had some other method. LIke, going on Omegle (RIP) with a checklist?
Socially, I had the opposite problem and learning this involved relaxing a different skill.
I went from taking way too long to answer social conversation cues to basically normal speed. The training regime for that involved a lot of games[1] and improv practice. By forcing myself to move fast and allow myself to make mistakes I wouldn't have made if I'd thought about it for longer, I got faster at speaking in conversation.
One of the games I played was chess, and chess clocks are decent practice for this! I looked up chess puzzles and spent five minutes with a clock trying to solve them, or played games with wonky clocks where I was once-a-game allowed to pause for five minutes to think and that pause didn't count against the regular clock. (A lot of these games were against the computer, which was very patient.) A thing that I think I'm better at now, but could not put into useful words yet, is thinking within the box of five minutes. (or other timers.) There's an idea of. . . noticing wasted motion and not repeating loops, of deliberately splitting the time into brainstorming and winnowing and refining and murphyjutsu, that I improved by slow trial and error.
As for taking the time out in conversations these days, it's largely tied to a weird little alarm in the back of my head that goes off when it notices longer term or large resource commitments about to happen, plus the ability to turn off (or rather, stop maintaining) a different set of alarms for when I'm about to violate social norms. The trickiest part for me happens at the interface, and I suspect that's unusually unique to how my head is configured.
- ^
I used to play RTS games like Age of Empires by hitting the pause button whenever I wanted to think for longer, often taking three or four times as much time in pause as I did letting the clock advance.
Promoted to curated: I thought this post was both quite well-written, and it covers one of the things in the LessWrong rationality catalogue that I found most valuable (though of course other schools of thought also have their own versions of this, I remember some references in Feynman's autobiographical books to similar cognitive habits).
The converse of fast paced conversations leading people to say stupid things is: if someone says something foolish it maybe not be a lie nor a tendency to BS nor stupidity. They may have responded faster than they thought. You can correct them not by refuting what was said but allowing them a moment to reconsider. A liar, politician, and fool are hard to reform or work with. “Stopping to think” may be easier trained and much easier used to keep a conversation on track.
Love this post. I've also used the five-minute technique at work, especially when facilitating meetings. In fact, there's a whole technique called think-pair-share that goes something like:
- Everyone think about it for X minutes. Take notes.
- Partner up and talk about your ideas for 2X minutes.
- As a group, discuss the best ideas and takeaways for 4X minutes.
There's an optional step involving groups of four, but I'd rarely bother with that one unless it's a really huge meeting (and at that point I'm actively trying to shrink it because huge committees are shit decision-makers).
Thank you for the addition! Pairing up to talk through the ideas seems like a good group technique, I like the sequence you outline!
I agree with the main points this post is making. However, I do want to disagree with something that I think is more tangential.
In a real time conversation with another person, I don't think that you need to pause for dozens of seconds of silent thinking if you don't "know the answer". Instead, you can think out loud. Whatever the inner dialog would be during the silent thinking, you can verbalize it. Something like this: "Hm, I'm not immediately sure. That's an interesting question. Let's think about it. My first thought is that I'm confused about X..."
I'm not saying that it's always appropriate to think out loud. Sometimes it's more appropriate to sit and think silently.
The skill to practice is to spot that you just tried to think very quickly, and to check if this is a decision which you would like to think about more. The technique you would like to train is that instant of noticing
Hm. My first thought was: "Yes. This. Very much this." But after thinking about it some more, I'm not so sure.
I was playing racquetball yesterday, and something really stuck with me. I was complaining to the guys about how my body isn't the same as it was when I was younger. There's various aches and pains and old man stuff (I'm 31). For example, I went to the gym two days before to do some very light cardio on the stair machine, and 10 minutes into it I had this weird pain in my knees. But it was in this weird spot on the inside of my knees. Same for each knee. And I've never had it before. And two days later playing racquetball, I feel totally fine.
Anyway, this guy interrupted me and said "Y'know what it is? Wheat."
"Wheat?" some other guy responded. Yeah. The first guy explained that his doctor told him something about how white people process wheat in some way that leads to knee pain, and that when he stopped eating wheat his knee problems went away.
This obviously seems very dubious. There's a lot wrong with his thinking. But I went with the objection of "why this inner location of my knees?". It's not general knee pain. It's pain in a very specific area. Why would wheat cause pain in that specific area?
It wouldn't. Was what we all realized. But why would the first guy be so confident that he knew the cause of my knee problems to begin with? I dunno. Lots of reasons. But -- and here's my point -- imagine that this wasn't the case. Imagine that the guy was a perfect Bayesian who always updates his beliefs the appropriate amount in response to pieces of evidence.
Well, then he would just be uncertain about the cause of my knee pain, and if he wanted to reach a higher level of certainty, Thinking By The Clock kinda follows naturally from that. More generally, if you 1) aren't that confident and 2) want to be more confident, then perhaps there's no need for the Thinking By The Clock skill.
Actually, let me backpedal a bit. None of us are anything close to the sort of perfect Bayesian I asked you to imagine. In reality, like the guy at racquetball, we're all prone to, mistakenly jumping to conclusions. I think the Thinking By The Clock skill helps to guard against this. Like, if you notice 1) that a thing is important and 2) you didn't spend the amount of time that seems like it should correspond with the level of importance (ie. buying a house after 30 minutes of research), hopefully those two things would trigger a "let me set a timer and really think about this".
There's one thing I don't understand. How the hell did you find introductory circuit theory interesting? Setting up and then solving all those simultaneous linear equations was tedious as hell and bored me to tears. :/
Dunno what to tell you. I played the lottery of fascinations and got lucky. Electricity is our world's equivalent of mana, and if I'm smart enough to inscribe the right patterns of sigils in metal the right way I can command the forces of the universe. Who can turn that offer down?
Well, I liked learning about mathematics, programming, and physics a lot more; I basically ended up low-grade traumatized by some of my electrical engineering classes, and circuits classes were among them. If I ever have to calculate another voltage or current, I'm first going to scream, and then if it's anything but a trivial calculation, I'm going to make a computer do it for me.
[See corrections in replies; "think for five minutes" was in EY posts as far back as 2007, the HPMOR chapter was in 2010, and the first CFAR retreat (though not under that name) was 2011 IIRC. Still curious to know where he got it from.]
Before HPMOR, "think for five minutes by the clock" was a CFAR exercise; I don't recall where they picked it up from.
The LessWrong Review runs every year to select the posts that have most stood the test of time. This post is not yet eligible for review, but will be at the end of 2024. The top fifty or so posts are featured prominently on the site throughout the year. Will this post make the top fifty?
I.
I'm sure Harry Potter and the Methods of Rationality taught me some of the obvious, overt things it set out to teach. Looking back on it a decade after I first read it however, what strikes me most strongly are often the brief, tossed off bits in the middle of the flow of a story.
This was the very first lesson of LessWrong-style Rationality I actually started trying to deliberately teach myself as a result of my contact with HPMoR and the sequences. This is the powerful technique of actually Thinking By The Clock.
I used to call it Thinking For Five Minutes, but that technique name is a misnomer. It's practically a lie-to-children really. Sometimes I think for much less time, about thirty seconds. Sometimes I think for much more time, like a couple of days. Still, in the way that when you first learn martial arts you might stand in an awkward, stiff stance without turning or stepping I first learned to think by the clock in increments of exactly five minutes.
II.
When I first went to a gym to lift weights, I did it with a friend. I didn't think it was going to work very well (I was a pretty skinny guy) but I wanted to humour them. I sat down on the bench they pointed me at, got a good grip on the heavy thing they wanted me to grab, and lifted it up and down for a while. When they said stop, I stopped. "That seemed kind of fast," I recall saying, "are we done?" Dear reader, we were not done. This pattern repeated when I first started going jogging with a different friend. I somehow expected the whole running thing to last, you know, until we got bored, which happened pretty quickly.
(If I may say a word in defense of younger!me, he really wasn't as unfit as this sounds. Soccer was fun and interesting, and I ran around plenty playing that. Stacking haybales got me paid, and I was quite willing to be paid to lift heavy things as long as I was told.)
So it may not come as a surprise to you that when I first encountered a hard intellectual task that was neither entertaining nor immediately profitable, I kind of bounced off. Going by memory, that was probably Calculus. I hated Calculus. I'd sit down at the table to do my homework or study for a test, and find myself reading the problem a couple of times and then glancing at the clock or looking longingly at the Circuit Theory textbook. (My definition of "entertaining" surprised a lot of people.) When the TA asked however, I'd say that I studied Calculus for a few hours. I was sitting at the table, wasn't I?
Just as sitting on a barn stool in the haybarn[1] will completely fail to get the bales stacked no matter how long you do it, sitting at the desk staring at the clock will completely fail to get the idea of derivatives into my head.
III.
But you know, that's not exactly the problem Fred and George had in the quote above, was it? They were presumably doing some thinking in those two seconds. So let me talk about a neat bit of cultural anthropology.
When two people are talking, there's a gap between when one person finishes and the other picks up. Since neither of them know in advance when they'll be finished talking and periods don't actually get pronounced, the listener has to wait a short while before starting to speak. If the listener doesn't wait long enough, they interrupt and talk over the other person. If the listener waits too long, you can get an awkward silence.
I used to be really bad at figuring out how long to wait. I'm told when I was a child it was not uncommon for someone to ask me how I was, and then to have to wait for half a minute (thirty whole seconds!) or more before I would begin to speak. When I did speak, what I'd say would be composed and unexpectedly thought-through. I think what was going on was, I had no idea that you were supposed to respond quickly, within the back-and-forth flow of conversation.
I had to learn that, deliberately forcing myself to start speaking quickly. For a while I made great use of non-committal openings such as "you know, I was thinking about it and..." which allowed me to gain an extra few precious seconds to figure out what I wanted to say next. Eventually though I got decent at the skill and now I find the silences as awkward as most people. It's often easy for me to pattern match a shallow, accurate-enough comment that fits the tone of the conversation and just say that. If I do, then often the conversation flows smoothly along. Yay, success!
Then this simple exchange in some funny Harry Potter Fanfiction made me realize that I'd learned to cut off my thoughts after a second or so if I was in conversation, and this sometimes made me say really stupid things. Once you're aware of this, sometimes you can get people to say stupid things themselves by setting up a slightly too-quick cadence of conversation and gently guiding them away from stopping to think.
Did I say that was something you could do to other people? Well guess what- people can do it to you too.
Perhaps the most dramatic occasion where I have used the skill of Thinking By The Clock was during a job interview. It had been an interview with multiple rounds with multiple different people, and at this point I was talking to the HR department. We had a good conversation going about what I thought about the office layout, what kinds of work I was excited about, and then they asked me without any indication that this was a different kind of question what kind of salary I was looking for.
Reader, I could see the obvious, accurate-enough response that would keep this friendly conversation going. I had put a bit of thought into this before arriving at the interview, I could just pull that cached number out quickly and we could keep talking. I was about to. I'd practiced the skill of keeping a conversation flowing often enough. Fortunately though, I'd also practiced the skill of noticing an important question I shouldn't blithely continue past. I didn't stop the conversation completely cold, but I did say a variation on "that's an important question, and I don't think it would be an obstacle if we decided we were a good fit. Let me think it over and I'll get back to you? I'm curious what range you had in mind."[2][3] They pushed back a bit, and I did stop the conversation for thirty seconds to actually think through my answer.
There are many moments in an ordinary person's life when you will be asked a question in conversation which will require you to think. The norms of conversational flow will suggest you think for a few seconds at most. These norms are lovely when talking about the weather or how the kids are doing, and they are antithetical to questions of how to best achieve a long range project or how to solve a difficult question.
I have used Thinking By The Clock in work meetings. (Adam: "How do we add this feature to the UI?" Bella: "How about as a sidebar?" Adam: "Great, that sounds good." Me: "I've got an idea. Lets each think about this for two minutes and write down our ideas, then we share the ideas?")
I have used Thinking By The Clock in discussions with my girlfriend. (No, I'm not giving you an example right here. If you're that curious how my romantic life works, I'm poly, feel free to ask me on a date sometime.)
You don't actually need the social pressure of the conversational gap to make this a useful technique. That's just where I see people make this mistake most often. Selection bias though, obviously I don't see as many mistakes other people make when they're alone.
The skill to practice is to spot that you just tried to think very quickly, and to check if this is a decision which you would like to think about more. The technique you would like to train is that instant of noticing, the tiny alarm bell saying you should give your reply or it'll be an awkward silence when that alarm bell is full of lies. If this was one movement of a kata in the martial art of rationality, that movement would start by being posed a question and would let you open your mouth to give a quick answer before catching yourself and taking the time to actually Think By The Clock.
Start with five minutes. Train yourself to use the time. Then practice smaller increments. Do not allow yourself to goof off or be distracted.
IV.
All off that is about thinking too fast and needing to slow down. What about the other direction?
Sometimes I hear someone say they've been thinking about a difficult problem for weeks, months, or even years. Sometimes that person is me! It can be both true that someone's been thinking about it for years and that good thinking is happening, but I often doubt it.
My mother gave me a cheap guitar as a holiday present when I was around thirteen. I found the instrument fascinating, playing with it to try and find out what sounds it could make. I had heard guitar music of course, but nobody in my family or immediate circle played. I kept the guitar with me for the next fifteen years, pulling it out of my closet or from under the bed in my dorm room or from the back of my car to fumblingly tune it, pluck a few strings, look up a chord or two, and enjoy the sound of it filling the room.
You could have said I played guitar for years, and it would have been true, but it would have been exceptionally misleading. Over the first decade I would be surprised if I spent more than a hundred hours with the guitar on my lap, and only a dozen of those hours would have been with an instructor or a teacher showing me what to do. Last year I decided I wanted to actually be able to play some music on the guitar, and I've probably spent more time actively practicing in the last month than I spent touching the instrument at all in the second year I owned it.
Before I read those lines in HPMoR, I too often thought about something for years in the sense of pulling the thought out of the back closet and noodling around with it for a bit before putting it away. I'll still use language like that sometimes, but now I have and additional version for when I actually want to find an answer or make something with my thoughts.
Think about the problem like you're practicing guitar, and you need to do the music for Secular Solstice next month. [4]
If you're like me, that just threw some things into sharp relief.
Don't think about it while distracted. Have some way to take notes or carry the information forward to the next session. You need a clear idea of what "good enough" looks like to know if you're getting there. Be ruthless about discarding lines that won't lead to the goal, but also don't keep banging your head against the same brick wall you aren't making progress on. Don't say to yourself that you're working on it when you're staring at the beguiling Circuit Theory textbook, so full of secrets like Kirchhoff's Laws and Thevenin's Theorems (Calculus eventually forced me to give proper battle) just waiting for you to stray.
This is how I thought about the decision to make a wild career swerve in my thirties. This is the approach I used when deciding whether to move out of my home state to the big city. I picked a few weekends, kept those weekends as clear of distractions as I could, and used a stopwatch to actively work on the problem for two whole days by the clock. I expected real progress to be made, even if that progress was in the form of ruling things out or crossing off blind alleys, and I kept trying.
Start with five minutes. Train yourself to use the time. Then practice larger increments. Allow time for breaks and for maintaining your life, but do not confuse making dinner for focusing on a problem.
V.
Thinking By The Clock is one of the more obvious mental moves which I can cleanly ascribe to Yudkowsky's writing. I wish I could describe how to do it to myself when I was younger, as I would have benefited from the technique. Lacking time travel, the next best thing I can do is explain it in far more words than Yudkowsky used, in case someone did not absorb the core idea and generate the steps to train it into reflex just by reading that conversation between Harry and the Weasley twins. Write A Thousand Roads To Rome and maybe one of them will lead someone there when all the other roads failed.
That is the hard part I think. Not to just absorb the idea, not to realize it's important, but to practice the move until you do it even against social pressure or when you're thinking of other things. My best and favourite sensei once told me a story of slipping on the ice while checking his mailbox in the morning, and finding himself moving into a perfect rearward fall. That is the level of ingrained use that a proper martial art of rationality would aim for, I believe.
To be able to think by the clock, even when taken by surprise before you've finished your morning coffee. That's what I aspire to. Ars Longa – the art is long! Tsuyoku Naritai – I want to become stronger!
. . .okay, this is the inaccuracy that's going to bug me if anyone points it out. No, there usually wasn't a stool in the haybarn.
Lest that sound too practiced, I will admit that I also rehearsed that answer a bit thanks to Patrick McKenzie's Salary Negotiation advice.
I'm actually irritated at myself that I didn't do this in my last job interview. I saw the moment, I flagged the thing to think about, and I didn't do it both because I valued the other person's time quite highly and because I considered us to be on a more cooperative footing than is typical for job interviews.
Fortunately for both Boston and New York City, I am not the main musician for either city's secular solstice. I'm both hopeful and currently unprepared for the circumstance where I get to play something at Boston's Solstice. If you happen to visit, feel free to say hello!