Posts
Comments
Wow, that's very cool! This is actually the closest existing thing to Arbital I've heard about. I guess to Eliezer's point, it does sound like they built this system in one go, without doing much of MVP-kind of iteration. But also... damn! From the sound of it, there are a lot of complicated components, and a lot of manual work that went into the whole thing. If those are the minimum requirements to build something like that, then I'm not sure we ever had a chance.
But the good news is: if this really works, then I think Arbital is totally doable and would be just the online version of this. The remaining difficult part would be monetization, but in retrospect it's quite possible Arbital should have been a non-profit.
Are you doing homeschooling? How is that going? I’m just very curious because I’m thinking about it for my children too. What is forest school?
I feel like the obvious answer here is to look at the underlying intention.
This clip (from The Office) reminds me of when people suggest various patches to AI after it violates some safety concern / fails an experiment.
I have a one month old baby girl. I think the most important skill I can teach her is how to learn. There’s of course a bunch of sub-skills there: beginner’s mindset, paying attention, research, asking for help, checking, etc....
As you might imagine it’s a personal thing that does not translate very well. I do want to share it with people but I have not yet thought of a way to do it .. correctly.
Honestly it’s the hardest thing I’ve ever had to do. One month in and it’s getting a bit easier. She sleeps about 2-3 hours at a time.
I was walking along the same line of thought about 10 years ago. And a year ago it finally led me to a very very interesting update. Happy adventuring!
It’s a good question, and this scenario is something I’ve been thinking about as well. I don’t have a good answer. One obvious consideration is how easy is it to travel. By car you can go to Mexico or Canada. Otherwise you need to fly or take a boat.
I’m now a father of a 3 weeks old girl. So your parenting posts are extra useful and appreciated!
Would you / when would you recommend Brian Skyrms's Signals: Evolution, Learning, and Information?
Fermi paradox has a much simpler answer: https://slatestarcodex.com/2018/07/03/ssc-journal-club-dissolving-the-fermi-paradox/
Very similar to : https://www.lesswrong.com/posts/GaP3JK5kSABRa8GkK/the-mind-board-game-review
Can you convert the percents into the usd amounts. The percents are kind of meaningless since the budgets differ a lot.
I've still been thinking about this quote a lot. Do you think you can trace it more specifically? Is there a book / video / resource that explains it well?
This is one of my favorite games! I’ll definitely try the variant.
I’m getting more and more interested in this sequence. Very applicable to what I’m doing: analyzing market data.
I’m expecting a baby any day now. I guess this is the area I’m most concerned about, since historically I’ve felt very protective of my own time. But I’m looking forward to spending it with my kids, even in a “childcareish” way. :)
I second what Kaj said. If you’re curious I can go and copy the rules verbatim.
Rationality: not really. But the community: yes.
Below is my personal experience with the game. I decided not to include it in the post because I didn't want to spoil the experience for anyone else. So I do recommend you read this comment only after you've played the game yourself a few times and ideally got to the end.
Recently I've been on a huge board game kick. My wife and I have started to build up a small collection. I try to buy games that seem fun, but are also diverse and interesting. (I get most of my ideas from The Dice Tower channel, which I'd highly recommend). Anyway, that's how I ended up buying this game.
Our roommate Anna usually plays with us, so it was a three player game. The first night we played it, we got to level 5 twice. It was fun, interesting and challenging from the very first round.
One thing we learned almost immediately is that everyone's sense of "speed" is quite different. My wife, Parina, tended to be on the faster side, sometimes making large jumps. Anna tended to be a lot more cautious, which meant sometimes she'd be hesitant to play a card when she should have. Usually that meant that when the two of them would have a "conflict" (wanting to play a card before the other), it should be Anna who goes first. But even knowing that, it wasn't always easy to follow that rule (S2), since so much of the game is based in S1.
Another thing I also admire about this game is how there's no downtime. Even if you have a 100, which means you'll play last for sure, you still want to pay attention to the flow. It helps you calibrate, but it also means you could be a helpful tiebreaker if the other players are at an impasse. There were multiple amazing moments where two players would feel very uncertain, but they would turn to the third player to help them figure it out. (It's also just fun to observe.)
The second night we played it, we got noticeably better. First time we lost around level 5. But the next play through, everything clicked. We got to level 5 with more lives and shurikens than we've ever had before. The "flow" state was setting in. Basically every time we had close cards, we got the order right. I didn't think it was possible to do that, but it definitely wasn't luck. It felt like catching the perfect wave or like being in flow solving an amazing puzzle. (In a way it feels like the opposite of poker: you're trying to let the other players know what your cards are.)
Seriously, the feeling of getting close calls right is amazing!!! You're looking at 41 in the center. It's level 5, and your next lowest card is 47. You slowly try to put it down, when you notice the other player is doing it too. You lock eyes and enter a "communication dance". "No, you go" you gesture. "Eh, I'm not sure I should," they gesture. (And yes, at the peak experience it almost felt like I was hearing people speak, which is extremely rare for me. And to be clear, no drugs were involved.) You keep doing the dance. After a while you just know they have 48. Because if they had 46 they would have definitely played it by now. You take the chance and play your 47, and they follow by playing 48. The tension and the release is just pure crack. :P
Every level we completed after that we felt ecstatic! We celebrated and held hands. Sometimes we would make mistakes and lose a life. It was so emotional. In Anna's words: "I felt betrayed, alive, sad, angry, surprised." You get so in sync that it creates this strong illusion of one-mindedness that it almost hurts when it breaks.
Yet another thing I love about this game is that the collaboration works super well. If you've played cooperative board games, you've probably seen the common fail mode where one player just directs everyone else. This definitely can't happen here. Also when someone makes a mistakes, the mistake is symmetric. One person should have gone first and they didn't, whereas the other person should have waited longer and they didn't. So it's really hard to blame anyone in this game. You just learn and adjust.
We got to level 10 with 2 lives and no shurikens. A few cards in we lost another life. We were a hair away from losing the game. The next 8 cards were just pure adrenaline, but at the same time so calm and steady. No mistakes could be made. Every decision was approached with maximum consideration. When we finished the level the feeling was almost orgasmic. Like solving one of the most satisfying puzzles or hearing the most beautiful resolution to a song. There's not many things in this world that I've experienced like that.
One nice property of markets is that it only takes a small number of people to remove an inefficiency. Once the better, cheaper mousetrap hits the market, the old mousetraps become obsolete; the new mousetrap is adopted even if there are initially many more old-mousetrap-makers than new-mousetrap-makers. Same with the stock market: even if most investors don’t notice a pattern, it only takes a handful to notice the opportunity and remove it.
This is very very wrong.
Knowing that there's an inefficiency is only part of the battle. In order of importance to counter your point:
- You need to have enough volume to remove the inefficiency. For a trading strategy, this is the capital you can allocate to this strategy specifically. It's not obvious how many hedge funds, for example, are capital constrained. In the mousetrap example this means actually producing the sufficient volume of mousetraps.
- The opportunity costs matter. The people positioned best to notice the inefficiency are also positioned best to notice others. May be there is a good trading strategy but it's *less* than what you would make with your other strategies. So you wouldn't actually trade it, leaving it for others to pick up. Same with the mousetrap: may be you go build a cat trap, or may be you don't build anything but make money consulting.
- The inefficiency removal needs to stick. If you're trading a strategy and then stop, the inefficiency will likely come back. You might stop for a multiple number of reasons.
- The execution is important. Let's say you can analyze the data perfectly and come up with a clever algorithm. You'd still need to figure out how to execute on the actual trading. Part of it is not introducing too many exploitable opportunities for other trades (e.g. market makers).
Overall, I'd say even if an inefficiency is publicly known, it usually takes a non-trivial amount of people*effort to remove all of it.
I’m noticing this post is not getting many upvotes, so I thought I’d leave a comment to help with the lack of feedback.
Overall, I think this is a fine post. I agree with most points, with a few minor disagreements that overall don’t undermine the main point. It’s written pretty well too; very easy to read for me. But the content / main point is pretty well know, I think. And that’s why I’m guessing this post isn’t getting a lot of attention.
I’ve been aware and working on this dynamic for many years now, but I haven’t thought about it from this perspective. It feels pretty helpful! I think the underlying assumption is “of course I can deal with the first-order problem myself. But it would be helpful to not worry about the second-order effects while I do it.”
This is so so good! I think this is my favorite post of yours. It’s just teeming with possibilities and connections.
Melting/freezing and magnetization connections are very interesting. This almost seems similar to what happens with brain plasticity increase (due to mania, being excited, or substances).
First, I'll question some of the assumptions.
I do think there aren't enough leaders. (I'm happy to discuss whether Ender Wiggins was a good leader in comments.) Part of what makes someone a good leader is being able to recruit and work with other people, not just rationalists. In fact, I think for most rationalist projects, you need a rationalist leader, but not necessarily rationalist helpers. For example, Arbital didn't fail because we didn't have enough rationalists (we had 3/3), but because neither of us knew how to run that kind of a startup. However, Arbital very likely wouldn't have even happened unless I tried. And I'm pretty sure of that because after Eliezer approach me with the idea and I told him "no", six months later nobody was doing it. So I picked it up.
So, I think there are a lot of valuable projects out there that have nobody to do them. And if you look at most failed projects, rationalist or not, they don't fail because there weren't enough helpers, they fail because the person who was supposed to be the leader didn't do a good job.
Over the years I have collected a large pile of various failed or aborted projects in my wake. Things that just didn't work out or didn't go the way I had planned or hoped.
Yeah, welcome to the club. I think most people, even those who have been pretty successful have a long list of failed projects as well. If I understand you correctly, it seems like all of yours failed. I think that's still fine. Most (if not all, depending on how you count) of my projects failed too. I don't think that alone is a good enough reason to stop trying, though obviously you should adjust your strategy. (More on that in the next parts.)
Second, I'll give what I think is a helpful context to approach this from.
To me this question reads loosely as "Hmm, I'm trying things, but they aren't working out. May be the root problem is this one particular underlying assumption." And the umbrella question is "What should I do with my life?" As such, I think any answer has to feel right for you at multiple levels. (Jordan Peterson covers this idea very well.) So I'd caution anyone against hyper-focusing on one assumption / problem when it comes to answering a big question like that.
That said, if you ask yourself "What should I do with my life that feels right for me at all levels?", you often don't get an answer. So it's much better to go try something than not do anything at all. I guess, overall, I'd encourage people to approach this question from a more S1 / holistic feeling rather than S2 / systemic debugging.
I'm not sure this applies in your context, but it might apply to some readers, and it's an important enough point that I'll mention it anyway: when you're young, you're just not that capable. It takes a lot of learning and a lot of mistakes to get to the sweet spot. Some of the most helpful advice I've gotten was: "No, we don't want you here helping on this relatively unimportant project. Go out and learn and become stronger and better." It's a bit hard to swallow because it means going out there and failing. A bunch. But if you're the kind of person who can slog through that and learn, you're also likely the kind of person who can grow a tremendous amount because of that. And that seems a lot more valuable than being stuck helping on relatively unimportant projects, where in the end the project has moved forward, but you have been left behind.
Third, I'll share some of my relevant experience.
I actually have had a very similar question some years ago. "Hmm, Eliezer sure seems to produce a lot of good stuff. I bet if I could help him be even 20% more productive that would be a hell of a lot better than anything I could do." So I did. Around 2015, for about 3-4 months I helped Eliezer with a variety of things. I think it was marginally useful, may be even 20% if I stretch it. But it also became clear to me that it's not what I wanted to do. Even if that was the best I could ever hope to accomplish, that just wasn't a sustainable path. I'm italicizing that point because I think it's a very important realization and even at the risk of the typical mind fallacy, I do think it applies to basically everyone. For a path to be sustainable, you need for it to click on all levels. (Of course I didn't learn this lesson back then, it took another two "failures".)
When I worked on Arbital, I was the CEO. I didn't like most of the CEO-specific tasks like raising money. For my current project, I'm the CTO. And that feels much better for me. One could argue that the CEO is Daimyo and the CTO is the Samurai, but I think that really oversimplifies the relationship and much depends on the context. There're opportunities to lead and to follow, to teach and to listen. I'd say it's much closer to collaboration / interdependence than a hierarchy. The best thing is that I get to do the things I love (coding) and someone else gets to do the things they love (talking to investors).
Fourth, I'll answer the questions directly.
1) I think the question is not extremely useful. It's better to search for the intersection between what you enjoy, what you're good at, and what is valuable to do. Then do it in the best way you can given the situation. If that means leading because nobody is doing it, then do that. If it means joining an existing project, then do that. If it means convincing someone to do the project and then helping them or moving on, then do that.
2) I'd start by talking to people who know you and whom you trust. See in which directions they'd invite you to explore. Check with yourself to see what resonates. Give yourself a month or two to explore, but once exploration starts to slow down, just commit to the best task you've found.
3) I'm guessing that since you've asked these questions, that you are indeed very capable of making "a positive difference." And yes, I think you can improve it. Continue to aim high, learn, strive, accept failures, and do the best you can. I think that's all anyone can ask of you, including yourself. And with a bit of luck, in a decade or two, you're going to be a 1000 year old vampire capable of extraordinary feats.
I used to have video stuttering until I switched from 2.4 GHZ to 5. That helped a lot more than I expected at the time.
I stopped doing it. I also stopped learning things (like math) that required that system. I think at this point I'd probably just get a grid notebook if I started learning math again.
My wife's doctor said chloroquine (a treatment for malaria) was shown to be very promising in treating COVID-19. I guess it's public now as well: https://www.wired.com/story/an-old-malaria-drug-may-fight-covid-19-and-silicon-valleys-into-it/
The whole abstract:
For over a hundred years Finnish sauna has been documented as a physiotherapeutic method in respiratory diseases. Physiological changes of ventilation remain small (about 10%) in the sauna. Heat load, sauna air and sympathetic stimulation generally do not cause problems to the lungs. Electron microscopic studies have not shown irreversible damages to the airway epithelium. Sauna takers should avoid bathing during acute respiratory infections. Sympathomimetic drugs may provoke tachycardia and arrhythmias in the sauna.
So I think they're specifically talking about bathing; while the sauna (dry air) is actually good.
I can't find the entire paper but if someone send me a link, I'll look more into it.
A random thought: If the virus doesn't survive well in high temperatures would going to a sauna help treat it?
Yeah, I think this is spot on. I’ve had this idea for a few years and have been holding a few stocks of each company that I think has a chance at AGI.
Part of my interest behind this question is that recently I've been noticing the subtle (but also not so subtle) ways our choices affect the people and the social fabric around us. Also, listening to Jordan Peterson highlight historical cases where the moral fabric of society fails (e.g. Nazi Germany and Soviet Union). It makes me think there's a lot of under-appreciated value in being a morally upright person. Some of the value is in directly impacting the people around you, but the majority is in signaling that "this is the kind of society we live in."
Love your posts! I’d recommend publishing them at a slower pace so people get a chance to see and read them. It’s also a bit confusing on LW when several posts are published on the same day, and you can’t tell which one is first in the series (just from the front page).
One source I looked at was: https://parentsguidecordblood.org/en
Some highlights:
- There are three public cord blood banks in CA, but non in the bay area. So donating might be hard if that's where you're delivering.
- The expected cost of cryo preserving is a bit under $2,000 upfront and $150 / yr.
- It costs about $30,000 to obtain a cord blood collection from a public bank, and that cost is charged to the patient’s health insurance.
- Looks like the stem cells could be used by the family members too. (Assuming it's a match, I'm guessing?). But actually the child might not be able to use them. (Is it only while they are a child?)
- The main advantage of cord blood is that it does not have to be exactly matched to the patient like transplants from an adult donor.
- In the United States, about 80% of cord blood donations are discarded.
- In the United States, the net probability that a child will need any type of stem cell transplant by age 20 is 3 in 5,000 or .06%.
Can you expand on “ AI/ML capabilities research in finance” or shoot me a PM?
I’m definitely not an expert, but it does sound like a few people moved the goalposts. It sounds very similar to the "AI can't play chess" argument switching to "well, it's not really playing chess, it's just following rules" and then to *quietly mumbling something about it not having the True representation of chess*.
So, I agree with the core point. GPT-2 is writing! And it's writing pretty damn well. Whatever is left has more to do with the general cognition skill than writing.
Great question! I will just say that I had such an experience, but don’t know how to share it in a way that feels adequate for me.
I think you can lead a healthy and happy life playing tennis most of your time, while avoiding the (incorrect and unhealthy, in my opinion) belief that “getting a life is finding a singular passion and pursuing it to the end”. It seems to me like a motivated reasoning used to defend what the person is doing. However the rest of the post is closer to the real truth: “I enjoy playing and teaching tennis and I find meaning in it.”
Also I think the singular/plural purpose divide is pretty smooth with plenty of people on both sides. I’ve met lots of people who say that to live your life you basically need to specialize like crazy.
My invitation to the author would be to write a similar essay to steel man the other side.
Wow, this seems huge to me! Great find! I’m not in the business of constructing tests, but if I ever do any home schooling with my children, I’ll definitely try this method.
Your decision / advice is rooted in anxiety / fear. This is much more about you than about the thing you’re talking about.
This post convinced me to watch the movies. Was not disappointed: great world building indeed.
This was back in 2013. At that point I have been developing games (as a hobby and then professionally) for 13 years. During my free time in high school and college, that's basically all I did. And then right out of college I got a job in the game industry. I also developed and published my own game. In 2011 I moved to the bay area, joined a game startup, and it was acquired. (By Zynga, so not that exciting, but overall, I'd say things were going well.)
Around that time, I realized that there was no way making games would help with x-risk. So I left the industry. This involved letting go of the deepest passion and the most developed skill I've had at the time. It involved changing which circles I networked in. And it included abandoning all the knowledge I accumulated of the programming libraries and framework, game design, game lore, and all the half-baked game ideas I had and was hoping to develop some day.
I still stand by that decision. But I've also found a way to incorporate a bit of that old self into my present life. For example, sometimes I design a board game. Or write down and explore game ideas until I can "see" how it would be developed. Or dream about having enough money to just hire an entire studio and have them develop it.
Yes.
Found this video and it reminded me of this story: https://www.youtube.com/watch?v=gNVqRC4mcSI
Tangentially related: https://www.lesswrong.com/posts/RasFpce3fNZ8xp2T4/open-and-welcome-thread-november-2019?commentId=trSeGziLqLZ5jiAgv
I have read the entire "My big TOE" book/trilogy by Thomas Campbell. If at least three people promise to read my summary of it, I'll write it up. Let me know how long you want the summary to be.
Super quick summary of the book: Thomas Campbell seems like a decently smart guy; he has a PhD in physics and works for the government. He claims his mind/consciousness can exist in and travel between multiple realities. In this book he explains his version of how the universe works, how it got started, why it's here, and why we are here.
Personally, I found it mildly interesting. There are a lot of big claims, but it's very hard to verify them. Tom is pretty upfront about it: his goal is to lay out a conceptual framework so that everyone can develop their own TOE (Theory of Everything). He doesn't ask the reader to trust him or his version. At the same time, if you can't trust it... it's kind of like reading sci-fi.
I'm also happy to answer any questions about the book / author.
Kaj, first off, huge thank you for writing up all these posts on meditation. I don't have time to read a lot of these books, but it's good literature, so it's perfect to get the summary from someone I can trust.
"Introspective awareness" sounds like the right object. Or, more specifically, it definitely feels like it's describing my own experience. And my own, homegrown hypothesis was something like: consciousness is like an echo or picture-in-picture. We can get glimpses of "ourselves" because we can look at / load partial concepts of ourselves into the working memory.
Introspective awareness is its own type of mental object...
Wow, yes, yes, yes! My original fear with starting meditation was that it would remove certain experiences from my life. It felt like I'd lose something. And in the 7+ years I've done it, that just hasn't happened. And now I have the precise language to describe it. That is exactly what happens for me: like, I can still see the suffering and all that, but it's like it's wrapped in a bubble, which still allows me to see the emotion/sensation exactly like I would before, but it loses the ability to instantly propagate its agenda to the rest of my system.
This kind of a process also teaches you to pay attention to patterns of cause and effect in your mind.
Yeah, I have seen this emphasized in various teachings, but I never actually practiced it. Probably some low hanging fruit here for me. I'll try it out.
“first you resolve a lot of issues, but then you can get the ability to push down the rest” dynamic
One, I think the foremost goal for any type of meditation should be to learn to see what is there. (Just like with rationality.) So there's no pushing or pulling or trying to change anything. And sometimes with bad feelings or headaches, I've noticed that simply paying attention to it (instead of flinching away) resolves it. It doesn't make it "go away" or "block it", it actually just unties the knot, so to speak.
Two, I think there are two orthogonal skills that get bucketed into "spiritual development." One skill is insight (see MCTB for a good definition), but it's basically the ability to see what's going on in your mind. And that skill will eventually take you to enlightenment. The other skill is morality. And while some practices do empathize morality practice as well, I think in the classical western tradition (MCTB is a good example of this) we found paths to get to enlightenment without all other "unnecessary" stuff... like morality training. So you end up with people who are technically enlightened, but it doesn't automatically make them good human beings. (I think this also resolves the confusion in one of the recent SSC posts.)
Use the power wisely. :)
I think some big life events fall into this category: being married, having a kid, having a close loved one pass away, dying yourself.