Posts
Comments
Not sure what's going on, but gpt-4o keeps using its search tool when it shouldn't and telling me about either the weather, or sonic the hedgehog. I couldn't find anything about this online. Are funny things like this happening to anyone else? I checked both my custom instructions and the memory items and nothing there mentions either of these.
Small groups of mammals can already cooperate with each other (wolf's, lions, monkeys etc.). In mammals, I'd guess having a queen gives a bottleneck in how fast there can be off-spring. Also if there are large returns to division of labor in child-rearing, large animals are smart enough that both parents can do this together, while in wasps the males just die (why actually?). So wasps get higher marginal returns when evolving the first steps towards being eusocial. Also smaller animals have more diverse environments and need fewer years to "locked in" eusociality and workers get born without being fertile (eusocial groups where workers are still fertile are really unstable so prone to evolve away from eusociality again when circumstances aren't in favor anymore). Also fathers can't be as sure of their children and the other way around leading to less cooperation if new males join in, which termites overcome by having king and queen, ants just have a queen that stores her sperm, while naked mole rats are just fine with incest?
… that wasn’t enough to learn the pattern, though. Shortly out of college, reality was still hitting me over the head; that time the big idea was an efficient implementation of universal competitively-optimal portfolios. I lost a couple thousand dollars on wildly over-leveraged forex positions.
I am curious what that idea was and where it went wrong.
In that case also consider installing PowerToys and pressing Alt+Space to open applications or files (to avoid unhelpful internet searches etc.).
I like this sequence and am aware it is not finished yet. Here's my I am understanding so far. After reading the sequence, I think I can predict your response to the first 5 conundrums, so my previous confusion there (why cluster rather than factor) seems resolved. But I think I still disagree with the later examples that I was confused with before reading your sequence. One example of conundrums where I think I get what your reply would be:
-
"Why isn't factor analysis considered the main research tool?"
Factor analysis doesn't capture the main bottlenecks (people being depressed for different reasons, people are successful for different reasons etc.)
For others, I don't see how they connect well. My replies would be:
-
"What is gifted child syndrome/twice-exceptionals?"
I don't know why you focus on this one? My impression why there's a focus on this group is because helping them might be worth the investment? Or because the people writing and consuming that theorizing tend to be higher iq. Also, maybe "that phenomenon where desirable trait X and Y tend to be anticorrelated, because the others tend to not want to hang out with you as much, or you don't want to hang out with them" (writing and math being anticorrelated in the average US college)? I don't see the relation in the log-normals, other that maybe in your thinking you might want to single out that group, because it might have bottlenecks that are different?
-
"Why would progressivism have paradoxical effects on diversity?"
I am confused? I can see you making the argument that the diversity angle might sometimes be the correct one if it is the bottleneck for a person (black person being arrested for doing drugs in the US? While less of a bottleneck for a lot of other minorities?)
-
"What's wrong with symptom treatment?"
Do you think people's intuition here is correctly adjusting for something like the epsilon fallacy? Or to quickly jumping to simplistic conclusions like in this college cost post you link (in a different context), where someone might (in my view accidentally) see the increasing number of small courses as a cause rather than a symptom?
-
"What value does qualitative research provide?"
I am reminded of
and
I feel like so far this sequence has mostly told me what tools not to use and, in practice, I cannot think of a case where reading this sequence has helped pick a better tool, but I was already pretty fond of log normals.
So I should look out for that, e.g. by doing some manual fermi estimates or other direct checking about ABC or by investigating the strength of the steelman of reaction XYZ, or by keeping an eye out for people systematically reacting with XYZ without good foundation so I can notice this,
Accusing people in my head of not being numerate enough when this happens has helped, because then I don't want to be a hypocrite. GPT4o or o1 are good at fermi estimates, making this even easier.
I noticed the tag posts imported from Arbital that haven't been edited on LW yet can't be found when searching those tags from the "Add Tags" button above posts. Adding ineffective edits like spaces at the end of a paragraph seems to fix that problem.
I noticed the tag posts imported from Arbital that haven't been edited on LW yet can't be found when searching those tags from the "Add Tags" button above posts. Adding ineffective edits like spaces at the end of a paragraph seems to fix that problem.
I didn't downvote, but my impression is the post seems to hand-wave away a lot of problems and gives the impression you haven't actually thought clearly and in detail about whether the ideas you propose here are feasible.
Some people have been thinking for quite some time now that an AI that wants to be changed would be great, but that it's not that easy to create one, so how is your proposal different? Maybe checkout the corrigibility tag. Figuring out which desiderata are actually feasible to implement and how is the hard part. Same goes for your Matroshka bunkers. What useful work are you getting out of your 100% safe Matroshka bunkers? After you thought about that for 5 minutes+, maybe checkout the AI boxing tag and the AI oracle tag. Maybe there is something to the reversibility idea ¯\_(ツ)_/¯.
Also using so many tags gives a bad impression ("AI Timelines"? "Tiling Agents"? "Infinities in Ethics"?). Read the description of the tags.
Finding some some friend (or language model?) to play Zendo (the science game) with makes this really intuitive on a gut level. Guessing a rule based on whether a sequence of 3 integers is either accepted or rejected works pretty well via text.
NOTE: I posted this to LW and I'm new here so I don't totally know the cross-posting policies. Hope it's alright that I posted here too!
It seems you posted on LW twice instead or in addition to cross-posting to the EA forum.
I just tried this with o3-mini-high and o3-mini. o3-mini-high identified and prevented the fork correctly, while o3-mini did not even correctly identify it lost.
I can only see the image of the 5-d random walk. The other images aren't rendering.
I was already sold on singularity. For what it's worth I found the post and comments very helpful for why you would want to take the sun apart in the first place and why it would be feasible and desirable for superintelligent and non-superintelligent civilization (Turning the sun into a smaller sun that doesn't explode seems nicer than having it explode. Fusion gives off way more energy than lifting the material. Gravity is the weakest of the 4 forces after all. In a superintelligent civilization with reversible computers, not taking apart the sun will make readily available mass a taut constraint).
One thing I am pretty confident about is that methylation patterns are downstream, not upstream. Methyl group turnover time is far too fast to be a plausible root cause of aging. (In principle, there could be some special methyl groups which turn over slowly, but I would find that very surprising.)
My possibly wrong understanding here is that there are histone modifications and other proteins (like CTCF) that make methylation patterns way more stable? Which leads to some methylation patterns like imprinting for genes like IGF2 to be stable in most tissues over ~decades. Nevertheless, loss of imprinting and epigenetic marks still doesn't necessarily seem like the most likely root cause of aging to me.
This argument against subagents is important and made me genuinely less confused. I love the concrete pizza example and the visual of both agent's utility in this post. Those lead me to actually remember the technical argument when it came up in conversation.
I found Steven Byrnes valence concept really useful for my own thinking about psychology more broadly and concretely when reading text messages from my contextualizing friend (in that when a message was ambiguous, guessing the correct interpretation based on valence worked surprisingly well for me).
I ended up dodging the bullet of loosing money here, because I was a bit worried that Nate Silvers model might have been behind, because the last poll then was on the 23rd. I was also too busy with other important work to resolve my confusions before the election. My current two best guesses are:
- The French whale did not have an edge,
- The neighbour polling method is a just-so story to spread confusion, but he actually did have an edge
- I don't understand correctly how this neighbour polling method is supposed to work.
In any case, if Polymarket is still legal in 4 years I expect the prediction market on the election to be efficient relative to me and I will not bet on it.
I had a discussion with @Towards_Keeperhood what we would expect in the world where orcas either are or aren't more intellectually capable than humans if trained. Main pieces I remember were: Orcas already dominating the planet (like humans do), large sea creatures going extinct due to orcas (similar to how humans drove several species extinct (Megalodon? Probably extinct for different reasons, weak evidence against? Most other large whales are still around)). I argued that @Towards_Keeperhood was also underestimating the intricacies that hunter-gatherers are capable of, and gave the book review for the secret of our success as an example. I think @Towards_Keeperhood did update in that direction after reading that post. I also reread that post and funnily enough stumbled over some evidence that orcas might have fallen into a similar "culture attractor" for intelligence, like humans:
Learn from old people. Humans are almost unique in having menopause; most animals keep reproducing until they die in late middle-age. Why does evolution want humans to stick around without reproducing?
Because old people have already learned the local culture and can teach it to others. Heinrich asks us to throw out any personal experience we have of elders; we live in a rapidly-changing world where an old person is probably “behind the times”. But for most of history, change happened glacially slowly, and old people would have spent their entire lives accumulating relevant knowledge. Imagine a world where when a Silicon Valley programmer can’t figure out how to make his code run, he calls up his grandfather, who spent fifty years coding apps for Google and knows every programming language inside and out.
Quick google search revealed Orcas have menopause too! While chimpanzees don't! I would not have predicted that.
Typo in the linked document:
There is no one is coming to save us.
Can someone who is already trading on Polymarket or is planning to do so soon tell me if there are any hidden fees (or ways my money might be locked up for longer than I expect) if I trade on Polymarket? Four years ago I got hit by enormous ether gas fees on Augur, which still made my bet positive EV, but only barely so (I had to wait quite a while for the gas cost to go that low and was loosing out on investing the money and my attention). I plan to bet ~$3K-$7K and think Kamala Harris has a 45% chance of winning. Is that enough for all the transaction costs to vanish?
One confounder: depression/mania. Recently (the last ~two weeks) I have been having bad sleep (waking up 3-7 am and not feeling sleepy anymore (usually I sleep from midnight to 9). My current best guess is that the problem is that my life has been going too well recently, leading to a self-sustaining equilibrium where I have little sleep and mania. Reduced my medication today (~55mg instead of 70mg) which seems to have helped with the mania. I had another day with slight mania 1 month ago when sleeping little in order to travel to a conference, so in the future I'll probably reduce my medication dose on such days. Took a friend describing his symptoms on too much medication for me to realize what is going on.
I am also interested in finding a space to explore ideas which are not well-formed. It isn’t clear to me that this is intended to be such a space. This may simply be due to my ignorance of the mechanics around here.
For not well-formed ideas, you can write a Quick Take (can be found by clicking on your profile name in the top right corner) or starting a dialogue if you want to develop the idea together with someone (can be found in the same corner).
I feel like there should exist a more advanced sequence that explains problems with filtered evidence leading to “confirmation bias”. I think the Luna sequence is already a great step in the right direction. I do feel like there is a lack of the equivalent non-fiction version, that just plainly lays out the issue. Maybe what I am envisioning is just a version of What evidence filtered evidence with more examples of how to practice this skill (applied to search engines, language models, someone’s own thought process, information actively hidden from you, rationality in groups etc.).
adult augmentation 2-3std for the average person seems plausible, but for the few +6std people on earth it might just give +0.2std or +0.3std, which tbc I think is incredibly worthwhile.
Such high diminishing returns in g based on genes seems quite implausible to me, but would be happy if you can point to evidence to the contrary. If it works well for people with average Intelligence, I'd expect it to work at most half as well with +6sd.
I am a bit confused why some of these theories would be so hard to test? It seems like some core pathways that seem like they wouldn't be reversible even in naive stem cells under any circumstances (like transposons copying themselves successfully), could possibly be tested by checking if clones derived from older cells age faster or something along those lines? The same goes for children from older parents? (Not sure to which extent that test would be made harder by all the mechanisms keeping the germ line immortal)
I don't know where anger fits into this. Also I should look at how these behaviors manifest in other animals.
Hypothesis based on the fact that status is a strong drive and people who are on the outer ends of that spectrum get classified as having a "personality disorder" and are going to be very resistant to therapy:
- weak-status-fear==psychopathy: psychopathy is caused by the loop leading to fear of loosing status, being less strong than average or possibly broken. (psychopathy is Probably on a spectrum. I don't see a reason why little of this feeling would be less optimal than none.)
- strong-status-fear==(?histrionic personality disorder)
- weak-status-seeking-loop==(?schizoid personality disorder)
- strong-status-seeking-loop==(?narcissism)
Was thinking about Steven Byrnes agenda to figure out social drives and what makes a psychopath a psychopath. One clearly existing social drive that seemed to be a thing was “status-seeking” and “status-fear” (fear of loosing status). Both of these could themselves be made of several drives? The idea that status-seeking and status-fear are different came to me when trying to think of the simplest hypothesis explaining psychopathy and from introspecting that both of these feelings feel very different to me and distinct from other fears. These two could be made more mostly separate loops, but I can't complicate my fake framework even more just yet.
If someone is interested, I'd write a post how to stress test this fake-framework and what I'd expect in the world where it is true or isn't (Most interesting would be social drives that are distinct from the above? Or maybe they use some of the same sub-circuitry? Like jealousy seems obviously like it would fit under strong status fear, so histrionic personality would go with being more jealous)
Since I am already on the fancy note-taking train, I'd find examples of your actual note files way more interesting.
On my phone, rotating the screen by 180° quickly reverses the direction and then I rotate it back slowly.
I think from the perspective of a radical probabilist, it is very natural to not only have a word of where your current point estimate is at, but also have some tagging for the words indicating how much computation went into it or if this estimate already tries to take the listeners model into account also?
I misread your whole post by thinking your title implied "your post would question whether the entropy increased=> the post argues it decreases" and then I was reading sloppily and didn't notice my error.
Also you should halt and reevaluate your intuitions if they lead you to believe there is a perpetual motion machine.
Photosynthesis? Most of the carbon is bound from CO2 by using sun exergy.
Cool post. I agree with the many-shot part in principle. It strikes me that in a few years (hopefully not months?), this will look naive in a similar way that all the thinking on ways a well boxed AI might be controlled look naive now. If I understand correctly, these kinds of simulations would require a certain level of slowing down and doing things that are slightly inconvenient once you hit a certain capability level. I don't trust labs like OpenAI, Deepmind, (Anthropic maybe?) to execute such a strategy well.
If legibility of expertise is a bottleneck to progress and adequacy of civilization, it seems like creating better benchmarks for knowledge and expertise for humans might be a valuable public good. While that seems difficult for aesthetics, it seems easier for engineering? I'd rather listen to a physics PhD, who gets Thinking Physics questions right (with good calibration), years into their professional career, than one who doesn't.
One way to do that is to force experts to make forecasts, but this takes a lot of time to hash out and even more time to resolve.
One idea I just had related to this: the same way we use datasets like MMLU and MMMU, etc. to evaluate language models, we use a small dataset like this and then experts are allowed to take the test and performance on the test is always public (and then you make a new test every month or year).
Maybe you also get some participants to do these questions in a quiz show format and put it on YouTube, so the test becomes more popular? I would watch that.
The disadvantage of this method compared to tests people prepare for in academia would be that the data would be quite noisy. On the other hand, this measure could be more robust to goodharting and fraud (although of course this would become a harder problem once someone actually cared about that test). This process would inevitably miss genius hedgehog's of course, but maybe not their ideas, if the generalists can properly evaluate them.
There are also some obvious issues in choosing what kinds of questions one uses as representative.
It not being linked on Twitter and Facebook seems more like a feature than a bug, given that when I asked Gwern why a page like this doesn't already exist, he wrote me he doesn't want people to mock it.
> I really like the importance Tags, but what I would really like is a page
> where I can just go through all the posts ordered by importance. I just
> stumbled over another importance 9 post (iron rules) when I thought I had
> read all of them. Clicking on the importance tag, just leads to a page
> explaining the importance tag.
Yeah, that is a mix of 'too hard to fix' and 'I'm not sure I want to
fix it'. (I don't know how Hakyll works well enough to do it
'normally', and while I think I can just treat it as a tag-directory,
like 'meta/importance/1', 'meta/importance/2' etc, that's a little
awkward.) Do I *want* people to be able to go through a list of
articles sorted by importance and be able to easily mock it - avoiding
any actual substantive critique?
I went through Gwern’s posts and collected all the posts with importance 8 and higher as of 2024-09-04 in case someone else was searching for something like this.
10
- How Often Does Correlation=Causality?
- Why Correlation Usually ≠ Causation
- The Scaling Hypothesis
- Complexity no Bar to AI
- The Algernon Argument
- Embryo Selection For Intelligence
- Life Extension Cost-Benefits
- Colder Wars
- Metamagical Themas: Sanity and Survival
- The Existential Risk of Math Errors
9
- Iodine and Adult IQ meta-analysis
- Lithium in ground-water and well-being
- Melatonin
- Modafinil
- Modafinil community survey
- Nicotine
- Why Tool AIs Want to Be Agent AIs
- Machine Learning Scaling
- The Hyperbolic Time Chamber & Brain Emulation
- The Iron Law Of Evaluation And Other Metallic Rules
- Spaced Repetition for Efficient Learning
- Are Sunk Costs Fallacies?
- Plastination versus Cryonics
- Silk Road 1: Theory & Practice
- Darknet Market Archives (2013–2015)
- The Melancholy of Subculture Society
8
- The Replication Crisis: Flaws in Mainstream Science
- How Should We Critique Research?
- The Narrowing Circle
- Zeo sleep self-experiments
- When Should I Check The Mail?
- Candy Japan’s new box A/B test
- The Kelly Coin-Flipping Game: Exact Solutions
- In Defense of Inclusionism
- Bitcoin Is Worse Is Better
- Time-lock encryption
- Easy Cryptographic Timestamping of Files
- GPT-3 Creative Fiction
- RNN Metadata for Mimicking Author Style
- Terrorism Is Not About Terror
- Terrorism Is Not Effective
- Archiving URLs
- Conscientiousness & Online Education
- Radiance: A Novel
The recent post on reliability and automation reminded me that my "textexpansion" tool Espanso is not reliable enough on Linux (Ubuntu, Gnome, X11). Anyone here using reliable alternatives?
I've been using Espanso for a while now, but its text expansions miss characters too often, which is worse than useless. I fiddled with Espanso's settings just now and set the backend to Clipboard, which seems to help with that, but it still has bugs like the special characters remaining ("@my_email_shorthand" -> "@myemail@gmail.com").
In particular, I think you might need to catch many escape attempts before you can make a strong case for shutting down. (For concreteness, I mostly imagine situations where we need to catch the model trying to escape 30 times.)
So instead of leaving the race once the models start scheming against you, you keep going to gather more instances of scheming until you can finally convince people? As an outside reader of that story I'd just be screaming at the protagonists that clearly everyone can see where this is going where scheming attempt number 11 is just good enough to be successful. And in the worlds where we catch them 30 times successfully it feels like people would argue: this is clear evidence that the models aren't "actually dangerous" yet, so let's keep scaling "responsibly".
There is probably a lot of variation between people regarding that. In my family meds across the board improved people's sleep (by making people less sleepy during the day, so more active and less naps). When I reduced my medication from 70mg to 50mg for a month to test whether I still needed the full dose, the thing that was annoying the most was my sleep (waking up at night and not falling asleep again increased. Falling asleep initially was maybe slightly easier). Taking it too late in the afternoon is really bad for my sleep, though.
Things I learned that surprised me from a deep dive into how the medication I've been taking for years (Vyvanse) actually gets metabolized:
- It says in the instructions that it works for 13 hours, and my psychiatrist informed me that it has a slow onset of about an hour. What that actually means is that after ~1h you reach 1/2 the peak concentration and after 13 hours you are at 1/2 the peak concentration again, because the half-life is 12h (and someone decided at some point 1/2 is where we decide the exponential starts and ends?). Importantly, this means 1/4 of the medication is still present the next day!
Here is some real data, which fit the simple exponential decay rather well (It's from children though, which metabolize dextroamphetamine faster, which is why the half-life is only ~10h) 
- If you eat ~1-3 grams of baking soda, you can make the amount of medication you lose through urine (usually ~50%) go to 0[1] (don't do this! Your body probably keeps its urine pH at the level it does for a reason! You could get kidney stones).
- I thought the opposite effect (acidic urine gets rid of the medication quickly) explained why my ADHD psychologist had told me that the medication works less well combined with citric fruit, but no! Citric fruit actually increase your urine pH (or mostly don't affect it much)? Probably because of the citric acid cycle which means there's more acid leaving as co2 through your lungs? (I have this from gpt4 and a rough gloss over details checked out when checking Wikipedia, but this could be wrong, I barely remember my chemistry from school)
- Instead, Grapefruit has some ingredients known to inhibit enzymes for several drugs, including dextroamphetamine (I don't understand if this inhibitory effect is actually large enough to be a problem yet though)
- This brings me to another observation: apparently each of these enzymes is used in >10-20% of drugs: (CYP3A4/5, CYP2D6, CYP2C9). Wow! Seems worth learning more about them! CYP2D6 gets actually used twice in the metabolism of dextroamphetamine, once for producing and once for degrading an active metabolite.
Currently still learning more about basics about neurotransmitters from a textbook, and I might write another update once/if at the point where I feel comfortable writing about the effects of dextroamphetamine on signal transmission.
Urinary excretion of methylamphetamine in man (scihub is your friend) ↩︎
Looking forward to the rest of the sequence! On my current model, I think I agree with ~50% of the "scientism" replies (roughly I agree with those relating to thinking of things as binary vs. continuous, while I disagree with the outlier/heavy-tailed replies), so I'll see if you can change my mind.
The technical background is important, but in a somewhat different way than I'd thought when I wrote it. When I was writing it, I was hoping to help transmit my model of how things work so that people could use it to make their own decisions. I still think it's good to try to do this, however imperfectly it might happen in practice. But I think the main reason it is important is because people want to know where I'm coming from, what kinds of things I considered, and how deeply I have investigated the matter.
Yes! I think it is beneficial and important that someone who has a lot of knowledge about this transmits their model on the internet. Maybe my Google foo is bad, but I usually have a hard time finding articles like this when there doesn't happen to be one on Lesswrong (only can think of this counterexample I remember finding reasonably quickly).
Raising children better doesn't scale well. Neither in how much ooomph you get out of it per person, nor in how many people you can reach with this special treatment.
What (human or not) phenomena do you think are well explained by this model? I tried to think of any for 5 minutes and the best I came up with was the strong egalitarianism among hunter gatherers. I don't actually know that much about hunter gatherers though. In the modern world something where "high IQ" people are doing worse is sex, but it doesn't seem to fit your model.
So on the -meta-level you need to correct weakly in the other direction again.
I used Alex Turners entire shortform for my prompt as context for gpt-4 which worked well enough to make the task difficult for me but maybe I just suck at this task.
By the way, if you want to donate to this but thought, like me, that you need to be an “accredited investor” to fund Manifund projects, that only applies to their impact certificate projects, not this one.