I learned the hard way that epigenetics is too new a field to get a good understanding by reading the textbook, because being 13 years out of date is in fact making a big difference.
Skill-issue. There is a free introductory e-textbook on epigenetics on Amazon from 2021 that I somehow completely missed.
I for one know that I interact very differently with children with different personalities! (Or, for that matter, with adults with different personalities.) One classic example of this is that children who are naturally compliant and “easy” are disciplined/punished less, because there’s much less of a need to do so.
Yeah equating parenting with shared-environment can lead to confusion, but your example doesn’t necessarily end up in the non-shared part I think. If the personality of the child was mostly downstream of the genes, then I think that would still end up in shared environment and would not be a problem (You treat both twins the same, because they have about the same temperament). If some parents treat twins differently because of “random” things like which twin left the womb first and is considered firstborn, which baby hit their head, inherent contingency in personality etc., then yeah, I think that would end up in a non-shared environment if you do twin experiments.
One reason why I find Lesswrong valuable is that it serves as a sort of “wisdom feed” for myself, where I got exposed to a lot of great writing. Especially writing that ambitiously attempts to build long-lasting gears. Sadly, for most of the good writers on Lesswrong, I have already read or at least skimmed all of their posts. I wonder, though, to which extent I am missing out on great content like that on the wider internet. There are textbooks, of course, but then there’s also all this knowledge that is usually left out of textbooks. For myself, it probably makes sense to just curate my own RSS feed and ask language models for “The Matt Levine in Domain X” etc. But it also feels like it should be possible to create a feed for this type of category with language models? Sort of like the opposite of news minimalist. Gears maximalist?
As for successionists, and honestly utilitarians in general, but only when they apply it to situations which result in their own deaths, I cannot understand this viewpoint.
Quite a few people would die to save their children. I actually have never met someone who told me they can’t relate to any thought experiment where they would die for someone else. I would be curious if you can relate to any of the thought experiments in the fake selfishness post. Presumably there are quite a few people who would not sacrifice their entire future for anything, but you can also just sacrifice it partially by taking risks like driving a car.
The neuroscience/psychology rather than ml side of the alignment problem seems quite neglected (because it harder on the one hand, but it’s easier to not work on something capabilities related if you just don’t focus on the cortex). There’s reverse engineering human social instincts. In principle would benefit from more high quality experiments in mice, but those are expensive.
Also, it seems like there is not much of that in the field of alignment. I want there to be more work on unifying (previously frontier) alignment research and more effort to construct paradigms in this preparadigmatic field (but maybe I just haven’t looked hard enough)
I am surprised regarding the lack of distillation claim. I’d naively expected that to be more neglected in physics compared to alignment. Is there something in particular that you think could be more distilled?
Regarding research that tries to come up with new paradigms, here are a few reasons why you might not be observing that much: I guess that is less funded by the big labs and is spread across all kinds of orgs or individuals. Maybe check MIRI, PIBBBS, ARC (theoretical research), Conjecture check who went to ILIAD. More of these researchers didn’t publish all their research compared to AI safety researchers at AGI labs, so you would not have been aware it was going on? Some are also actively avoiding researching things that could be easily applied and tested, because of capability externalities (I think Vanessa Kosoy mentions this somewhere in the YouTube videos on Infrabayesianism).
If valence is “extremized” (i.e., almost every thought is either very positive valence, or very negative valence, but rarely anywhere in between), it should lead to a cluster of symptoms suspiciously close to Narcissistic Personality Disorder (NPD)
Also sounds like it could lead to OCPD? Just less focus on self and validation and more focus on the environment and correctness or something along those lines. Mostly spitballing after I listened to this podcast on OCPD. Best prototypical example is probably Sheldon from the big bang theory, which I always thought of as “popular conception of what autism looks like” and it hadn’t occurred to me there were actually people in this category.
I recently told my dad that I wasn’t trying to save for retirement. This horrified him far more than when I had previously told him that I didn’t expect anyone to survive the next couple of decades. The contrast was funny.
I had a similar experience where my dad seemed unbothered by me saying AI might take over the world and some other day I mentioned in passing that I don’t know in how many years we have AI that is a better software engineer than humans, but 5-10 years doesn’t sound strictly impossible. My father being a software engineer found that claim more interesting (He was visibly upset about his job security). I noticed I’ve kinda downplayed the retirement thing to my parents, because implicitly I noticed at that point they might call me insane, but explicitly thinking about it, it might be more effective to communicate what is at stake.
Make sense?
Yes, this part was obvious! What I meant with those bit-flips was the exponentially small probability of a sudden discrete transition. Why could accidental transitions like this not accumulate? Because they are selected against fast enough? I wouldn’t expect those epigenetic marks to have enough redundancy to reliably last to the end of an organisms’ lifetime, because methylated cytosine is prone to deaminate (which is why CG is the least frequent 2-mer at a ~1% frequency, rather than the ~6.25% you’d expect on baseline). I am confused how the equilibrium works here, but it seems like mutational load could explain why organisms who rely on methylating cytosine have less CG’s than would be useful to maintain epigenetic information. Things would be different in organisms that don’t rely so heavily on methylating cytosine.
The mistake: methyl groups usually do not stick around long-term; they turn over regularly. Here are two studies which measured the turnover. Turnover timescale varies by location on the DNA strand, but turnover every few days is typical.
I don’t know if someone else was making this particular mistake. I certainly find it quite tricky to describe these things with language unless I am extremely careful. I am still confused, but I do find your suggested pathway via transposons more plausible, and the story for why transposons fits into my head. I don’t think a correct story for why not epigenetic “bit-flips” currently fits properly into my head, and that correct story would have to be created in order to convince other people. Ideally, some testable predictions[1].
But if the methyl groups are instead part of a dynamic equilibrium, especially with high redundancy (and therefore stability), then that’s a whole different situation.
I don’t understand why this has to be such a dichotomy? What I hear you say is that these dynamic equilibria are too stable to be a root cause of aging. I think this is true in the case of X-inactivation[2], which is redundant across most of the X-chromosome. But as far as I can tell, the number of redundant parts that make up one “unit” that can bit-flip lies on a spectrum. Example: one methylated C is probably too unstable to encode something. But what if I have a local cluster of five of them where each one reinforces each other (because enzymes that change the methylation often check the methylation status of close-by C’s)? That’s the point of CpG islands, right? So some of these CpG islands are probably more stable than others, and some could be just stable/unstable enough to be a root cause of aging, no?
For example, my guess would be that the transposon story and the accumulating epigenetic chaos story make different predictions about which methylations would serve as an aging clock and how well. They would also predict different things if we compare aging in different species, because transposons and mitochondria are (quite?) universal in eukaryotes, while epigenetics between eukaryotes is quite different.
My understanding is X-inactivation almost never bit-flips in adults. If it were unstable on the timescale of years, you’d sometimes see one white hair in a spot where a cat has black hair, for example. Maybe it is more common in cells with faster turnover?
Yep! When I was in high school I was self-learning physics and was learning calculus via Khan academy, which made me learn that “sometimes”, “somehow”, you can for example solve a differential equation by getting all the dx and x’s on one side and dy’s and y’s on the other. I was always expecting when I would study this stuff more rigorously in university, this would be explained at some point. To my great disappointment, the mandatory class on this subject for my CS degree “real analysis”, did not in fact clear up why this works in the slightest!? I don’t know if every real analysis class is this bad (My linear algebra courses were excellent in comparison), but mine was mostly focusing on making students adopt “rigor”[1], presenting definition after definition and theorem after theorem (for short theorems proofs were included, but with zero sub-text what the motivation or idea of the proof was. Complicated proofs were just skipped). Starting studying during Corona and not being able to ask professors questions like that directly certainly didn’t help. Somehow I forgot, though, that I never got these questions answered that I had hoped to get an answer to. Also, the quotient in df/dx still confuses me, and it might be due to my confusion about division? Looking at the wikipedia article you link, maybe I should take a look at differential forms?
As my secondary subject for my bachelor, I chose physics, but that mostly involved applying Euler-Lagrange to particular systems. I remember becoming really confused why this can work at this point and noticing that I can’t fit the explanation in my head (I am in fact still confused and would love if you can point me to useful resources!). I remember thinking why can you differentiate with respect to speed and treat acceleration etc. as not dependent on speed, weird. Clearly, when framing them all as a function of time, it made no sense (and still doesn’t). I was looking it up on Wikipedia and apparently “Differential Geometry” was the answer to my questions, so I got 2–3 textbooks on Differential Geometry (for physics), but just skimming, I couldn’t find something that made this click.
Probably I should have looked more aggressively for people to tutor me. Now with language models it feels worth taking another stab at this though. For example when I told GPT4-o the rough areas I am confused about (link to full chat): there are these people talking about scaling laws in the context of “scaling laws” in engineering, and it seems like I might be missing some prerequisites, because I’ve bounced off this topic a few times even though this seems extremely interesting. Which led it to introduce dimensionless analysis to me, and then I asked more questions and ultimately gave me some textbook recommendations: “Street fighting mathematics”, I had already heard praise for in “Biology by the numbers”, so then I checked it out and discovered this gold.
Another question I am still confused by: how does your choice of units affect what types of dimensionless quantities you discover? Why do we have Ampere as a fundamental unit instead of just M,L,T? What do I lose? What do I lose if I reduce the number of dimensions even further? Are there other units that would be worth adding under some circumstances? What makes this non-arbitrary? Why is Temperature a different unit from Energy?
Also, I notice all the formalizations of dimensional analysis that Terence Tao mentions here strike me as ugly? Is that me being stupid, thinking that there must exist something nicer?
Which I found mostly annoying and trivial, but made me glade I had spent ~60h before university just practicing proving things. How to proove it is a great book!
Thinking about this thread on “How could I have thought of that faster?”. In practice, I noticed the phrasing doesn’t work well for me and I prefer “How could I have seen this faster”? I especially like this framing for when I have identified a “blindspot”, “developmental milestone” or noticing someone’s wizard power is clearly hinting at a powerful learnable skill or concept I wasn’t aware of and that I don’t yet possess.
I feel this frame helps me to better find areas where there are cached beliefs to correct, and remember hints at the shape of the thing in the past. I do feel like I have an easier time correcting object-level believes, compared to finding good ways to hone heuristics about my own thinking and how I could avoid some types of “going in circles”. I find it hard to account for hindsight, and also memory can be trickery.
Looking back at the original thread:
I currently don’t have a fast perceptual view like that, even though I think I understand the concepts mentioned above. I don’t feel like thinking of my mind as an engine helps me in post-mortems. So either I lack a skill here or this was hinting at a point I’d find elementary.
Funnily enough, cluncs and the do operator are easily implemented in nix (Merging attribute sets is quite important in nix). Only problem is that nix doesn’t support even the most basic numeric operations (like sqrt, pow (so I gave up on implementing your causal examples)).
let
do = model: overrides: let self = model (self // overrides); in self;
quad = self: {
x = 4;
constant = 3;
linear = 2 * self.x + self.constant;
result = self.x * self.x + self.linear;
};
quadDefault = do quad { };
quad2 = do quad { x = 2; };
quadNoLinear = do quad { linear = 0; };
in {
default = quadDefault.result;
overridden = quad2.result;
noLinear = quadNoLinear.result;
}
It is time that I apply the principle of more dakka and just start writing (or rather publishing) more. I know deliberate practice works for writing. If you want to be good at something, you need to do it badly and then just keep going with doing the thing is very common advice. I still find it very hard to do.
Things I ran into when writing before that stopped me:
It’s hard to decide what to write about
I get anxiety about writing something that is not good enough
Topics that go into my head about what to write about revolve mostly about how it is hard to write.
Some of the writers that I admire would see it as below their standard to write a post like that? Lots of unproductive self-reinforcing status anxiety follows.
Perfectionism around writing and what I can do on the platform that I am writing on.
The real pain: All the small decisions at the end. “Do I also post this on my blog or is it too much work?”, deciding to try and in the worst case I can eliminate that part of the process tomorrow. Noticing those annoyances could also be added to the post, then noticing how this is taking over the post
At the end, I fell for the trap of editing again (I could have been done in 1h, and now it is 1h 46 minutes).
Things that worked in my favor this time, so I ended up publishing this.
I decided to stick with 1 hour of writing and just publishing what I have at that point sounds safe to my brain.
I decided to just put it on Twitter and hope for the best on the distraction front.
Instead of editing my earlier drafts, I gave each draft an iteration number (feels more like progress!) and then I just started writing again and completing things with copilot. (This was draft 3)
I decided to stick with a nested list format, which is my favorite format for writing.
I decided no actually this nested list doesn’t work on Twitter, so I will put this in an lw shortform
gave up on adding more relevant links. That is valuable in general, and maybe I can find ways to make this easier and less distraction prone (I end up searching for a specific blogpost I like and then don’t end up finding it and read something else interesting from that blog).
That is a high false positive and false negative rate, but not fatally so if we get more coverage in return.
Actually, if we are willing to tradeoff coverage for higher probability of false negative and false positives, bisulfite sequencing might also be on the table again, because you can just run the reaction for shorter which would keep more DNA intact (presumably with some kind of bias that might not be worth the headache).
why do stimulants help ADHD? well, they short circuit the part where your brain figures out what priorities to trust based on whether they achieve your true motives
I view taking stimulants more as a move to get the more reflective parts of my brain more power (“getting my taxes done is good, because we need to do it eventually, now is actually a good time, doing my taxes now will be as boring as doing them in the future, rather than playing magic the gathering now”) in steering compared to my more primitive “true motives” that tend to be hyperbolicly discounted (“dosing in bed is nice”, “washing dishes is boring”, “doing taxes is boring”). Maybe I am horrible at self-modelling, but the part where the self model is out of sync as an explanation why the self-reflective parts have less steering power seems unnecessary.
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.
One of the first things that (shy?) people use to gauge each other’s interests before or instead of talking about anything explicit is eye contact. So I think that wearing your glasses puts you at a disadvantage unless you take them off when you are flirting. I’m not sure why you’re wearing them, but taking them off in itself could be a flirty move. I am not particularly good at flirting. But I remember in 9th grade a girl I had flirted with for like half an hour at an event via eye contact. We didn’t exchange more than ~3 sentences in person (there were no innuendos). Then she called me later that same day, asking me out explicitly if I wanted to be her boyfriend.