LESSWRONG
LW

All of rai's Comments + Replies

Scaling Wargaming for Global Catastrophic Risks with AI
rai30

Would love that! You can send them to the email above, comment here, whatever works for you. 

Reply1
Introducing Squiggle AI
rai30

This is cool. I asked it a question in a field I'm not knowledgable about at all but my roommate is. I think having the structure there to make it maximally easy to elicit his expertise is super valuable.

Reply1
Communications in Hard Mode (My new job at MIRI)
rai50

Wow, somewhere along the way I got the impression that MIRI is a quiet org that occasionally speaks loudly… so much so that I never thought to check if they put out a newsletter.

Reply
Thoughts on seed oil
rai130

It was gluten.

Reply
Thoughts on seed oil
rai90

This is shockingly similar to what I'm going through.  And the fries that fucked me up the other night are indeed fried in canola oil. I'm cautiously optimistic but I know how complicated these things can be -_-. Will report back!

Reply1
rai130

It was gluten.

Reply
1David Cato
I wish you the best and look forward to hearing how it goes.
Orexin and the quest for more waking hours
rai10

Ok so modulo @GeneSmith's comments about this gene potentially being red herring, if there were a genetic basis to this, we'd expect to see prevalance of this trait increase due to genetic drift and removal of the selection pressure right?

I don't have a good sense of how quickly we'd expect to be able to detect those population level differences with the rate at which calories have become more available and the rate of growth of the variant.

Figure 7 in @guzey's post shows that it's directionally been increasing but a lot has changed about our environments ... (read more)

Reply
2ChristianKl
I don't see why genetic drift would increase this particular mutation. On average I expect genetic drift to destroy the body's ability to produce substances and not increase their production of them.  If the selection pressure is due to certain risk-taking behavior, it's not clear that we see a removal of selection pressure. In humans, we certainly don't have an environment which is similar to those cavefish as far as the selection pressures go.
Significantly Enhancing Adult Intelligence With Gene Editing May Be Possible
rai23

That matrix goes a long way in showing that there isn't much correlation between diseases in the natural distribution. What is the reason to believe those correlations will remain low when you are making edits resulting in an extremely unlikely genome?

Reply
5kman
We'd edit the SNPs which have been found to causally influence the trait of interest in an additive manner. The genome would only become "extremely unlikely" if we made enough edits to push the predicted trait value to an extreme value -- which you probably wouldn't want to do for decreasing disease risk. E.g. if someone has +2 SD risk of developing Alzheimer's, you might want to make enough edits to shift them to -2 SD, which isn't particularly extreme. You're right that this is a risk with ambitious intelligence enhancement, where we're actually interested in pushing somewhat outside the current human range (especially since we'd probably need to push the predicted trait value even further in order to get a particular effect size in adults) -- the simple additive model will break down at some point. Also, due to linkage disequilibrium, there are things that could go wrong with creating "unnatural genomes" even within the current human range. E.g. if you have an SNP with alleles A and B, and there are mutations at nearby loci which are neutral conditional on having allele A and deleterious conditional on having allele B, those mutations will tend to accumulate in genomes which have allele A (due to linkage disequilibrium), while being purged from genomes with allele B. If allele B is better for the trait in question, we might choose it as an edit site in a person with allele A, which could be highly deleterious due to the linked mutations. (That said, I don't think this situation of large-conditional-effect mutations is particularly likely a priori.)
rai's Shortform
rai30

Kernel of something that might inspire someone else who knows more than I.

Assuming weights that have “grokked” a task are more interpretable, is there use in modifying loss functions to increase grokking likelihood? Perhaps by making it path dependent on the updates of the weights themselves?

Reply
A No-Nonsense Guide to Early Retirement
rai20

I think it'd be good to add an endnote mentioning that while saving for your FIRE number is relatively straightforward, withdrawing it can be much more complicated. I know my intuitions didn't serve me at all when thinking about that phase. https://earlyretirementnow.com/2018/06/27/ten-things-the-makers-of-the-4-rule-dont-want-you-to-know/

Reply