42Joined Feb 2015


aka Alex K Chen


Also how does one stress-test tendency to reveal infohazards when one is under severe stress, like anger, despair, desperation, or mania?

Ideally one would stress-test as little as possible..

Does pretending that one isn't employed by X (one can have plausible deniability when one says that "one is employed by a stealth AI organization" - and there are enough stealth AI organizations out there that one can pretend to be employed by any of them), when one is in fact employed by X, help reduce the temptation to spread infohazards? (in which case - if someone is secretly employed by institution X but does not want to say that they're employed by X, how would they pretend not to be a NEET so they wouldn't face the social stigma associated with being NEET?)

FYI - the  emotional toll of keeping secrets can be lower for some people with trolling/shitposting tendencies b/c there are fewer expectations for them to produce a directionally truthful answer to prompts about their main project. roon is kind of a shitposter, and many people don't expect him to provide a truthful answer on what projects he is working on. Trolls/shitposters (or just people with crazy ADHD) also have less difficulty with announcing alternative unserious projects they're working on (which can serve as decoys)! [or they may also present themselves as having a relationship with X very different from their actual relationship with X]. ALSO trolls/shitposters just know how to distract people away from temptation (this does not always work esp in front of extremely high-agency players, but works enough to be worth mentioning).

Also, DO NOT MAKE INFOHAZARDS HILARIOUS (the more hilarious, the more tempting it is for some to share them).

The volatility of some people drastically decreases when given permanent economic security - for some other people they are inherently volatile whether or not they have economic insecurity. Some people have something they're really missing that (once filled up with economic security or once their emotional neediness needs are filled) => makes them much more reliable to not get into emotionally volatile situations long-run. 

I'm also curious if psychedelic use is predictive of leaking (or not leaking) infohazards. I would guess that some people are much more risky with them, while it does not affect the ability of others to leak infohazards (they increase volatility in some and decrease volatility in othe. It could be that some people may only be employable at infohazard-y corporations once they swear off all unsupervised psychedelic use (just like I know EAs who quit psychedelic use upon entering politics).

Also as a final note - I think others are more tempted to  to leak secrets if they think they have access to highly desireable information...


How does one translate mathematical/high-level agenty-foundations guidelines into code/instructions that an RL agent (or any AI agent, including a scaling laws one) can follow?

Has anyone noticed that nightmares are also associated with "bad trips" from psychedelics, given that psychedelics are "waking dreams"? (the fear of a bad trip is what interferes with many of the profoundly beneficial effects of psychedelics transpiring to a broader population => significantly blunts their "healing potential")

Just out of curiosity - how much time is wasted on evaluating half-ass'ed proposals?

>People are often grateful to you for granting them money. This is a mistake.

Sometimes they're resentful if you reject them (though this depends on community and is probably highly asymmetric).

Categorizing one's favorites (or putting them in folders) so one doesn't have to scroll through them all to the beginning.

At what point can large language models start to do distillation, especially of the early LW sequences?

more on mosaicism -


João Pedro Magalhães

The large number of mutations with age recent studies are finding in some human tissues showcase how difficult it will be to significantly intervene in aging because we can't easily get rid of mutant cells and replace them by pristine cells. is very deep too - actually it hints that many older cells are dominated by pro-growth/pro-survival mutations that don't complete all the necessary conditions for cancer (but it just shows how cancer is the adaptive response of A LOT of other responses that are pro-growth/survival in ordinary cells that USUALLY don't result in cancer...)

It's not just tau/junk that contributes to cytoskeleton damage - the cytoskeleton is made of proteins that are easily oxidizeable in the same way that nuclear pore complexes are, and damage to NPCs don't have tau as their primary culprit.

Load More