LessWrong 2.0 Reader

I'm inclined to bite this bullet too, though it feels somewhat strange. Weird implication: you can increase the amount of reality-fluid assigned to you by giving yourself amnesia.

I explored a similar line of reasoning here: Magic by forgetting [LW · GW]

I think that yes, the sameness of humans as agents is generated by the process of self-identification in which a human being is identifies herself through a short string of information "Name, age, sex, profession + few more kilobytes". Evidence for this is the success of improv theatre, where people quickly adopt completely new roles through one-line instructions. 

If yes, then we should expect ourselves to be agents that exist in a universe that abstracts well, because "high-level agents" embedded in such universes are "supported" by a larger equivalence class of universes (since they draw on reality fluid from an entire pool of "low-level" agents).

I think that your conclusion is valid. 

If I’ll probably see them again, I don’t miss people. I thought people saying they miss you were just being overly polite.

Yeah I definitely agree you should start learning as young as possible. I think I would usually advise someone starting out to learn general math/CS stuff and do AI safety on the side, since there's way more high-quality knowledge in those fields. Although "just dive in to AI" seems to have worked out well for some people like Chris Olah, and timelines are plausibly pretty short so ¯\_(ツ)_/¯

Yes this world.

Interesting post! I’ve wondered the same thing before.

I’m doing a much more half-arsed version, as a casual quantum foundations enjoyer alongside a technical writing job, and also getting endlessly distracted by other things I find interesting, so my output is not impressive. But it’s a pretty fun hobby and I’m surprised more people don’t try this!

Hi Neel,

you're absolutely right, all research in the gemmascope paper was performed on the open source Gemma 2 model. I wanted to group up all research that my paper was based on in a concise sentence and by doing so erroneously put you in the 'proprietary LLMs' section. I went ahead and corrected the mistake.

My apologies.

I hope you still enjoyed the project and thank you for your great research work at DeepMind. =)

Could you post coordinates next time? I can't find the entrance on Elizabeth St. you're referring to

The corresponding arbital page is now (apparently) dead.

A link appears to have broken, does anyone know what “null” was supposed to link to in “policy  null ” (note the extra spaces around “null”

I had to write several new Python versions of the code to explore the problem before it clicked for me.

I understand the proof, but the closest I can get to a true intuition that B is bigger is:

• Imagine you just rolled your first 6, haven't rolled any odds yet, and then you roll a 2 or a 4.
• In the consecutive-6 condition, it's quite unlikely you'll end up keeping this sequence, because you now still have to get two 6s before rolling any odds.
• In the two-6 condition, you are much more likely to end up keeping this sequence, which is guaranteed to include at least one 2 or 4, and likely to include more than one before you roll that 6.

I think the main think I want to remember is that "given" or "conditional on X" means that you use the unconditional probability distribution and throw out results not conforming to X, not that you substitute a different generating function that always generates events conforming to X.