←
prerat33,166 posts
prerat avatar

prerat

@prerat

social media influencee

◷ Joined August 2020▧ 4,374 Photos and videos
prerat avatar
prerat@prerat·Nov 16

I've never played GTA but I might want to try this one, since it's the last big human-coded game ever. Like going to an art museum to see the hand-carved statues

prerat avatar
prerat@prerat·Nov 13

Nvidia is nothing without its people 💚

WIRED@WIREDNov 13

An internal post urging the board to consider employee sentiment about Nvidia’s former CEO has drawn more than 1,600 comments, the majority in full support of reinstalling him. buff.ly/iaZuiLF

prerat avatar
prerat@prerat·Nov 13

My kid's school sent this home. What % AI do you think this fox is?

Post attachment
prerat avatar
prerat@prerat·Nov 11

Polymarket this is so close to being useful Please make two CONDITIONAL markets for NVDA price: one if Jensen stays, the other if Jensen out Then you (& shareholders) can see if the markets think ousting Jensen will raise or lower the price, before actually deciding

Post attachment

75% chance Jensen is out as CEO this year. poly.market/2gD4gqP

prerat avatar
prerat@prerat·Nov 9

smh frontier models still think 5.2 > 5.5

prerat avatar
prerat@prerat·Nov 8

Either you understood how Duck Hunt on NES worked, or you were gonna be confused anyway

Post attachment

Is it kind of a little messed up that Nintendo straight up went "Hmm, it's a little unintuitive that the pointer sensor is in the Wiimote, how about we just lie and tell them the lightbulbs are a sensor bar" like just because I'm stupid doesn't mean you can play these mind tricks

prerat avatar
prerat@prerat·Nov 7

iiuc CDT is fine at voting, it's just sometimes computationally confusing because the probability of being pivotal is tricky but since UDT can be implemented via CDT+SIA (from @jessi_cata), you *can* turn out to vote using CDT just as well as using UDT lesswrong.com/posts/5bd75cc5…

prerat avatar
prerat@prerat·Nov 7

Guy who deeply understands the idea of saying things that are technically "true" but misleading the public on national TV and so on

crying this is so elaborate

3:00
prerat avatar
prerat@prerat·Nov 6

This post was 17 years ago when Dario was a grad student. He did end up giving the child vaccine charity $10,000 and he was a student not a billionaire yet. In the post he writes about how it's worth doing but still emotionally hard to put numbers on things like this

Post attachment

whew nytimes.com/2026/11/05/techn…

Post attachment
prerat avatar
prerat@prerat·Nov 5

(ok it's a stretch but it felt related)

prerat avatar
prerat@prerat·Nov 5

*taps mic* people wanna be told how to be artists themselves more than they wanna see anyone else's art

there's a joke in photography spheres even before the internet that if you want to make money in photography, your best bet is to make a paid workshop called "how to make money with photography", where you explain how to sell paid workshops instead

prerat avatar
prerat@prerat·Nov 3

Tentative Plan A scenario: - Solve deep interpretability for current-ish models and create a DEEPLY TRUSTWORTHY but only-slightly-superhuman agent (no strong-super generalization yet) - Etch it into some chips, temp ban all other models - Hopefully it can solve superalignment

Basic idea is just Do Things That Don't Scale This kind of alignment very likely WOULDN'T persist across self-improvements or work at all on a sufficiently intelligent AI BUT could it get us a *somewhat* superhuman researcher that is at least as trustworthy as a human upload?

prerat avatar
prerat@prerat·Nov 3

Basic idea is just Do Things That Don't Scale This kind of alignment very likely WOULDN'T persist across self-improvements or work at all on a sufficiently intelligent AI BUT could it get us a *somewhat* superhuman researcher that is at least as trustworthy as a human upload?

prerat avatar
prerat@prerat·Nov 3

Difficulties include that it's an alien made of linear algebra we don't understand But it's deterministic & we can read its explicit thoughts & inspect all its neurons for implicit thoughts & even do arbitrary brain surgery experiments etc When it's NOT SUPER, is this enough?

prerat avatar
prerat@prerat·Nov 3

I think an under-discussed question is how hard it would be to create a deeply well aligned ONLY MODERATELY HIGH CAPABILITIES agent. Superalignment has all these extra difficulties including reflexive stability etc But what about like a Fable 8 that we trust >= a good human?