Chelsea Sierra Voss

engineeress ✨ fmr member of technical staff @openai // ilu @martinshkreli // prev @pilothq, @sendwaveapp, @khanacademy, @MIT, IOL, 2x IBO

[Chelsea Sierra Voss @csvoss]

@iamkim @deepfates @OpenAI oh thank goodness it's configurable now, you can click on it there to delete it! 🙏

[Chelsea Sierra Voss @csvoss]

@iamkim @deepfates @OpenAI heya – I'm on version 26.930.51102 and I'm not seeing an option to configure the shortcut in Settings > Keyboard Shortcuts, or anywhere else. was it removed?

[Chelsea Sierra Voss @csvoss]

@ashokaraj_ congrats!!

[Chelsea Sierra Voss @csvoss]

@ciphergoth Fair enough! They do say “less room” so that tracks.

[Chelsea Sierra Voss @csvoss]

“The seeds were generated in a very simple way: a common pseudorandom number generator initialized from the clock, which anyone can now reproduce. That leaves much less room for subversion and that we can be confident the parameters are secure.”

—my friends, right before I beat them at a game of GitHub Nomic back in 2019 by brute-force-searching through putative merge timestamps until I found one that hashed to my liking

https://t.co/NvEt2WOZYb

[Steve Weis @sweis]

CLAUDE FOUND THE SEEDS
https://t.co/m8N6OPQoCQ

[Chelsea Sierra Voss @csvoss]

@kipperrii guess it’s mandatory offline time 😂

[Chelsea Sierra Voss @csvoss]

@kipperrii this is the second time I’ve received this advice this morning but how am I gonna mute the words “you,” “can,” “see,” and “everything” 🫨

[Chelsea Sierra Voss @csvoss]

@QiaochuYuan aw I don’t get this at all, it’s highly readable to me. curious whether there’s some specific jargon I could help unpack

[Nick @nickcammarata]

interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had

with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works

https://t.co/zHij5zwKGT

[Chelsea Sierra Voss @csvoss]

@GuiveAssadi @tszzl @Turn_Trout (fwiw I was originally thinking only of features in my first reply, which seem clearly mechinterp? — but steering vectors do seem like a valid superset of what I meant (every feature is steerable, but not every steerable thing is a feature), so, happy to pivot/concede to those)

[Chelsea Sierra Voss @csvoss]

@GuiveAssadi @tszzl @Turn_Trout okay but “as you said above” is wrong though — I do think they’re useful! that was the whole rhetorical point of my reply :)

something can be useful while also being not yet effectively productized to its full untapped potential. like GPS before ubiquitous smartphones

[Chelsea Sierra Voss @csvoss]

@diviaprocessing You can just have nice things!

[Chelsea Sierra Voss @csvoss]

*sic (I wrote this first, then convinced myself it was wrong and deleted it. smh)

[Chelsea Sierra Voss @csvoss]

ugh it might finally be time to sicc Astra on making an up-to-date, complete, and BEAUTIFUL version of this poster so that I can finally OWN ONE

[zeta @zeta_globin]

okay but also the entire!! Roche map is just like ~5% of known human metabolism (you should be excited!!! about how much chemistry inside of you is completely unknown!!!) and Roche stopped updating it for exactly that reason

[Chelsea Sierra Voss @csvoss]

every time I think about how they stopped selling them I get mad that I don’t have one lol

[Chelsea Sierra Voss @csvoss]

@lydialaurenson @tracewoodgrains apparently you can't get sorted into Whole Earthers anyways because they're the ~ancestors~!

(found this out while trying to see what it would take to get sorted into Hackers, and failing...)

[Chelsea Sierra Voss @csvoss]

@igypham @dlbydq @micsolana tbh I’m mostly belatedly reacting to a time in September when I saw a moot whom I usually respect & agree with being obtuse in Tyler’s replies and failing to grasp what I saw as various basic concepts 😂 but it wasn’t a good time for me to try to wade into the debate then

[Chelsea Sierra Voss @csvoss]

I don’t parse Tyler’s argument as implying everyone must short

the example corroborates the argument as I understood it, which is that given one’s own beliefs, you can use a combination of financial instruments to tease apart one’s disagreement with the market

cf. the quote at the end of the OG article that kicked off this clade of discourse, which seemed to be what many on Twitter (mysteriously to me!) took issue with:

> Overall I am extremely skeptical of arguments that essentially take the form of “what I am concerned about is too big and too important to show up in any market prices.”  Pick your market prices!

https://t.co/33ucqlE0V9