jessica dai
@jessica.bsky.social
go bears!!!
jessicad.ai
kernelmag.io
jessicad.ai
kernelmag.io
Pinned
jessica dai
@jessica.bsky.social
· Jul 1
individual reporting for post-deployment evals — a little manifesto (& new preprints!)
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
hey wasn't this the same company that made a beautiful shiny "research" post about how AI evals should include error bars or something like that. or did they decide the CLT didn't apply here
August 6, 2025 at 3:20 AM
hey wasn't this the same company that made a beautiful shiny "research" post about how AI evals should include error bars or something like that. or did they decide the CLT didn't apply here
individual reporting for post-deployment evals — a little manifesto (& new preprints!)
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
July 1, 2025 at 11:38 PM
individual reporting for post-deployment evals — a little manifesto (& new preprints!)
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
tldr: end users have unique insights about how deployed systems are failing; we should figure out how to translate their experiences into formal evaluations of those systems.
Reposted by jessica dai
@jessica.bsky.social on individual reporting as a means to build collective knowledge.
Individual experiences and collective evidence
Jessica Dai on theory for the world as it could be
www.argmin.net
June 24, 2025 at 2:46 PM
@jessica.bsky.social on individual reporting as a means to build collective knowledge.
where are the bullshit "x% of experts believe" polls when you need them lol
April 24, 2025 at 5:31 PM
where are the bullshit "x% of experts believe" polls when you need them lol
perhaps this is a stupid question but given that ai safety has been a pretty vibrant (+ well funded) field for the last 5-10 years... how should we be thinking about the concern that (ai) catastrophe still is, allegedly, imminent
April 18, 2025 at 5:42 AM
perhaps this is a stupid question but given that ai safety has been a pretty vibrant (+ well funded) field for the last 5-10 years... how should we be thinking about the concern that (ai) catastrophe still is, allegedly, imminent
in middle school we were asked to write a short story in the style of edgar allan poe. as you might expect, all of our little pieces (even, especially, the ones the students thought were "good") were hilariously bad. anyway, i had forgotten about that homework until now
x.com/sama/status/...
x.com/sama/status/...
Sam Altman on X: "we trained a new model that is good at creative writing (not sure yet how/when it will get released). this is the first time i have been really struck by something written by AI; it got the vibe of metafiction so right. PROMPT: Please write a metafictional literary short story" / X
we trained a new model that is good at creative writing (not sure yet how/when it will get released). this is the first time i have been really struck by something written by AI; it got the vibe of metafiction so right. PROMPT: Please write a metafictional literary short story
x.com
March 11, 2025 at 8:09 PM
in middle school we were asked to write a short story in the style of edgar allan poe. as you might expect, all of our little pieces (even, especially, the ones the students thought were "good") were hilariously bad. anyway, i had forgotten about that homework until now
x.com/sama/status/...
x.com/sama/status/...
im ngl i think this kinda just means u are stupid
February 10, 2025 at 6:03 PM
im ngl i think this kinda just means u are stupid
how to have a good time in a phd
(not real advice)
letterstomyfriends.substack.com
January 1, 2025 at 7:04 PM
i don't work well under deadline pressure but i also don't work well without it. therefore,
December 18, 2024 at 1:01 AM
i don't work well under deadline pressure but i also don't work well without it. therefore,
Reposted by jessica dai
The plan: Post your dissertation abstract online to rekindle a decades-long controversy about the utility of the humanities, turning your paper into the most-read publication in the history of your field
December 3, 2024 at 11:28 PM
The plan: Post your dissertation abstract online to rekindle a decades-long controversy about the utility of the humanities, turning your paper into the most-read publication in the history of your field
i just know these people would have been the biggest fans of japanese internment
December 1, 2024 at 2:23 AM
i just know these people would have been the biggest fans of japanese internment
Reposted by jessica dai
one year ago today i got to break the news that kissinger died to the entire reboot editorial board. an absolutely thrilling moment
November 29, 2024 at 9:11 PM
one year ago today i got to break the news that kissinger died to the entire reboot editorial board. an absolutely thrilling moment
Reposted by jessica dai
foreign interference is when women don't want to continue talking to me for more than 30 seconds
November 28, 2024 at 6:12 PM
foreign interference is when women don't want to continue talking to me for more than 30 seconds
i am NOT seeing enough trolling about this "did you know that sometimes you can use the CLT for evals" be so for real right now
November 22, 2024 at 4:33 PM
i am NOT seeing enough trolling about this "did you know that sometimes you can use the CLT for evals" be so for real right now
ok im going to revive this thread
insufferable ai twitter (all sides) is not here (yet) so in the meantime going to workshop my hot takes into the void
November 20, 2024 at 12:43 AM
ok im going to revive this thread
ok i guess i might be coming back here
November 18, 2024 at 5:52 PM
ok i guess i might be coming back here
this is too mean of a take for twitter but that one thread where people are like "don't do a phd in AI, the field moves too fast, you'll be stuck, academia sucks"...
sounds like lots of ppl are motivated primarily by glory/ ego and prob better for them not to be in the academic community anyway :^)
sounds like lots of ppl are motivated primarily by glory/ ego and prob better for them not to be in the academic community anyway :^)
November 2, 2023 at 5:25 PM
this is too mean of a take for twitter but that one thread where people are like "don't do a phd in AI, the field moves too fast, you'll be stuck, academia sucks"...
sounds like lots of ppl are motivated primarily by glory/ ego and prob better for them not to be in the academic community anyway :^)
sounds like lots of ppl are motivated primarily by glory/ ego and prob better for them not to be in the academic community anyway :^)
poetry and manifestos in an anarchist hackerspace.. still so psyched about the turnout for kernel magazine x lit crawl :^D
October 24, 2023 at 7:46 PM
poetry and manifestos in an anarchist hackerspace.. still so psyched about the turnout for kernel magazine x lit crawl :^D
The Artificiality of Alignment
How are we actually “aligning AI with human values”?
joinreboot.org
August 19, 2023 at 8:07 PM
let me just say that talking about the importance of values pluralism is... well, clearly, not the same thing as actually being open to conflicting values and alternative proposals for the future ...! 🤨
July 15, 2023 at 12:17 AM
let me just say that talking about the importance of values pluralism is... well, clearly, not the same thing as actually being open to conflicting values and alternative proposals for the future ...! 🤨
this piece was so horrendous it singlehandedly cured my imposter syndrome about having anything interesting to say about AI
https://www.newyorker.com/science/annals-of-artificial-intelligence/can-we-stop-the-singularity
https://www.newyorker.com/science/annals-of-artificial-intelligence/can-we-stop-the-singularity
June 3, 2023 at 11:34 PM
this piece was so horrendous it singlehandedly cured my imposter syndrome about having anything interesting to say about AI
https://www.newyorker.com/science/annals-of-artificial-intelligence/can-we-stop-the-singularity
https://www.newyorker.com/science/annals-of-artificial-intelligence/can-we-stop-the-singularity
Reposted by jessica dai
how come male CS grad students all dress like the seaborn default color palette
May 30, 2023 at 6:21 AM
how come male CS grad students all dress like the seaborn default color palette
there is nothing a good gummy vitamin cannot solve
May 7, 2023 at 1:13 AM
there is nothing a good gummy vitamin cannot solve
somehow ai twitter got even worse so uh, if you have a good ai opinion pitch me? lol
insufferable ai twitter (all sides) is not here (yet) so in the meantime going to workshop my hot takes into the void
May 5, 2023 at 4:27 AM
somehow ai twitter got even worse so uh, if you have a good ai opinion pitch me? lol