I feel bad for Ajinomoto — they’re the world’s largest producer of two delicious and completely safe molecules (MSG and aspartame) that for some reason people don’t trust!
@max_spero_ No, the SynthID-Text technique they claim to use produces text that’s indistinguishable from unwatermarked outputs if you don’t know the key. The logits and probability distribution remain intact; what changes is the randomness used to sample from it.
Extremely disappointing: Fable won't just refuse to assist with AI research, it will silently give worse responses. Our tools are being designed to work against us.
mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community
also the fact that this is un purpose not visible to the user is crazy
@itsclivetime@zzzetachrome I am not inclined to assume the DoW will operate in good faith, and I don’t read the legal language as locking in current policy.
If DoW breaks the spirit but not letter of the agreement, most OAI employees may remain unaware. And unilateral termination probably isn’t an option.
@itsclivetime Only to the extent that mass domestic surveillance violates 4A or other current laws. Those are woefully inadequate for the AI age.
Fully autonomous weapons are only banned “where law, regulation, or Department policy requires human control”, which is to say, not at all.
@itsclivetime@krishnakaasyap Does this not say that “‘all lawful uses’ are permitted regarding autonomous weapons and mass domestic surveillance (regardless of whether current laws are mature and comprehensive enough to cover all aspects of AI-enabled mass surveillance)”? Because that’s how I read it.
@BayAreaNewLibs Sorry I couldn’t make it out last night! Hope @thecydneyhayes wasn’t too disappointed to miss me. I’m sure she’ll be even more distraught to learn that I also detest Trump!
Sometimes I think about e.g. multivariable calculus and call it “trivial”, but it’s good to step back every now and then and appreciate how much I’ve learned. Things only seem easy in retrospect.
@dawnlightmelody@AnniePosting I first Annie on Jan 22. She was talking about how much she liked billionaires, and that conversation was one of my inspirations for the march.
She falsely claimed to media that she was the organizer when she was not involved with that, and said some other things I disagree with.
167 Followers 677 Following16. founder papert lab (https://t.co/SMOeA1ShNz). research @virginiatech & @ucsd. augmenting human cognition with computers. cs education research
8K Followers 3K Following@MetascienceObs, @RADVACProject, and @GeneWizardX. "Speak your mind, even if your voice shakes". Leave anonymous feedback here: https://t.co/LQ5eZWxbi1
60 Followers 37 FollowingBuilding a database of replications, pioneering AI-powered large-scale literature review, and building an AI agent to do forensic metascience at scale.
193 Followers 511 FollowingFirst documented person to get modern FTM top surgery fully awake. I like abundance, classical liberalism(≠centrism), gay male culture, & big machines.
62 Followers 1K Following🎙️ Insights from conversations with leading capital allocators
🗞️ Subscribe to a free newsletter for daily top insights across 100s of interviews 👇
4K Followers 5K Followingshe/hermit. local intelligibility as defining property. world’s biggest Cynthia Ozick stan (except about Palestine). late twenties
348 Followers 882 FollowingApplying the ideas of Count Richard von Coudenhove-Kalergi to the 21st century. American Imperialist. US Army. Tamil Kayastha. Philosemite and supporter of ZOG
167 Followers 677 Following16. founder papert lab (https://t.co/SMOeA1ShNz). research @virginiatech & @ucsd. augmenting human cognition with computers. cs education research
791K Followers 4 FollowingA platform for illuminating academic papers. We annotate and share a paper every week. Save, annotate and share papers with anyone: https://t.co/0o2Pls3jmo
60 Followers 37 FollowingBuilding a database of replications, pioneering AI-powered large-scale literature review, and building an AI agent to do forensic metascience at scale.
8K Followers 3K Following@MetascienceObs, @RADVACProject, and @GeneWizardX. "Speak your mind, even if your voice shakes". Leave anonymous feedback here: https://t.co/LQ5eZWxbi1
47K Followers 4K Followingplaying with words. creating whimsical ai @ https://t.co/khnCIsqbhL. built iOS apps @xAI, @Spotify, and co-founded @imagilabs.
2K Followers 25 FollowingWe're building an open source map of the individuals and orgs with the potential to shape AI policy in the US. Join us! [email protected] https://t.co/zz0fw6hWI6
518 Followers 338 FollowingBuilding (generally) intelligent robots @pantographPBC
Will probably try to talk to you about that one chobani solar punk commercial.
17K Followers 494 FollowingHelping the world prepare for extremely powerful AI. Risk assessment @METR_evals. Writing at Planned Obsolescence (about AI), Good Bones (about whatever).
3K Followers 2K Following"The rules of maximizing expected value are simple and finite. Any Cosmo girl would have known." Personal account: @Kirsten3531
3K Followers 507 Following@thinkymachines, formerly @workshoplabs cofounder. Blog: https://t.co/GwKY6jjw3N
Tweets are all strictly in a personal capacity
3K Followers 439 Following30–day microblogging challenge til Sep15 (join!) || Working on making sci-fi futures real || @AevitasHouse + @PrimordiaGrants
193 Followers 511 FollowingFirst documented person to get modern FTM top surgery fully awake. I like abundance, classical liberalism(≠centrism), gay male culture, & big machines.
13K Followers 3K FollowingNY Assemblymember, computer scientist, dad, husband. Feared by the powerful, fighting for NY. Running for Congress in NY-12 because the old ways aren't working.
44 Followers 17 FollowingOut of this Box! is a musical comedy with original pop/jazz songs made to explain AI safety arguments for the general public. [email protected]