Ajeya Cotra @ajeya_cotra
AI could get really powerful soon and I worry we're underprepared. Analysis+grantmaking in AI alignment @open_phil (views my own), editor+writer @plannedobs. planned-obsolescence.org Berkeley, CA Joined October 2017-
Tweets225
-
Followers6K
-
Following286
-
Likes37
We just clarified eligibility criteria for our agent benchmarks RFP (openphilanthropy.org/rfp-llm-benchm…) and included a link to METR's task development resources (metr.github.io/autonomy-evals…) which many applicants may find helpful
We just clarified eligibility criteria for our agent benchmarks RFP (openphilanthropy.org/rfp-llm-benchm…) and included a link to METR's task development resources (metr.github.io/autonomy-evals…) which many applicants may find helpful
It's unfortunate how discourse about dangerous capability evals often centers threats from today's models. Alice goes "Look, GPT-4 can hack stuff / scam people / make weapons," Bob goes "Nah, it's really bad at it." Bob's right! The ~entire worry is scaled-up future systems.
I thought this was an interesting intuition pump for how superhuman AI could conduct physical R&D
I thought this was an interesting intuition pump for how superhuman AI could conduct physical R&D
Two updates on these funding opportunities: 1) We replaced our longer applications with a 300 word EOI 🎉 2) The broader RFP on studying+forecasting LLM impacts will be on hiatus after May 3 (the agent benchmarks RFP is still going strong!)
Two updates on these funding opportunities: 1) We replaced our longer applications with a 300 word EOI 🎉 2) The broader RFP on studying+forecasting LLM impacts will be on hiatus after May 3 (the agent benchmarks RFP is still going strong!)
Katharine Collins @Kat_a_Collins
881 Followers 1K Following Program Officer - Global Heath R&D @open_phil | views my own. Malaria researcher | R21 inventor #malariavaccines #malariatransmission #CHMI | DPhil @UniofOxford
Maheep Chaudhary | �.. @ChaudharyMaheep
41 Followers 509 Following MS @NTU || Collab w/ Stanford || Ex-MIT Driverless, UIUC.
Tejas Kamtam @tejaskamtam
76 Followers 208 Following he wept, for there were no more worlds to conquer
Victor Oluwatuyi @VOluwatuyi42011
8 Followers 104 Following
Paylz @paylza
137 Followers 2K Following The best online market for digital downloads with best prices.
Janvi Ahuja @jn_ahuja
575 Followers 2K Following PhD @bdi_oxford and @FHIOxford. Into cause-agnostic charity 👾
seed @tiabija
671 Followers 4K Following Grow, pollinate, evolve, fruit. @[email protected] is my backup - no posts yet
Paolo Lenzarini @PaoloLenzarini
96 Followers 849 Following
Bart Miller @BartMil92122695
177 Followers 5K Following
Matthew Feeney @M_feeney
3K Followers 1K Following Mostly at the other place. https://t.co/iWFMvuZvAx
Alice Taylor @_alice_taylor_
36 Followers 334 Following
Tom Myers @tommyers37
40 Followers 25 Following
Shirin @acodingtraveler
1K Followers 4K Following Software Engineer👩🏻💻 | Former Tech Transfer Specialist + Innovative Partnerships @NASA Tech+Design | ML+DS | @UCLA @Caltech | Writer | Musician
Claudia Richoux @_laudiacay
2K Followers 343 Following @banyancomputer is decentralizing the cloud // ex @protocollabs @trailofbits @uchicago
Anubrata Das @d_anubrata
579 Followers 2K Following Ph.D. Candidate @UTiSchool I think about Human-Centered NLP | #NLP + #HCI
Xiaoyang Yu哲学 @_xiaoyangyu
891 Followers 5K Following Phil of causality/science/mind. Cosmos is a (fatalistic) state machine: https://t.co/thoe63alMX Hard determinim: https://t.co/TY6NP7Beow
Tal Linzen @tallinzen
16K Followers 893 Following Professor @nyuling and @NYUDataScience, research scientist @GoogleAI
Michelle Nie @michellesnie
127 Followers 539 Following MPA Candidate @ Sciences Po Paris | AI Policy & Governance | Strategy & Research
Sneha Revanur @SnehaRevanur
1K Followers 503 Following Mobilizing my peers for a safe, equitable AI future @EncodeJustice
Reza Kakooee @rezakakooee
129 Followers 913 Following AI in ETH Zurich. Love reinforcement learning, from all its aspects: deep learning, theory, optimal control, neuroscience, and applications.
Jay Bee @JayBee123460856
201 Followers 1K Following
Susan AdD @DiPsyops
385 Followers 511 Following USC 🇺🇸 GDI Far Right Conspiracy Theorist NO WEF NO WHO NO WAY No DM’s No porn I"m getting the blue checkmark, i'm authentic, I support Elon
Edgard Minete @EdgardMinete
26 Followers 248 Following
Rasmus Fonnesbæk @fonnesbaek
1K Followers 4K Following Inclusive tech, digital + data/AI for the world @InstituteGC. Fan of democracy and skill-based migration. Ph.D. in PoliSci. Views always my own.
ELON MUSK🌐 @elonreeevvemusk
2 Followers 167 Following CO-FOUNDER OF NEURALINK AND OPENAI; AND PRESIDENT OF THE MUSK FOUNDATION. THE CEO TECHNOLOGY OFFICERS OF SPACEX AND TESLA.🌏🚀🚀
Fusunkathy @fusunkathy2977
360 Followers 4K Following
Wei Dai @weidai11
7K Followers 82 Following wrote Crypto++, b-money, UDT. thinking about existential safety and metaphilosophy. blogging at https://t.co/mBVFhriJVf
Shivam Pandey @ShivamPR21
175 Followers 4K Following Past: Research Engineer Intern @_FiveAI | SR. Student Research Associate @ IITK - SERB | ADAS Intern @BoschGlobal | BTech - MTech GeoInformatics, @IITKanpur
Ja @bharath4jinda
289 Followers 2K Following
Clark Benham @ClarkBenham2
42 Followers 404 Following
James Lucassen @JamesLucas71486
4 Followers 22 Following
Marina Levay @emergentodyssey
2 Followers 53 Following
Yury Orlovskiy @yvorlovskiy
9 Followers 40 Following
Matt @Matt_F_Rogers
201 Followers 272 Following Thoughts & opinions on freedom, individuality, fin-tech, crypto, opportunity creation, payment systems, personal development, growth, success, & wealth creation
Victoriayiyiyi @jnwangyi
128 Followers 2K Following
Ethan @EthanKosakHine
297 Followers 374 Following
Elias Romualdez @eliasromualdez
1 Followers 327 Following
Karthik Narasimhan @karthik_r_n
3K Followers 448 Following Assistant Professor @PrincetonCS, Head of Research @SierraPlatform. Ex @OpenAI, @MIT_CSAIL, @iitmadras
Katherine Lee @katherine1ee
6K Followers 931 Following understanding ourselves and our models. senior research scientist @GoogleBrain, @genlawcenter and @CornellCIS, formerly @Princeton @[email protected]
Cognition @cognition_labs
123K Followers 19 Following Makers of Devin, the first AI software engineer. We are an applied AI lab focused on reasoning, and code is just the beginning. Join us: https://t.co/tpfZwEwGiq
Daniel Kang @daniel_d_kang
3K Followers 84 Following Asst. professor at UIUC CS. Formerly in the Stanford DAWN lab and the Berkeley Sky Lab.
Joelle Pineau @jpineau1
10K Followers 352 Following AI researcher. VP AI Research (FAIR), @AIatMeta. Professor of Computer Science, @mcgillu. Core academic member, @Mila_Quebec
METR @METR_Evals
672 Followers 1 Following Model Evaluation and Threat Research (METR) works on building evaluations to empirically test whether cutting-edge AI systems could pose catastrophic risks.
Ana Marasović @anmarasovic
4K Followers 604 Following Asst prof @UUtah · Ex @allen_ai @uwnlp postdoc @HD_NLP PhD · she/her 🇭🇷
Sam Rodriques @SGRodriques
4K Followers 327 Following Director and CEO at FutureHouse. Building an AI scientist. https://t.co/rQYoPOxsYo
Internal Tech Emails @TechEmails
526K Followers 900 Following Internal tech industry emails that surface in public records. 🔍
Eric Newcomer @EricNewcomer
41K Followers 3K Following Founder + author of Newcomer – a startups and VC newsletter & events company. Host summits, including Cerebral Valley AI Summit Subscribe https://t.co/ptjq4pjVS9
Ashlee Vance @ashleevance
145K Followers 3K Following Writer at Bloomberg. 2X NYT best-seller. Emmy-nominated host+writer of Hello World. Doc producer @HBO. NYT, El Reg alum. My new book https://t.co/12lAgpqoYo
Julian Michael @_julianmichael_
1K Followers 122 Following Researching stuff @NYUDataScience. he/him
Jeremy Kahn @jeremyakahn
6K Followers 2K Following AI Editor @Fortunemagazine Ex-Bloomberg @business. Views my own. Retweets are not endorsements.
Gabriel Recchia @mesotronium
276 Followers 307 Following Cognitive scientist, previously at @Cambridge_Uni 's Winton Centre for Risk and Evidence Communication, now working on LLM capability evaluation & alignment
Princeton PLI @PrincetonPLI
1K Followers 19 Following Princeton University initiative enhancing fundamental understanding of AI, enabling its use in academic disciplines, and examining AI's societal implications.
Luca Righetti @lucafrighetti
720 Followers 254 Following Research @Open_Phil, co-host @HearThisIdea, prev. @FHIOxford. Views my own. I like econ, evals, and long walks on the beach.
Daniel Ziegler @d_m_ziegler
440 Followers 346 Following
Adam Marblestone @AdamMarblestone
7K Followers 5K Following Technologist, Scientist Co-founder and CEO @Convergent_FROs Tweets do not represent employers
Irina Rish @irinarish
9K Followers 994 Following prof UdeM/Mila; Canada Excellence Research Chair; AAI Lab head https://t.co/UzlrC7ZrGF; INCITE project PI https://t.co/0rV7szd7rH; CSO https://t.co/XDhj6MEtUj
Seth Lazar @sethlazar
7K Followers 2K Following ANU Philosophy Prof working on normative philosophy of computing. This place is bad. Find my work at linktree below
Charles Foster @CFGeek
2K Followers 256 Following 🪄 Tensor-enjoyer 🧪 @FinetuneLearn. Mastodon: @[email protected] Bluesky: https://t.co/rJecB0pvkA
Joel Eriksson @OwariDa
5K Followers 4K Following Offensive security researcher and entrepreneur -Kernels, browsers and all that jazz- Also: - AI/ML/DL - AR/VR/XR - CTFs (pwn/re/crypto) + Cicada 3301, Boxen etc
Max Nadeau @MaxNadeau_
385 Followers 346 Following Advancing AI honesty, control, safety at @open_phil. Prev Harvard AISST (https://t.co/xMMztyYR3O), Harvard '23.
Amanda Ngo @manda_ngo
1K Followers 240 Following prev product at https://t.co/3dW4Mn5BkI • building https://t.co/VHkAJ4RcST
alex lawsen @lxrjl
3K Followers 745 Following AI Grantmaking @ Open Philanthropy Previously 80,000 Hours, teaching, forecasting, poker. Views my 🐒's
Lennart Heim @ohlennart
3K Followers 823 Following huh? | AI (Compute) Governance @GovAI_ | Also @EpochAIResearch |
Paul Crowley ⏸️ @ciphergoth
5K Followers 1K Following Personal trivia, cryptography, and the future of all humanity. I block on a first offense for rudeness or sarcasm. https://t.co/XJfeVUjaCm
Molly ⏸️ Hickman @celloMolly
390 Followers 595 Following Data scientist @ Forecasting Research Institute. Formerly @ nLine, Inc. measuring power quality at outlets around the world. 🧬 Samotsvety Forecasting
Coffeezilla @coffeebreak_YT
574K Followers 2K Following Internet Detective Exposing Scams. Follower of Jesus. Support Investigative Journalism: https://t.co/ATpU5zUzUY
Scott Alexander @slatestarcodex
138K Followers 30 Following I have a place where I say complicated things about philosophy and science. That place is my blog. This is where I make terrible puns.
Inner City Press @innercitypress
267K Followers 3K Following Matthew Russell Lee for/as Inner City Press covers SDNY, UN Gate, banks & IMF. books https://t.co/xHL0pGID4n https://t.co/VTEqaLISDB
Molly White @molly0xFFF
116K Followers 2K Following crypto researcher & critic, software engineer, wikipedian • @web3isgreat creator • subscribe to my newsletter at https://t.co/WftJCrCfSY
Zac Goldsmith @ZacGoldsmith
115K Followers 2K Following Ex UK Foreign Office Minister for the International Environment, Climate, Forests, Ocean - in the Lords [email protected] Galajaj
Forecasting Research .. @Research_FRI
582 Followers 22 Following Research institute focused on developing forecasting methods to improve decision-making on high-stakes issues, led by chief scientist Philip Tetlock.
Jacob Andreas @jacobandreas
14K Followers 958 Following Teaching computers to read. Assoc. prof @MITEECS / @MIT_CSAIL (he/him). https://t.co/5kCnXHjtlY https://t.co/2A3qF5vdJw
Ethan Mollick @emollick
211K Followers 552 Following Professor @Wharton studying AI, innovation & startups. Democratizing education using tech Book: https://t.co/CSmipbJ2jV Substack: https://t.co/UIBhxu4bgq
Kara Swisher @karaswisher
1.5M Followers 2K Following “Vitriolic” and now “shrill”media lady, though dogs can hear me loud and clear
Lilian Weng @lilianweng
95K Followers 148 Following Working on AI safety, past on robotics, applied research @OpenAI; Writing ML blogs to help myself & others to learn; Ideas my own.
Josh Albrecht @joshalbrecht
2K Followers 56 Following Helping machines learn. CTO at Imbue (formerly Generally Intelligent)
Imbue @imbue_ai
6K Followers 14 Following Making AI systems that reason, so they can safely accomplish larger goals for us in the real world. Join us: https://t.co/9UF0rR6YUr
Kanjun 🐙🏡 @kanjun
17K Followers 488 Following understanding human & machine minds to build a creative abundant future. CEO @imbue_ai. support founders @outsetcap. co-organize https://t.co/H1aXYk96ja.
Matthew Kenney @baykenney
3K Followers 3K Following Co-Founder @aspect_labs | Previously: Senior ML Engineer at @Apple, Asst. Research Prof at @DukeU Duke Data Science | @penn_state '15 | @Cornell '11
Deb Raji @rajiinio
30K Followers 588 Following AI accountability, audits & eval. Keen on participation & practical outcomes. Fellow @mozilla, CS PhDing @UCBerkeley. forever @AJLUnited, @hashtag_include ✝️
Geoffrey Irving @geoffreyirving
8K Followers 259 Following Research Director at the UK AI Safety Institute (AISI). Previously DeepMind, OpenAI, Google Brain, etc. @[email protected]
david rein @idavidrein
2K Followers 983 Following Sentio ergo sum. AI alignment research at NYU, early employee @cohere
Cameron Jones @camrobjones
544 Followers 599 Following CogSci PhD Student at UCSD in Ben Bergen's Lab. Interested in whether matrix multiplication has a theory of mind.
Sara Hooker @sarahookr
39K Followers 7K Following I lead @CohereForAI. Formerly Research @Google Brain @GoogleDeepmind. ML Efficiency at scale, LLMs, @trustworthy_ml. Changing spaces where breakthroughs happen.
Many views on AI timelines have shortened since @ajeya_cotra's 2020 biological anchors report, but what I think of as the conservative core of her argument (even if you need enough compute to re-run evolution wholesale, good chance of AGI this century) is still not priced in IMO.
Our new paper: 1. LLMs are finetuned for alignment on examples of good behavior. 2. But they also see descriptions of bad LLMs in training. Can these descriptions subtly influence the LLM at test time?
Sam Altman had been given a lot of power, the cloak of a nonprofit, and a glowing public profile that exceeds his more mixed private reputation. He lost the trust of his board. We should take that seriously. newcomer.co/p/give-openais…
As AIs improve at persuasion & argumentation, how do we ensure that they help us seek truth vs. just sounding convincing? In human experiments, we validate debate as a truth-seeking process, showing that it may soon be needed for supervising AI. Paper: github.com/julianmichael/…
These are two really well thought-through and consequential requests for proposals—very interested to see what comes from them.
Experts disagree wildly on how powerful LLMs could get in the near future. Very excited to announce two new @open_phil RFPs for projects that could shed light on this. 1st, a call for realistic+difficult benchmarks for LLM agents (e.g. OpenAI's "GPTs"): openphilanthropy.org/rfp-llm-benchm…
I am calling for a complete and total shutdown of AI policy news until we can catch up on reading and figure out what the hell is going on.
I have so much respect for how Bengio publicly changes his mind, discusses (what he now sees as) mistakes, and emotional struggles & reflection. theglobeandmail.com/opinion/articl…
My take on "responsible scaling policies" (RSPs) 🧵 The good: having labs publicly commit to thresholds where they'd pause scaling is good! Much better than not having thresholds! The bad: governments could see voluntary commitments / self regulation as sufficient. They're not
It's a serious problem that people's perceptions of what AI is capable of today are still being shaped by GPT 3.5.
Brilliant terminology coinage by Ajeya and Kelsey. If anyone wants to learn more about quantifying schlep, I recommend checking out @pvllss work on trading off training and inference. epochai.org/blog/trading-o…
In this post, @KelseyTuoc and I introduce a distinction we've found helpful for thinking about LLM progress: you can make LLMs more useful both by *scaling* them up, and by doing *schlep* to turn the "raw models" into full-blown AI *systems*: planned-obsolescence.org/scale-schlep-a…
Also, more work in the effects of schlep is coming soon-ish from @EpochAIResearch !
Philosophy: "Our low-hanging research fruit has been plucked for the last 2000 years, so to make progress you must build a ladder." AI: "And over here we have the fruit gun. It pelts you with low-hanging research fruit every few seconds. Try not to let it distract you."
@ajeya_cotra @KelseyTuoc In fairness to the forecasters, my recollection is that the platform originally limited prediction distributions to +/- 15% or so, which meant that for uncertain events it was very likely to look more surprising than it actually was.
@AISafetyMemes I mean, this argument is basically the same as your chimpanzee argument, which I thought you agreed was not very good. It’s not like all animals got together and decided to try to build aligned humans, and failed. We are actively designing the upbringing & reward circuitry of AI
@ryancareyai @ajeya_cotra The rhythm up until now it has not been that fast however. It's more like 3x per year. It's a rapidly changing situation though, and I expect ~faster growth this year! epochai.org/trends#investm…
I’m excited about this and daunted by the responsibility: openphilanthropy.org/research/alexa…
Barbie is really good. I laughed a lot. Would recommend.
@ajeya_cotra @mllichti @daniel_271828 But I agree that we are likely to soon enter a window of opportunity to leverage AI systems—that aren’t in a position to pull off any kind of coup—as the engine of building a global immune system against global catastrophic risks (including from AI, from humans, and from nature).
@davidad @daniel_271828 @ajeya_cotra In the hard takeoff scenario, AI makes the equivalent of centuries of scientific progress in a few minutes. It can then quickly take over using advanced technologies like nanotech. In a slow takeoff, they have to use regular robots or human allies, which is much easier to stop.
Trends for United States
You might like
