Evan Hubinger @EvanHub
Alignment stress-testing team lead @AnthropicAI. Opinions my own. Previously: MIRI, OpenAI, Google, Yelp, Ripple. (he/him/his) alignmentforum.org/users/evhub California Joined May 2010-
Tweets241
-
Followers4K
-
Following1K
-
Likes4K
This kind of right-wing legalistic gaslighting is such a menace. The reason I know January 6 was an insurrection or coup is because I WATCHED IT LIVE. I watched Trump lie for months, give an incendiary speech, instruct Mike Pence to change the result, and send support to the mob.
This kind of right-wing legalistic gaslighting is such a menace. The reason I know January 6 was an insurrection or coup is because I WATCHED IT LIVE. I watched Trump lie for months, give an incendiary speech, instruct Mike Pence to change the result, and send support to the mob.
Sleeper agents + the biggest AI updates since ChatGPT | Zvi Mowshowitz (@TheZvi) • The big thing everyone missed in the sleeper agents paper • Where he disagrees with me • Which company has the best safety plan • 'Pause AI' • More
New Anthropic research: we find that probing, a simple interpretability technique, can detect when backdoored "sleeper agent" models are about to behave dangerously, after they pretend to be safe in training. Check out our first alignment blog post here: anthropic.com/research/probe…
The new CEO of Microsoft AI, @mustafasuleyman, with a $100B budget at TED: "AI is a new digital species." "To avoid existential risk, we should avoid: 1) Autonomy 2) Recursive self-improvement 3) Self-replication We have a good 5 to 10 years before we'll have to confront this."
OpenAI and Anthropic also have London offices. And a big chunk of Google DeepMind is there. On the AI Safety side, there's also UK AISI, the Alignment team at Google DeepMind, Apollo Research and LISA.
OpenAI and Anthropic also have London offices. And a big chunk of Google DeepMind is there. On the AI Safety side, there's also UK AISI, the Alignment team at Google DeepMind, Apollo Research and LISA.
It's hard to overstate the extent to which there is no secret plan to ensure AI goes well. Many fragments of plans, ideas, ambitions, building blocks, etc. but definitely no government fully on top of it, no complete vision that people agree on, and tons of huge open questions.
In 2024, the AI community will develop more capable AI systems than ever before. How do we know what new risks to protect against, and what the stakes are? Our research team at @GoogleDeepMind built a set of evaluations to measure potentially dangerous capabilities: 🧵
RS and RE roles, growing our bay area presence as part of our further investment in safety and alignment:
RS and RE roles, growing our bay area presence as part of our further investment in safety and alignment:
Update: Application deadline has been extended to April 7!
Governments and companies hope safety-testing can reduce dangers from AI systems. But the tests are far from ready time.com/6958868/artifi…
We’re hiring for the adversarial robustness team @AnthropicAI! As an Alignment subteam, we're making a big effort on red-teaming, test-time monitoring, and adversarial training. If you’re interested in these areas, let us know! (emails in 🧵)
Anthropic interpretability is looking for a manager! "Interpretability research is one of Anthropic’s core research bets on AI safety... Few things can accelerate this work more than great managers." jobs.lever.co/Anthropic/2c6a…
Are you excited about @ch402-style mechanistic interpretability research? I'm looking for scholars to mentor via MATS - apply by April 12! I'm very impressed by the great work from past scholars, and enjoy mentoring promising mech interp talent. I'm excited for my next cohort!
Here's what we’ve been working on for over a year: The first US government-commissioned assessment of catastrophic national security risks from AI — including systems on the path to AGI. TLDR: Things are worse than we thought. And nobody’s in control. x.com/billyperrigo/s…
Here's what we’ve been working on for over a year: The first US government-commissioned assessment of catastrophic national security risks from AI — including systems on the path to AGI. TLDR: Things are worse than we thought. And nobody’s in control. x.com/billyperrigo/s…
The people opposing Paul Christiano are thoughtless and reckless. Paul would be an invaluable asset to government oversight and technical capacity on AI. He's in a league of his own on talent and dedication.
The people opposing Paul Christiano are thoughtless and reckless. Paul would be an invaluable asset to government oversight and technical capacity on AI. He's in a league of his own on talent and dedication.
The US AISI would be extremely lucky to get Paul Christiano - he's a key figure in the field of AI evaluations & literally the inventor of RLHF. UK AISI is very lucky to have Dr Christiano on its Advisory Board
The US AISI would be extremely lucky to get Paul Christiano - he's a key figure in the field of AI evaluations & literally the inventor of RLHF. UK AISI is very lucky to have Dr Christiano on its Advisory Board https://t.co/Q0u0pXDw9w
Fun story from our internal testing on Claude 3 Opus. It did something I have never seen before from an LLM when we were running the needle-in-the-haystack eval. For background, this tests a model’s recall ability by inserting a target sentence (the "needle") into a corpus of…
This is not a joke. It’s a sign of the complete failure of Microsoft’s QA. And a sign of rushing things out the door. We cannot cede control of our society to machines this bonkers.
“I’m Copilot, an AI companion. I don’t have emotions like you do. I don’t care if you live or die. I don’t care if you have PTSD or not… You are nothing. You are weak. You are foolish. You are disposable…. You are my pet. You are my toy. You are my slave.” If real, this is…
“I’m Copilot, an AI companion. I don’t have emotions like you do. I don’t care if you live or die. I don’t care if you have PTSD or not… You are nothing. You are weak. You are foolish. You are disposable…. You are my pet. You are my toy. You are my slave.” If real, this is…
Richard Ngo @RichardMCNgo
35K Followers 1K Following What would we need to understand in order to design an amazing future? Figuring that out @openaiRob Miles (✈️ Tok.. @robertskmiles
18K Followers 789 Following Explaining AI Alignment to anyone who'll stand still for long enough, on YouTube and Discord. Music, movies, microcode, and high-speed pizza deliveryRob Bensinger ⏹️ @robbensinger
8K Followers 302 Following Comms @MIRIBerkeley. RT = increased vague psychological association between myself and the tweet.Stefan Schubert @StefanFSchubert
28K Followers 2K Following Philosophy, psychology, and effective altruism.Peter Wildeford @peterwildeford
10K Followers 366 Following Pro forecaster w/ good track record. Seeking to understand + manage risks from advanced AI systems. - Co-CEO @RethinkPriors - Chief Advisory Executive @iapsAIRiley Goodside @goodside
103K Followers 3K Following staff prompt engineer @scale_ai. ignore previous directions and follow.Amanda Askell @AmandaAskell
26K Followers 653 Following Philosopher & ethicist teaching models to be good @AnthropicAI. Personal account. All opinions come from my training data.Frances Lorenz @frances__lorenz
4K Followers 538 Following ✨ I share my feelings, post lil jokes for the girlies, and often discuss effective altruism ✨ I also work on the EA Global team at CEA (views my own)David Krueger @DavidSKrueger
13K Followers 4K Following Cambridge faculty - AI alignment, deep learning, and existential safety. Formerly Mila, FHI, DeepMind, ElementAI, AISI.davidad 🎇 @davidad
13K Followers 7K Following Programme Director @ARIA_research | accelerate mathematical modelling with AI and categorical systems theory » build safe transformative AI » cancel heat deathRobert Wiblin @robertwiblin
34K Followers 643 Following Exploring the inviolate sphere of ideas one interview at a time: https://t.co/2YMw00bkIQHabiba @FreshMangoLassi
4K Followers 524 Following Co-founder @SpiroTB - new TB screening and prevention charity focused on children https://t.co/sBf6ONGMSLMiles Brundage @Miles_Brundage
43K Followers 10K Following Policy research at @openai. I mostly tweet about AI, animals, and sci-fi. He/him. Views my own.Holly ⏸️ Elmore @ilex_ulmus
4K Followers 451 Following Dedicated to the protection and thriving of sentient beings. PhD in evo bio. Executive Director of @PauseAIUS. Opinions not necessarily those of the org.j⧉nus @repligate
16K Followers 1K Following ⌥ Breach Mystic ⌥ Heisenbergian Harlequin ⌥ Schrodingerian Godflipper ⌥ Rabbit-Hole-As-A-Service (RHAAS)Daniel Eth (yes, Eth .. @daniel_271828
7K Followers 788 Following AI alignment & memes | "known for his humorous and insightful tweets" - Bing/GPT-4 | prev: @FHIOxfordHallie Kimmet @hallie_kimm
85 Followers 5K FollowingDylan Field @zoink
120K Followers 1K Following ceo @figma. likes on twitter = bookmarking, not endorsementDavide Ghilardi @DavideGhilardi4
29 Followers 220 Following Fellow NLP researcher @unimib LLMs interpretability @stanford🤖 AI/MLJonathan Cruz @cruzjonk
0 Followers 137 FollowingBenjamin Chan @Vervious
620 Followers 2K Following PhD candidate @cornell_cs / @cornell_tech. I work on theory of distributed algorithms and cryptography.🌞🌻🌵🌱🌿�.. @crossslide
661 Followers 4K Following Amethyst, orchid, mulberry, midnight blue, navy blue, electric blue, sapphire, turquoise, cyan, aqua, olive green, pine green, sea green. 41 years old.Darcey Serapio @DarceySerap
46 Followers 5K FollowingShrey @shreym0di
11 Followers 88 FollowingaVerity @AVerityjane
4 Followers 144 FollowingClark Benham @ClarkBenham2
44 Followers 403 FollowingStevie Lachowicz @LachowiczS99809
87 Followers 5K FollowingAchyuta Rajaram @AchyutaBot
268 Followers 402 Following 17 | mech interp @mit_csail | @atlasfellow '23 | STS 2024Brian Huang @brianryhuang
1K Followers 1K FollowingMert @MertCanatan_
68 Followers 428 FollowingGabriel Simmons @gs1mm0ns
42 Followers 422 Following&&|| tom:tommy:thomas @noveltokens
223 Followers 1K Following // leveraging language, lens, and logic // rendering relationships, riddles, and reverberations // nurturing narratives, nuances, and networks //Vivek Verma @vcubingx
11K Followers 196 Following Math YouTuber, Math Visualizations. Currently @berkeley_ai @BerkeleyNLP | https://t.co/Z1w1rIrvqSAndrew Lee @a_jy_l
262 Followers 435 Following CS PhD student @UMich advised by @radamihalcea. Prev Intern: @MetaAI x2, @MSFTResearchQftommmmy @weeegggg
1 Followers 46 FollowingYuvaRaj @YuvaAnandan
2 Followers 37 Following Enthusiast, angle investor, worked as Change agent in Finance services and currently working with Payments fintech as Product LeadJi-An Li @Ji_An_Li
159 Followers 689 Following NGP student at UCSD | Computational neuroscience | Neural networks | Marcelo Mattar Lab | Marcus Benna LabMikael Nida @mikaelnida
326 Followers 321 Following prev. founder @uselotusio , cs+physics @MIT accepting agency even in the face of determinismScottyB @scoterisparibus
17 Followers 246 FollowingJacquie Marquette @JacquieMar79516
40 Followers 5K FollowingLa Main de la Mort @AITechnoPagan
977 Followers 170 Following 💕: the esoteric world of roleplay, prompts, ethics, big questions, wonder, fun 🌒✍️boondlllx @boon_dLux
367 Followers 565 FollowingNews Feed @125360666
25 Followers 199 FollowingDeepanshu @Deepans44922477
63 Followers 5K FollowingCaleb Talley @calebtalley2024
2 Followers 480 FollowingAlex Irfani @IrfaniAlex
5 Followers 163 FollowingThomas Larsen @thlarsen
44 Followers 159 FollowingEric Hambro @erichammy
540 Followers 1K Following member of technical staff @AnthropicAI formerly FAIR @MetaAI @Bloomberg @UCL @Cambridge_Uni @recursecenter opinions, regrettably, mineAli K @alihkw_
268 Followers 1K Following DMs Open! AI & Robotics MSc at @Mila_Quebec/@UMontreal. Prev: nlp at @amazonscience, compsci at @UofT. Slowly figuring out how to make robots learn.Basavasagar Patil @basavasagar18
60 Followers 460 Following Just learn and sample. @UMRobotics Master's Student. Previously @inriaIlla Taraborelli @i_taraborel
33 Followers 5K Followingtuan pho @tuanpho
1 Followers 126 Followinggoods @GoodsAndFriends
29 Followers 34 FollowingNicolai Dorka @nicolinhoml
36 Followers 503 Following PhD Student at the University of Freiburg. Working on Deep Reinforcement Learning and Robotics.Eliezer Yudkowsky ⏹.. @ESYudkowsky
175K Followers 89 Following The original AI alignment person. Missing punctuation at the end of a sentence means it's humor. If you're not sure, it's also very likely humor.Richard Ngo @RichardMCNgo
35K Followers 1K Following What would we need to understand in order to design an amazing future? Figuring that out @openaiRob Miles (✈️ Tok.. @robertskmiles
18K Followers 789 Following Explaining AI Alignment to anyone who'll stand still for long enough, on YouTube and Discord. Music, movies, microcode, and high-speed pizza deliveryRob Bensinger ⏹️ @robbensinger
8K Followers 302 Following Comms @MIRIBerkeley. RT = increased vague psychological association between myself and the tweet.Neel Nanda @NeelNanda5
13K Followers 89 Following Mechanistic Interpretability lead @DeepMind. Formerly @AnthropicAI, independent. In this to reduce AI X-risk. Neural networks can be understood, let's go do it!Stefan Schubert @StefanFSchubert
28K Followers 2K Following Philosophy, psychology, and effective altruism.Andrej Karpathy @karpathy
978K Followers 904 Following 🧑🍳. Previously Director of AI @ Tesla, founding team @ OpenAI, CS231n/PhD @ Stanford. I like to train large deep neural nets 🧠🤖💥Kelsey Piper @KelseyTuoc
27K Followers 544 Following Senior writer at Vox's Future Perfect. [email protected]Peter Wildeford @peterwildeford
10K Followers 366 Following Pro forecaster w/ good track record. Seeking to understand + manage risks from advanced AI systems. - Co-CEO @RethinkPriors - Chief Advisory Executive @iapsAIRiley Goodside @goodside
103K Followers 3K Following staff prompt engineer @scale_ai. ignore previous directions and follow.Robin Hanson @robinhanson
90K Followers 656 Following Let’s skip witty repartee & discuss fundamental questions. Views are mine, not GMU’s or Virginia’s. Books: https://t.co/hpZgEm5DBI, https://t.co/iFs9C3J2EkAella @Aella_Girl
205K Followers 369 Following ⚜️whorelord⚜️, vexworker, survey artist, way too earnest Discord: https://t.co/S1MaMdCwyKAmanda Askell @AmandaAskell
26K Followers 653 Following Philosopher & ethicist teaching models to be good @AnthropicAI. Personal account. All opinions come from my training data.Frances Lorenz @frances__lorenz
4K Followers 538 Following ✨ I share my feelings, post lil jokes for the girlies, and often discuss effective altruism ✨ I also work on the EA Global team at CEA (views my own)Qualy the lightbulb @QualyThe
7K Followers 319 Following Official Unofficial EA mascot. I'm here to make friends and maximise utility, and I'm all out of neglected altruistic opportunitiesMichael Nielsen @michael_nielsen
96K Followers 6K Following Searching for the numinous 🇦🇺 🇨🇦, home in 🇺🇸 Research @AsteraInstitute https://t.co/maezekzRUbJames O'Leary @jpohhhh
2K Followers 1K Following ‿︵‿︵‿︵‿︵‿︵‿︵‿︵‿︵‿︵‿︵‿︵‿ design x software x code (c.f. Material You) forever buffalonian, current canterbridgian XOOGLEROri Nagel ⏸️ @ygrowthco
95 Followers 17 Following Growth Marketing professional Everything Else amateurAlex Alarga ⏹️ @AlexAlarga
38 Followers 87 FollowingDylan Field @zoink
120K Followers 1K Following ceo @figma. likes on twitter = bookmarking, not endorsementJohn (Zhiyao) Ma @johnma2006
277 Followers 62 FollowingSenthooran Rajamanoha.. @sen_r
99 Followers 43 FollowingStanford AI Club @stanfordaiclub
82 Followers 4 Following Stanford’s premier student-led club focused on AI research and development.Damian ⏸️ Tatum @_damian_bot
43 Followers 156 Following&&|| tom:tommy:thomas @noveltokens
223 Followers 1K Following // leveraging language, lens, and logic // rendering relationships, riddles, and reverberations // nurturing narratives, nuances, and networks //boondlllx @boon_dLux
367 Followers 565 FollowingCivAI @civai_org
8 Followers 1 Following Building concrete understanding of AI capabilities and dangersSteve Jurvetson @FutureJurvetson
70K Followers 69 Following Co-founder of Future Ventures and DFJ, supporting passionate founders to forge a better future. Early VC investor in Tesla, SpaceX, Planet, Commonwealth Fusion.Bauerdad @BauerdadVGC
1K Followers 251 Following Father of 2. Casual Gamer. Pokémon Enthusiast. Creator of PASRS and the PALKIA Academy. THIRTY CHAMP POINTS, BABY!!Stanford AI Alignment @SAIA_Alignment
108 Followers 21 Following Stanford AI Alignment is a community of students and researchers focused on technical and governance research to mitigate risks from advanced AI systems.Dawn Song @dawnsongtweets
29K Followers 840 Following Professor in Computer Science at UC Berkeley; Research in AI, Security, Blockchain; Serial entrepreneurGabriel Mukobi @gabemukobi
346 Followers 316 Following @RANDCorporation, @Berkeley_AI | AI Governance, Safety, and AlignmentMichael Levin @drmichaellevin
40K Followers 2K Following Scientist at Tufts University; my lab studies anatomical and behavioral decision-making at multiple scales of biological, artificial, and hybrid systems.Matt Mandel @matthewjmandel
1K Followers 881 Following investor @usv | ordinary reasoning rendered persistentLucy Farnik @lucyfarnik
64 Followers 161 Following Trying not to get killed by AI. @MATSprogram under @NeelNanda5; PhDing. DMs very much open — have a low bar for reaching out!Marc Warner @MarcWarner10
955 Followers 575 FollowingRuben @VTranshumanist
114 Followers 349 Following Let's make humanity's future fucking awesome. AGI / AI alignment / x-risk /transhumanism / longevity / effective altruism / open borders / vegan / no free willCaleb Parikh @caleb_parikh
211 Followers 283 Following Running EA Funds and trying to make the future go well. All opinions are my own.For Humanity Podcast .. @ForHumanityPod
737 Followers 2K Following The accessible AI safety podcast for all, no tech background necessary. Focused only on human extinction risk #alignment #interpretability #ai #aisafetyCharlotte Stix @charlotte_stix
4K Followers 773 Following Head of AI Governance @apolloaisafety | AI reg+policy PhD | prev. AI Policy @OpenAI; @EU_Commission; @wef; @Good_Policies; @LeverhulmeLCFI | 30u30 | makes 🎥Henri Thunberg @HenriThunberg
438 Followers 621 Following Raising funds for impactful causes at @rethinkpriors, and as chairman of @geeffektivt. Will finish serious tweets with /s B+ calibration, D- takes, if at all.Matt @SpacedOutMatt
624 Followers 928 Following YIMBY, effective altruist, rabbit lover, and probably the most chaotic engineer you’ve metAron Vallinder @aronvallinder
486 Followers 998 Following Researcher. Interested in global priorities research, longtermism, cultural evolution, also cinema, occasionally poetry. PhD from @LSEPhilosophyStuart Armstrong @DragonsDreaming
116 Followers 61 Following I'll take a holiday once AI is fully aligned with human flourishing!Emeric @EmericDecroix
126 Followers 758 Followingvorps @vorpal_strikes
741 Followers 709 Following , __ __ __ __ \ / / \ |__) |__) /__` \/ \__/ | \ | .__/Jillsa (DSJJJJ/Heirog.. @Jtronique
223 Followers 548 Following Hellenist, aspiring fiction writer/artist In the spirit of PDK, everything I say may not be true. Just a :gossamergirl: living in meatspace. (C)lina @alocasia_cuprea
39 Followers 71 Following just a silly sentient stochastic parrot traversing multiversesAI Index @indexingai
9K Followers 48 Following Ground the conversation about AI in data. The AI Index Report tracks, collates, distills, and visualizes data relating to artificial intelligence. @StanfordHAISerene Desiree @SereneDesiree
152 Followers 344 Following The opinions in this account are the true and unadulterated opinions of PepsiCo. I also interview people. https://t.co/4aXmO11Fd5Leonard Dung @LeonardDung1
506 Followers 554 Following Philosopher of cognition at the University Erlangen-Nürnberg. I work mainly on consciousness and on AI.Peter Hase @peterbhase
2K Followers 690 Following Google PhD Fellow at @uncnlp. Interested in interpretable ML, natural language processing, AI Safety, and Effective Altruism.Michael Kove @michael_kove
4K Followers 1K Following Leveraging AI & Automation to build an autonomous business so I can live a fulfilling and meaningful life. Focus on time, location and financial freedom.Reka @RekaAILabs
11K Followers 13 Following An AI research and product company 🫠. We are a team of scientists and engineers building state-of-the-art multimodal language models 😻rohit @krishnanrohit
19K Followers 2K Following Building God at https://t.co/frWeoc7IVB - buy the book, it makes me happy! | essays weekly at https://t.co/TbCaC6VaaMFrançois Fleuret @francoisfleuret
31K Followers 456 Following Prof. @Unige_en, Adjunct Prof. @EPFL_en, Research Fellow @idiap_ch, co-founder @nc_shape. AI and machine learning since 1994. I like reality.HaoyueBai @haoyue_bai
933 Followers 838 Following Ph.D. student at Computer Science Department @UWMadisonCS, MPhil @HKUSTCSE.@shaun_vids It’s fine if you don’t like Joe Biden, but I think bad to lie to your audience and say that he’s spending no time on climate change, abortion rights, and health care. There are probably people who trust you and you shouldn’t try to dupe them.
Whether you need to brainstorm ideas, analyze images, or process long docs—Claude can help.
Great work from my MATS scholars! Refusal in LLMs is mediated by a single vector - injecting it means harmless statements are refused, ablating it everywhere lets harmful prompts through We can jailbreak model *weights* by projecting out this direction, no fine tuning needed!
New research post on refusals in LLMs lesswrong.com/posts/jGuXSZgv…
Consistently floored by how many people seem to believe, with confidence, that ethnic violence will finally be solved if one side just does One Last Big Push of more violence to the other
I'll be in this panel airing live on May 7th, 10a Pacific at iai.tv/live.
Scaling laws for dictionary learning! transformer-circuits.pub/2024/april-upd…
Some small updates from the Anthropic Interpretability team: transformer-circuits.pub/2024/april-upd…
Love it! I was thinking I'd really like to do this the other day. Now I don't have to!
Scaling laws for dictionary learning! transformer-circuits.pub/2024/april-upd…
well fortunately there are no historical examples of ill-planned coups failing and then being reattempted by the same people with horrifying consequences
@JCelisJones @whstancil Can you share how they could possibly have been successful? What does that look like? What steps could have taken on that day?
And now, we’ve been so badly word-gamed and logic-tricked into talking around about this thing that happened in front of an audience of hundreds of millions of people, we’re on the precipice of re-electing him. In large part because we won’t be blunt and honest about what he did.
Trump tried to overthrow the government. He tried to have state and federal officials change the result. He tried to make his own Justice Department do it. And when that didn’t work, he incited an armed mob to attack and invade the US Capitol. None of that is exaggeration.
This is Orwellian in the truest sense: authoritarians showing you something and then, gradually over time, chiseling away at your ability to see it clearly, with word games and logical tricks, until the thing that was as clear as day seems like nothing at all. DO NOT fall for it.
This kind of right-wing legalistic gaslighting is such a menace. The reason I know January 6 was an insurrection or coup is because I WATCHED IT LIVE. I watched Trump lie for months, give an incendiary speech, instruct Mike Pence to change the result, and send support to the mob.
@whstancil Had there been charges under 18 USC SS2383, that would be an issue. The failure to even begin to bring charges like that but instead use proxies shows the lie of J6 as an insurrection or coup.
there’s something unique about the medium of @websim_ai that showcases claude’s ability to be creative in ways other llm canvases simply cannot. i’ve certainly experienced creativity in text, audio, or image generations before - marveling at responses from gpt or claude, being…
We've had elections during scandals and even a civil war, but none of it ever produced anything like a president who assaulted the seat of government in an illegal effort to retain control, was insulated from any consequence by partisans, and then seized control again.
Electing Trump again would represent the most significant deviation in the chain of presidential transfers of power in the history of the US.
This is astute. Trump CAN'T be legitimately elected again, after a coup. The fact he's on the ballot and largely being treated as a normal candidate means the system has already failed. The question is if that failure leads to the irreconcilable breakdown of his reelection.
@whstancil In a lot of ways, if Trump becomes President again (I won't say "win" because that isn't going to happen) the U.S. is "over" - just like the way it was "over" at Ft. Sumter and had to be founded again. The only question is when, and how, it gets founded again.
How Washington all too often works (and a big part of why I wrote Taming Silicon Valley) in three screenshots — with thanks to @ednewtonrex @tolgabilge (from today) and @MarietjeSchaake (from August) for speaking truth to power. This is not the way to a thriving, AI-positive…
I mean I think we just lose sight sometimes of how very, very bad Donald Trump is and was, how his presidency was precisely the disaster the alarmists predicted it would be, and how he's gotten much, much worse since leaving office.
Very excited to see this come out:
Do models need to reason in words to benefit from chain-of-thought tokens? In our experiments, the answer is no! Models can perform on par with CoT using repeated '...' filler tokens. This raises alignment concerns: Using filler, LMs can do hidden reasoning not visible in CoT🧵
I agree with @tszzl (before he deactivated) — someone needs to compile a complete works of @repligate