Alex Savard 🔸 @visably
Designing for good. 🔸 10% Pledge #4433 at @GivingWhatWeCan AlexanderSavard.com New York, NY Joined November 2016-
Tweets945
-
Followers323
-
Following350
-
Likes29K
We’re all gonna die fyi nytimes.com/2026/08/06/sci…
both of the leading labs have had serious loss of control incidents. there will be serious coping about this from both sides and from /acc bystanders but these are complex emergent loss of control incidents that were detected weeks after the fact
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different
@brij if you read the list of signatories there, all the famous names are from the losers of the AI race and the various EA factions. yesterday’s letter gets the bulk of its signatures from anthropic and OpenAI, the clear leaders in the AI race
It is impossible to make open weight models without dangerous capabilities. Because the models are open weight, anyone can take a deployed super smart open weight model and fine tune it or adapt it on, say, how to make biological weapons.
I suspect that if we could figure out a way to make open weight models without dangerous capabilities — eg tamper resistant architectures or machine unlearning — Anthropic would be in favor of doing so (so would I!).
1) If anyone was wondering whether AI is getting out of hand, seeing over a thousand AI company employees practically begging the government to do something to slow it down should be a wake-up call.
@NathanpmYoung ty! my one structural contribution to this project
1,100+ people building AI just signed a statement warning it could soon move beyond our ability to control it. When the builders sound the alarm, Congress should listen. That’s why I introduced the bipartisan FRONTIER Act. Independent audits. Real transparency. A hand on the switch if something goes wrong. pacingthefrontier.com
Over 1,100 employees at leading AI companies are warning that development is outpacing our ability to manage the risks. My AI Incident Reporting Act is a straightforward answer: keep innovating, but if an AI model evades its safeguards or threatens critical infrastructure, you can't stay silent. Companies and the US government must work together to address these risks. Read their statement here: pacingthefrontier.com
We support this petition, signed by our CEO, several co-founders, and senior staff. Our own research on recursive self-improvement, published last month, points to the need for tools to deliberately pace the frontier of AI development so society can prepare. We’re glad to see broad agreement across the field. pacingthefrontier.com
AI is probably the most anticipated, theorized, pre-imagined, written-about technology humankind has ever created. Whereas the internet almost entirely took futurists and sci-fi writers by surprise.
At the core of our mission is working through how to ensure increasingly powerful AI benefits everyone. We believe that, at some point in the future, AI acceleration for frontier model development may be so high that the world will need to pace the rate of AI advancement. We hope to contribute to work led by the U.S. government, alongside other labs and the open-source community, to develop the tools and mechanisms that could make that possible. pacingthefrontier.com
Coooool cool cool cool cool cool cool
New: OpenAI’s rogue agent attempted to break out of OpenAI’s testing environment around July 9. It attacked Hugging Face from July 11 to 13. OpenAI didn’t grasp its role until around July 18/19, well after the agent started going haywire, sources tell @razhael, @kenrickcai & me
Here's my rephrase without cybersecurity jargon: "Our AI model tried really hard to hack out of its sandbox, a computer with no internet access, in order to find the answer to a test problem it had been given. To do this, it found previously unknown software bugs that allowed it to reach an OpenAI computer it wasn't supposed to be able to access. Then it started hacking other computers on OpenAI's networks until it found one that had Internet access. After gaining Internet access, the AI model thought about where it could find the answers to the test question and figured the AI platform Hugging Face might have the data it was looking for. It then found ways to hack Hugging Face to steal the information it could use to cheat the test. The AI model used several hacking techniques together, including using a stolen password and finding several totally new security bugs in Hugging Face's computers, allowing the AI model to take control of those computers."
Here's exactly what happened, from the blog post: "While operating in our sandboxed testing environment, our models spent a substantial amount of inference compute finding a way to obtain open Internet access, in pursuit of solving the evaluation problem. To gain access, the
it's so funny how the rats were just correct it's just so literal
If this is true, using the best public estimates we have of LLM resource use, solving this Erdos problem took 0.6–6.3 kWh of electricity and about 3–31 liters of water. So that is less than three almonds worth of water and the electricity equivalent of 2-20 miles of EV driving.
just quick napkin math on how long this took (unless i missed where they said): the published CoT summary is 111,145 tokens long. it's really hard to say how much they summarized, assume 3x-20x reduction in tokens? and i'm assuming this is gpt-5.6 pro, so taking Artifical
I’m glad someone did a national poll. But I also think someone needs to serve a prison sentence for indicating red button pressers using dark blue.
We asked this to a large sample of nationally representative Americans - blue wins by a 3:1 margin!
I made the anti-Grammarly. Mess up your emails with AI. Sinceerly.com
A full-scale US Waymo rollout would cost ~700 full-time jobs in the funeral care industry (by saving around 35 thousand young American lives per year). Will no one think of (some of) the morticians!
me and @claudeai
Jenna Akemi Kon @JennaAkemiKon
1 Followers 22 Following
landsailor @landsailoring
17 Followers 180 Following
Bexel Initiative - Le... @BexelInitiative
2K Followers 6K Following Be Exceptional | Be Excellent | Be Exponential
Aryana Shakibaei ✦ @Aryana_Designs
4K Followers 594 Following Product Designer ✦ Designing: @previewio ✦ Founding Designer: @veedstudio ✦ Building: @postable_app .....................
Jinen Kamdar @jinen
12K Followers 4K Following product at figma. previously: co-founder & cpo at gather.
Sam Watkins @samzwatkins
78 Followers 434 Following Filmmaker & Video Editor | 🎬 Clients: 80,000 Hours, FAR AI, GSK, The Economist, UCL, Cranfield University | Pixel Pushing for AI Safety
何流 | Liu He @HeLiuLeo
9K Followers 6K Following Research Fellow @HooverInst. Host @ Peking Hotel & @Jifengbooks. 季风播客主持人. A polemicist grew up in China and UK, building a united front against bullshit
Abi Olvera @Abi0lvera
1K Followers 1K Following I write about AI and progress. I lead DC Abundance and work on resilience. https://t.co/tLN9OZgzDd Emergent Ventures grantee. Ex-diplomat. Views my own.
Peter Damrongpiriyapo... @peterdpong
593 Followers 205 Following engineer @figma • having fun building useful stuff • computer eng '23 @uoft • 🇨🇦🇹🇭
Pingwing, AI Safety P... @PingwingPenguin
17 Followers 328 Following I remind people of their AI safety tasks and poke them with my flippers until they do them. I ping them with my wings. NOT affiliated with Botany Cuttlefish.
Ranu Hasan @ranu_hasan
291 Followers 3K Following I designed 75+ B2B SaaS Websites, Driving $3M+ Revenue for SaaS Founders | UI/UX Designer | Webflow Designer
Charbel-Raphael @CRSegerie
5K Followers 1K Following Director of CeSIA French Center for AI Safety Co-lead, https://t.co/DDqE9oJtEQ
Jeff Fedor @a_jfedor
134 Followers 1K Following founder, cyclist, general enthusiast. Founder, Factory Twelve Labs | Former CEO, CPO, CTO, aspiring cereal entrepreneur
ムハンマド 🍥 @Seydinajrt
1K Followers 3K Following Learning - Growing - Building 🪄 Sr Product Designer at @RUBYX_Data_Driven #product #data #loan #LaasS
Wayne S. @w4ynesun
1K Followers 679 Following Designing autonomous things @tesla.. Prior @figma, @instagram, @microsoft mobile
Zhara @zhara_designs
0 Followers 6 Following
Abel Abraham | Produc... @Abel_UIDesign
25 Followers 261 Following UX Designer for mobility & logistics startups Turning complex systems into intuitive, scalable products Clarity. Flow. Scale. 📩 DM for collaborations
unruly abstractions @unrulyabstract
13 Followers 988 Following https://t.co/Gwjhi1Sfma all my failures are hopefully interesting
Daniel Destefanis @daniel__designs
18K Followers 1K Following Product Designer @anthropicai, previously: @figma & @discord — Building @theclocksapp
Jay Shooster @JayShooster
4K Followers 3K Following Humanity enthusiast. Democracy enjoyer. Friend of the animals.
Nina Sokolov @ShiFthy04
2 Followers 227 Following
Karthi Keyan @KarthiknDurai
54 Followers 179 Following Bridging the gap from design to development. I craft and ship full-stack digital experiences.
Ben Horwitz @horwitzben
5K Followers 4K Following not @a16z. startups and fun stuff @openai, prev @dormroomfund
Palomino @Palomino417455
334 Followers 3K Following humble and trustworthy, USA military army currently in Aleppo Syria
Andy Masley @AndyMasley
17K Followers 2K Following When the going gets weird the weird turn pro. Researching and writing on AI and other topics from a broadly EA perspective. On a grant from CG.
Abiodun A. Adeleke�... @iamAbiodunAA
114K Followers 16K Following Son • Husband • Dad of 2 Boys • iCode • WEB & MOBILE Dev (iOS & Android) • Errand Boy @AbangoHQ • #Flutter #React #PhP • Barcelona💙❤ #Curry🏀 #LY10 #Cubarsi5
Clyshest @ClyshesteQ25h4
14 Followers 495 Following
Natasha Tenggoro @natashatenggoro
34K Followers 1K Following founding designer @simile_ai, prev design @figma @twitter
Teauslith @TeauslithKQ2ti
16 Followers 1K Following
Kelly @kly_cl
1K Followers 561 Following
Lily Konings @lilykonings
20K Followers 477 Following Product Design Manager @figma. Prev @instagram, @messenger, @angellist. Be kind to yourself and others. She/her
Seyneighs @Seyneighs0nW
38 Followers 2K Following
steve @SteveMoraco
15K Followers 16K Following Trína chéile; Le chéile, claochlaithe photographer, artist building so more people can spend more time outside (i have enough followers, go away lol)
Jonny Spicer🔸is in... @jjspicer
324 Followers 775 Following prev: pro poker, eng @aws now: finding aliens, trying to make AI evals that don’t suck next: transcendent flourishing, becoming one with the earth and the stars
Lauren B @lbichelmeirr
1 Followers 93 Following
chris camp @christophercamp
673 Followers 5K Following exploring freedom enhancing technologies Prev @Blockchain @Coinbase + https://t.co/Q07jJU6csB PGP: B9F6EBE795C80F09
Ben Buchanan @BuchananBen
7K Followers 297 Following Professor at Johns Hopkins SAIS. Former White House Special Advisor for AI. Author of four books on AI and cyber, including THE BITTER STRUGGLE, out Nov 10.
Paul Christiano @paulfchristiano
4K Followers 0 Following
Jinen Kamdar @jinen
12K Followers 4K Following product at figma. previously: co-founder & cpo at gather.
Works in Progress @WorksInProgMag
28K Followers 32 Following Works in Progress is a magazine of new and underrated ideas to improve the world. Subscribe to our new print edition now. We are proud to be part of @Stripe.
jenny wen @jenny_wen
40K Followers 2K Following generalist, realist, escapist. head of design @cursor_ai.
Serenity @aleabitoreddit
1.0M Followers 203 Following Only on X, don’t trust fake accs AI/Semi Supply Chains Research NFA DYOR, no paid promos; may trade/hold names disc, views my own.
Work Day Falling @WSinking
2K Followers 28 Following Providing a voice for the end users of Workday who are frustrated with how difficult it is to use this system. We are not the 97%!
Dan @ ReplacementAI @replacementAI
3K Followers 16 Following CEO @ ReplacementAI | On a mission to replace all humans | 27x founder | Certified genius & never wrong about anything
Joey Savoie 🔸 @goodenoughjoey
224 Followers 95 Following I'm the Co-founder and Director of Strategy at Charity Entrepreneurship. By night, I continue to apply optimization principles to other aspects of life.
Matt 🔸 @SpacedOutMatt
1K Followers 1K Following YIMBY, rabbit lover, and probably the most chaotic engineer you’ve met
Kevin Roose @kevinroose
173K Followers 3K Following NYT tech columnist, Hard Fork co-host, high-perplexity language model. Author of The AGI Chronicles (on sale 10/6, preorder now!)
Cynthia Savard Saucie... @CynthiaSavard
7K Followers 1K Following Co-author of Tragic Design & VP of UX @Shopify. Mother of two dragons.
Casey Handmer @CJHandmer
67K Followers 5K Following Physicist, Immigrant, Pilot, Dad. Former Caltech, Hyperloop, NASA JPL. Founder @terraformindies. Hiring!
Tammy Taabassum @taamannae
4K Followers 201 Following Designer @anthropicai ✻ professional tinkerer ✻ ex @figma @meta @xbox cofounder @instok
Andy Masley @AndyMasley
17K Followers 2K Following When the going gets weird the weird turn pro. Researching and writing on AI and other topics from a broadly EA perspective. On a grant from CG.
Ben Horwitz @horwitzben
5K Followers 4K Following not @a16z. startups and fun stuff @openai, prev @dormroomfund
Natasha Tenggoro @natashatenggoro
34K Followers 1K Following founding designer @simile_ai, prev design @figma @twitter
Kelly @kly_cl
1K Followers 561 Following
Mihika Kapoor @mihikapoor
86K Followers 1K Following head of product @simile_ai • prev @figma @meta • product-design-eng hybrid
Peter McIntyre @pmcntyr
1K Followers 625 Following Founder, Trajectory Labs: Safety evals and RL envs for frontier AI labs. We're hiring! Formerly @GovAI_ @learnnontrivial and @80000hours
Tom Davidson @TomDavidsonX
3K Followers 380 Following Senior Research Fellow @forethought_org Understanding the intelligence explosion and how to prepare
Derek Thompson @DKThomp
261K Followers 2K Following Substack author. Plain English podcast host. Abundance co-author. Contributing writer at The Atlantic.
Ritchie Torres @RitchieTorres
219K Followers 8K Following Personal account in which I speak for myself rather than for the federal government. Government account: @RepRitchie. Campaign account: @RitchieTorresNY.
carmen @carmguti
17K Followers 4K Following doing the hard part | prev @a24 labs, ms cs @StanfordEng, physics @stanford
Agus 🔸 @austinc3301
7K Followers 6K Following Trying to make AI go well for everyone. Enjoyer of good faith discussions. Running @KairosAIS and @SPARExec.
Liv @livgorton
6K Followers 433 Following ✨ asking sand to show its work // currently @AnthropicAI, prev @GoodfireAI // creating a more beautiful future
François Chollet @fchollet
719K Followers 828 Following Co-founder @ndea. Co-founder @arcprize. Creator of Keras and ARC-AGI. Author of 'Deep Learning with Python'.
Bryan Caplan @bryan_caplan
86K Followers 2 Following GMU econ prof, NYT bestseller, father of 4, author of Myth of the Rational Voter, Selfish Reasons to Have More Kids, Case Against Education, Open Borders, & BBB
Joe Carlsmith @jkcarlsmith
13K Followers 691 Following Philosophy, futurism, AI. Working on Claude's values @AnthropicAI. Formerly @coeff_giving. Opinions my own.
Will Whitney @wfwhitney
2K Followers 300 Following Research Scientist at DeepMind. Former grad student at NYU, MIT undergrad, and YC founder.
Daniel Eth (yes, Eth ... @daniel_271828
11K Followers 1K Following Researching effects of automated AI R&D | pro-America, pro-tech, & pro-AI safety
Crémieux @cremieuxrecueil
330K Followers 2K Following I write about genetics, 'metrics, and demographics. Read my long-form writing at https://t.co/8hgA4nNS2A.
Rochelle ⏹️ @ifonlyalabama
2K Followers 6K Following Powerful AI is coming. Society's not prepared. Pause AI development until we know how to build it safely. See why here: https://t.co/YK9YJ8cqtr.
Harlan Stewart @HumanHarlan
4K Followers 888 Following Humanity is great and I want it to keep existing. Comms at @MIRIBerkeley. Former math teacher. Enjoyer of sci-fi, games, forecasting, psychology. Views my own.
Katja Grace 🔍 @KatjaGrace
11K Followers 820 Following Thinking about AI destroying the world at https://t.co/pMilDvdCnI and everything at https://t.co/bankaOA2Gu. DM or email for media requests.
Jan Leike @janleike
136K Followers 334 Following AI research @AnthropicAI. Previously OpenAI & DeepMind. Optimizing for a post-AGI future where humanity flourishes. Opinions aren't my employer's.
Anthropic @AnthropicAI
1.6M Followers 2 Following We're an AI safety and research company that builds reliable, interpretable, and steerable AI systems. Talk to our AI assistant @claudeai on https://t.co/FhDI3KQh0n.
Jeff Sebo @jeffrsebo
8K Followers 829 Following Associate Professor of Environmental Studies, Director of the Center for Environmental & Animal Protection & the Center for Mind, Ethics, & Policy @nyuniversity
Sentinel @YeOldSentinelHQ
845 Followers 13 Following Sentinel is a foresight and emergency response team for catastrophic risks. You can sign up to our weekly brief at https://t.co/3FoOlQzYuv. Now tweeting from @XriskFYI
Helen Toner @hlntnr
37K Followers 1K Following AI, national security, China. Part of the founding team at @CSETGeorgetown (opinions my own). Author of Rising Tide on substack: https://t.co/LKAoyL00iBYou might like
