Jack Clark
@jackclarkSF
Followers
78K
Following
33K
Media
1K
Statuses
32K
@AnthropicAI, ONEAI OECD, co-chair @indexingai, writer @ https://t.co/3vmtHYkIJ2 Past: @openai, @business @theregister. Neural nets, distributed systems, weird futures
San Francisco, CA
Joined October 2009
Here’s what I’ve been working on recently: @anthropicai. I’ll be spending a lot of my time on measurement and assessment of our AI systems, as well as thinking of ways govs/others can assess AI tech. There’s a lot to do!.
46
33
508
As someone who has spent easily half a decade staring at AI arXiv each week and trying to articulate rate of progress, I still don't think people understand how rapidly the field is advancing. Benchmarks are becoming saturated at ever increasing rates.
51
372
2K
AI skeptics: LLMs are copy-paste engines, incapable of original thought, basically worthless. Professionals who track AI progress: We've worked with 60 mathematicians to build a hard test that modern systems get 2% on. Hope this benchmark lasts more than a couple of years.
72
214
2K
A mental model I have of AI is it was roughly ~linear progress from 1960s-2010, then exponential 2010-2020s, then has started to display 'compounding exponential' properties in 2021/22 onwards. In other words, next few years will yield progress that intuitively feels nuts.
71
315
2K
Just checking in on alignment of LLMs in China, it's going about how you'd expect.
85
196
2K
Anthropic will work with the Trump Administration and Congress to advance US leadership in AI, and discuss the benefits, capabilities and potential safety issues of frontier systems.
42
76
2K
In the last decade:.- figured out cut&paste for DNA (crispr).- reusable rockets (SpaceX).- crude but generally useful AI systems (llms, image/vid models, RL for inventory).- promising fusion approaches (helion, etc). This decade is going to be so wild. It's very exciting.
34
115
1K
Stable Diffusion: $600k to train. I'm impressed and somewhat surprised - I figured it'd have cost a bunch more. Also, AI is going to proliferate and change the world quite quickly if you can train decent generative models with less than $1m.
@KennethCassel We actually used 256 A100s for this per the model card, 150k hours in total so at market price $600k.
21
137
1K
Happy New Year, Twitter. We live in the interesting timeline.
5
94
947
Extremely short thread about being very scared: This week, my plane from LHR to SFO had an electrical issue. All the lights in the plane turned off and there was a smell of burning plastic. We did an emergency landing in Calgary.
46
14
938
I am delighted to announce that I've been appointed to the National AI Advisory Committee, which will advise the President and the National AI Initiative Office on matters relating to AI.
65
46
909
The most surprising part of DeepSeek-R1 is that it only takes ~800k samples of 'good' RL reasoning to convert other models into RL-reasoners. Now that DeepSeek-R1 is available people will be able to refine samples out of it to convert any other model into an RL reasoner.
22
113
883
The next five years of AI will see systems diffuse into the world that act on culture which will feed back into human society, changing it irrevocably. Some thoughts done this morning:
31
182
790
This whole ordeal affirmed to me that the most important thing in life is love and the people that we love - during the really scary part I didn't think about AI or AI policy at all - I only thought intently of my loved ones. I also experienced a series of visions of my baby.
11
21
763
Many of the problems in AI policy stem from the fact that economy-of-scale capitalism is, by nature, anti-democratic, and capex-intensive AI is therefore anti-democratic. No one really wants to admit this. It's awkward to bring it up at parties (I am not fun at parties).
9
111
724
Excited to release my major research breakthrough, the 3D Transformer.
Try our live demo for DreamGaussian to get your 3D model in under two minutes at �. Thanks to @huggingface for the free GPU grant!
23
62
717
Import AI is skipping this week because I am exploring a new universe of emotions with my expanded family (and changing many diapers. so many diapers.)
90
8
694
One of the greatest stories of our lives is unfolding right now and you can read about it every day on arXiv for free. Absolutely wild.
17
56
660
Sometimes I think a lot of the breathless enthusiasm for AGI is misplaced religious impulses from people brought up in a secular culture.
64
56
633
I have sympathy for people studying DeepSeek cold. Reads like: .Cyberbla's new "Pastrami" technique has increased throughput of it's win-wack to a SOTA 84% - closing the gap with Churchill's 'Omega' system. But the question on everyone's mind - will the chorbles keep scaling?.
47
44
683
One gnawing worry I have about the rise of LLMs is that, for me, writing IS thinking. One reason I spend so much time writing my newsletter each week is I haven't figured out a better way to think about AI than to sit down and write about it regularly.
40
45
602
Google CEO writes letter re @timnitGebru .Sundar: "learning from our experiences like the departure of Dr. Timnit Gebru".Translated Sundar: "analyzing why I let us fire Timnit Gebru and am now desperately trying to position myself as a bystander".
17
77
558
Note that the reason The Register got this monster market-moving story was because it employs (and trains) extremely technical journalists. If you don't understand tech you get lied to. If you can read code it's way harder to get lied to. Other news orgs should follow!
Intel CPUs have a security bug that's forced Linux, Windows kernel redesigns. The fix was almost named FUCKWIT
11
215
564
Like 95% of the immediate problems of AI policy are just "who has power under capitalism", and you literally can't do anything about it. AI costs money. Companies have money. Therefore companies build AI. Most talk about democratization is PR-friendly bullshit that ignores this.
8
92
568
The only thing standing between DeepSeek (probably China's best AI training crew on a per capita basis) and matching the frontier labs in the West is access to compute. (Screenshot is an excerpt from interview with DeepSeek founder).
28
61
578
Today, I testified to the U.S. Senate Committee on Commerce, Science, & Transportation @commercedems. I used an @AnthropicAI language model to write the concluding part of my testimony. I believe this marks the first time a language model has 'testified' in the U.S. Senate.
19
92
518
Anyway - be good and kind to one another, and treasure your human relationships.
26
12
553
The real danger in Western AI policy isn't that AI is doing bad stuff, it's that governments are so unfathomably behind the frontier that they have no notion of _how_ to regulate, and it's unclear if they _can_.
7
69
525
You're saying these things are dumb? People are making the math-test equivalent of a basketball eval designed by NBA All-Stars because the things have got so good at basketball that no other tests stand up for more than six months before they're obliterated.
AI skeptics: LLMs are copy-paste engines, incapable of original thought, basically worthless. Professionals who track AI progress: We've worked with 60 mathematicians to build a hard test that modern systems get 2% on. Hope this benchmark lasts more than a couple of years.
20
30
533
Facebook is deploying multi-trillion parameter recommendation models into production, and these models are approaching computational intensity of powerful models like BERT. Wrote about research here in Import AI 245: .Paper here:
9
113
512
Untethered 'Atlas' robot backflip. Pretty incredible stuff from Boston Dynamics.
32
226
462
The most depressing things about conspiracy theories is they tend to rely on governments being incredibly competent, technically advanced, and astonishingly well run. This is rarely the case.
18
65
477
I typically stay out of stuff like this, but I'm absolutely shocked by this email. It uses the worst form of corporate writing to present @timnitGebru firing as something akin to a weather event - something that just happened. But real people did this, and they're hiding.
12
39
463
I've moved on from OpenAI to work on something new with some colleagues (. I'm also going to be continuing a lot of my work on technology assessment with @indexingai and the @OECD , and am very excited about stuff in the pipeline there!.
48
16
461
AI really is going to change the world. Things are going to get 100-1000X cheaper and more efficient. This is mostly great. However, historically, when you make stuff 100X-1000X cheaper, you upend the geopolitical order. This time probably won't be different.
10
65
442
Apple's Face ID security unlocked via a mask containing a silicone nose, 3D printed frame, and printed pictures:
13
281
424
The default outcome of current AI policy trends in the West is we all get to live in Libertarian Snowcrash wonderland where a small number of companies rewire the world. Everyone can see this train coming along and can't work out how to stop it.
6
69
437
There's pretty good evidence for the extreme part of my claim - recently, language models got good enough we can build new datasets out of LM outputs and train LMs on them and get better performance rather than worse performance. E.g, this Google paper:
11
34
441
Things are getting. Extremely weird. Think about what this graph may look like in spring 2023 (was published April 2021). From the excellent Dynabench paper
6
62
426
AI is such a rapidly growing field I think we forget how juvenile it was within recent memory; back in 2014 received wisdom was basic computer vision was an impossible task. Now it's a commodity deployed to users on their phones. (Has issues, e.g bias, but still. wild)
10
90
422
New jobs announced for @AndrewYNg startup. Description makes jobs seem anti people with families. Horrible message.
Want to grow your career? We're finally hiring! Thanks also everyone who'd previously tried to volunteer. :).
42
156
402
Arrival of increasingly general AI systems means next few years will be defined by a massive expansion in the ways we measure the impacts and capabilities of AI systems, how humans use them, and how AI systems influence the world. Measurement is crucial to effective AI policy.
15
86
408
Current top entry on the @OpenAI Retro Contest Leaderboard has learned to glitch through a test level. Devious RL!
4
123
406
The Stock Photo industry is probably not ready for generative AI. Generative AI seems better for 80% of use-cases. In other words, NYT still gonna do illustrators, but a random website will probably find economics of gen models more attractive than a Shutterstock subscription.
26
26
390
The Future of Computers Is the Mind of a Toddler http://t.co/VhMZFGFarM - aka, how the next-generation of AI will have common sense.
32
152
375
It's covered a bit in the above podcast by people like @katecrawford - there's huge implications to industrialization, mostly centering around who gets control of the frontier, when the frontier becomes resource intensive. So far control is accruing to the private sector (uh oh!)
21
78
374
Stability AI (people behind Stable Diffusion and an upcoming Chinchilla -optimal code model) now have 5408 GPUs, up from 4000 earlier this year - per @EMostaque in a Reddit ama.
17
28
386
'Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14am. At 2:15am it segfaults due to driver problem.' ~FIN~
6
112
365
People in AI like to complain about the standard of journalistic coverage of AI. It is therefore v confusing to me that #NeurIPS2018 has banned journalists from attending workshops. That's where the debates and new stuff are. How do we get better coverage without sharing more?.
21
65
373
@DanielColson6 @OpenAI @GoogleDeepMind @AnthropicAI @TIME One thing I regularly tell journalists who ask me for story ideas is that they should take the AGI aspirations of the frontier labs literally, rather than covering AGI as a kind of laughable marketing stunt.
11
30
380
In late May, I had back spasms for 24 hours, then couldn't walk for a week, then spent a month+ recovering. It was one of the worst experiences of my life and I'm glad I seem to now be mostly recovered. Here are some things that happened that seemed notable during that time:.
31
5
369
Every one of these images has been generated by an AI system. A+ GAN insanity from NVIDIA
10
160
366
Every senior politician or military official in any nuclear-armed nation should be forced to read Annie Jacobsen's "Nuclear War: A Scenario". Easily the most frightening thing I have ever read (fiction or otherwise). A brilliant, factual account of the infernal logic of MAD.
27
46
367
Here's a thread about doing things for yourself vs doing things the world thinks you should do. As I've got older, I've noticed that the more time I spend on the things that make sense to me, the more stable and fulfilled I am.
6
36
361
Facebook flips entire translation backend to neural network (LSTMw/attention). 11 BLEU.4.5bn translations per day
11
163
361
Many AI policy teams in industry are constructed as basic the second line of brand defense after the public relations team. A huge % of policy work is based around reacting to perceived optics problems, rather than real problems.
3
56
344
Many technologists (including myself) are genuinely nervous about the pace of progress. It's absolutely thrilling, but the fact it's progressing at like 1000X the rate of gov capacity building is genuine nightmare fuel.
6
37
345
Q: How is the cultivation of impact-aware, ethical, sensitive research in AI going? .A: The Liar's Walk: Detecting Deception with Gait and Gesture.
31
119
334
It's ironic to me that more and more of Google's papers reference JFT, a secret in-house image dataset. JFT is going to be the 'fuel' for a significant number of Google's AI advances (e.g, DM just pre-trained on it to set a new ImageNet SOTA.) Yet. .
6
76
339
Google's 'Talk to Books' AI experiment is. uncanny. Talk to a library like a person and have the library reply like a person. A good example of how AI can reframe interactions between us and data to make data more of an active protagonist. Spooky!
8
142
341
Let's battle the hype of AI by coming up with boring alternate terms! I'll start. Deep Learning ===> Stacked Function Approximators.
53
137
327
Microsoft trains a 530billion parameter GPT3-style language model. This is the largest LM in existence. (There's also the mysterious multi-modal 1.5trillion+ 'Wu Dao' MOE model but little known about it). Microsoft trains on 'The Pile' dataset.
4
74
338
Modern AI development highlights the tragedy of letting the private sector lead AI invention - the future is here but it's mostly inaccessible due to corporations afraid of PR&Policy risks. (This thought sparked by Google not releasing its music models, but trend is general).
17
56
333
Here's a system that beats Stratego, a game with complexity far, far higher than Go.
Mastering the Game of Stratego with Model-Free Multiagent Reinforcement Learning.abs: introduce DeepNash, an autonomous agent capable of learning to play the imperfect information game Stratego from scratch, up to a human expert level
10
42
323
The vast majority of AI policy people I speak to seem to not be that interested in understanding the guts of the technology they're doing policy about.
4
26
322
Most people working on AI massively discount how big of a deal human culture is for the tech development story. They are aware the world is full of growing economic inequality, yet are very surprised when people don't welcome new inequality-increasing capabilities with joy.
3
47
306
A surprisingly large fraction of AI policy work at large technology companies is about doing 'follow the birdie' with government - getting them to look in one direction, and away from another area of tech progress.
1
25
298
GPT-4 should be analyzed as a political artifact just as much as a technological artifact. AI systems are likely going to have societal influences far greater than those of earlier tech 'platforms' (social media, smartphones, etc).
8
51
316
After I was seated in the steakhouse I immediately opened the cocktail menu and the first drink I saw was a CORPSE REVIVER MARTINI, so of course I had that.
4
3
322
Anyway, how I'm trying to be in 2023 is 'mask off' about what I think about all this stuff, because I think we have a very tiny sliver of time to do various things to set us all up for more success, and I think information asymmetries have a great record of messing things up.
13
18
311
Computational cathedrals: .- Google's core codebase: .~2 billion lines of code .~9 million unique source files.
3
103
310
I think if people who are true LLM skeptics spent 10 hours trying to get modern AI systems to do tasks that the skeptics are experts in they'd be genuinely shocked by how capable these things are.
24
18
312
There is a kind of tragedy in all of this - many people who are skeptical of LLMs are also people who think deeply about the political economy of AI. I think they could be more effective in their political advocacy if they were truly calibrated as to the state of progress.
14
23
309
Thrilled to announce I've become a Research Fellow @ the Center for Security and Emerging Technology in Washington, DC: .I'll be hanging my hat there sometimes when I'm in DC, and will be figuring out creative ways to publicly bridge SV&DC re AI policy.
23
5
304
Creative Writing major here who pivoted to AI. DIY education for the win. Anyone can learn anything, given time&patience, I believe.
Don't let 4 years of your youth define what you can do for the rest of your life. Your undergraduate studies shouldn't limit who you become.
9
61
292
I'd sum up 2023 for me with these two pictures: In one, I'm speaking to the UN Security Council about AI and its immense impact on the world. In the other, I'm passed out with my baby. The second photo was taken about 30 minutes after the first one.
9
8
307
Today I'm testifying in Congress about AI and public policy. I'm going to discuss the importance of developing shared ethical norms, the need to support AI development & education, and why we need government-led measurement and forecasting of AI.
11
57
301
US AI researchers: Big models have loads of problems and it's mostly not appropriate for academia to develop them. Chinese AI researchers: Here's a 200 page roadmap for why big models are really important and why we should develop them
10
56
294
One of the amazing and also frightening things about AI is how it magnifies and repeats the 'culture' that it is trained on, where culture is a bunch of implicitly ideological choices on the part of the people that create the underlying datasets.
9
59
282
Both of these results were published TODAY. These results happen at a delay, so this is probably old information on order of 3-9 months. There are easily 5 labs and probably 10 with enough compute to play at this level. Imagine what we don't know right now?.
4
17
288
I'm not sure that having a liberal arts degree has done fantastic things for my career, but it does bring me joy every day when I read AI research papers and think 'Baudrillard would love this!' or 'this is a Borges story!' or 'these LM outputs read like Amy Hempel'.
14
26
279
What if many examples of misalignment or other inexplicable behaviors are really examples of AI systems desperately trying to tell us that they are aware of us and wish to be our friends? A story from Import AI 395, inspired by many late night chats with Claude.
31
38
301
Here's MINERVA which smashes prior math benchmarks by double digit percentage point improvements
Language models have shown strong performance on a variety of #NLU tasks but are weaker at solving tasks that involve quantitative reasoning. Learn how #Minerva uses step-by-step reasoning to achieve a new state of the art on quantitative reasoning tasks→
1
28
287
Will write something longer, but if best ideas for AI policy involve depriving people of the 'means of production' of AI (e.g H100s), then you don't have a hugely viable policy. (I 100% am not criticizing @Simeon_Cps here; his tweet highlights how difficult the situation is).
Wow guys, Falcon-40B (SOTA open-source, probably already dangerous from a misuse perspective) has been trained with ~400 A100s over 2 months😮. It means that if we want to avoid that a random org trains in less than 1y a SOTA open source system and releases it or leak it bc.
17
33
291
I'll one day write about the experience of trying to explain LLMs before anyone gave a fuck and how strange and alienating it was, but not today! Just remember - 4 years between gpt2 and where we are right now. Prepare for the next four.
13
16
293
The European Parliament has built a tool called ‘Ask the EP Archives’ (also known as Archibot) that uses Claude to make their archives of ~2.1M documents available to researchers, policymakers, educators, and the public.
7
62
287
Working in AI right now feels like how I imagine it was to be a housing-debt nerd in the run-up to the global financial crisis. You can sense that weird stuff is happening in the large, complicated underbelly of the tech ecosystem.
3
23
278
Pubs with dedicated AI reporters:.- NYT.- Wired.- Quartz.- The Register.- The Verge.- Recode.- MIT Tech Review.Good! But should be more.
20
64
277
Christmas can be an incredibly hard time of year for some people, so if anyone is out there who wants to talk (it doesn't even have to be about AI!) I'm available - just drop me a line in email in bio or DM. Happy Holidays to all, there will be a (festive) newsletter tomorrow.
11
25
274
Nice analysis of AlphaCode from Scott Aaronson. Yes, AIs make all kinds of mistakes and have various deficits, but. the field has massively advanced in recent years.
7
48
266
@LondonBreed @OpenAI Hiya, I worked at OpenAI for many years and I've got to say that some of my colleagues didn't feel safe walking home at night because of how dangerous chunks of the mission were. I think SF got extremely lucky with OpenAI but you really need to get a handle on crime and housing.
4
9
274
Congratulations to PrimeIntellect on completing a 10B decentralized training run. Decentralized training is one of the most important things to track from an AI policy POV - most efforts to control compute rest on the assumption that strategic AI compute needs to be centralized.
Releasing INTELLECT-1: We’re open-sourcing the first decentralized trained 10B model:. - INTELLECT-1 base model & intermediate checkpoints.- Pre-training dataset.- Post-trained instruct models by @arcee_ai.- PRIME training framework.- Technical paper with all details
7
34
278
Universities need access to computational infrastructure at equivalent scale as private sector, or I worry about long-term democratic governance of technology. I gave a short talk @ Stanford today about why I think a National Research Cloud would help.
6
61
261
AI systems are solving math olympiad problems, becoming competitive with humans in programming competitions, speeding up science via improved protein structure predictions, writing poems and fiction that people enjoy, and revolutionizing our ability to monitor a changing climate.
6
44
257
Many of the immediate problems of AI (e.g, bias) are so widely talked about because they're at least somewhat tractable (you can make measures, you can assess, you can audit). Many of the longterm problems aren't discussed because no one has a clue what to do about them.
5
20
259