XOutlier
Download .zipAdd to Chrome

Jason Wei

@_jasonwei115k followers

ai researcher @meta, past: openai, google 🧠

Open on X
Top outlier
150×
Best organic post
Peak views
2.8M
Single post
Bookmarks
42k
Across 99 organic posts
Last scan
Oct 16
Community cache
Posts 99 posts
#DatePost↓ OutlierViewsLikesRepostsBookmarks
1Dec 13, '24Yall heard it from the man himself150×2.8M3.9k4241.2k
2Oct 19, '23Enjoyed visiting UC Berkeley’s Machine Learning Club yesterday, where I gave a talk on doing AI research. Slides: In the past few years I’ve worked with and observed some extremely70.4×1.3M1.4k2521.7k
3Feb 20, '24My typical day as a Member of Technical Staff at OpenAI: [9:00am] Wake up [9:30am] Commute to Mission SF via Waymo. Grab avocado toast from Tartine [9:45 am] Recite OpenAI charter.48.9×909k3.6k2911.2k
4Dec 20, '24o3 is very performant. More importantly, progress from o1 to o3 was only three months, which shows how fast progress will be in the new paradigm of RL on chain of thought to scale 32.8×610k1.7k198368
5Sep 7, '23One thing that I started doing at OpenAI is that I created a policy for myself to be *100% transparent* with my manager about everything. It seems obvious and weird to say aloud, b32.4×602k2k1581.5k
6Feb 13, '24An incredible skill that I have witnessed, especially at OpenAI, is the ability to make “yolo runs” work. The traditional advice in academic research is, “change one thing at a tim28.8×535k2.4k2021.1k
7Feb 14, '23Personal update: after two years at @Google Brain, I joined the #ChatGPT team at @OpenAI! Excited to continue working on large language models and can't wait to see the impact of A28.4×528k3.4k117172
8Sep 12, '24Super excited to finally share what I have been working on at OpenAI! o1 is a model that thinks before giving the final answer. In my own words, here are the biggest updates to the28.4×527k3.3k3331.2k
9Dec 6, '24Kickstarting my career in comedy after all this AGI stuff is over. And yes, this was actually the first time Sam, Hyung Won, and Max heard the joke25.8×479k1.5k67269
10Jul 16, '25New blog post about asymmetry of verification and "verifier's law": Asymmetry of verification–the idea that some tasks are much easier to verify than to solve–is becoming an import21.4×397k1.6k2511.5k
11Jul 16, '25Becoming an RL diehard in the past year and thinking about RL for most of my waking hours inadvertently taught me an important lesson about how to live my own life. One of the big 21.2×395k3.5k3582.1k
12Nov 18, '23From @sama I learned a lot about clear thinking and challenging conventional beliefs. @gdb inspires me to work harder and be a good person. When I first joined, I was working at th21×391k1.7k73205
13Jun 30, '25We don’t have AI self-improves yet, and when we do it will be a game-changer. With more wisdom now compared to the GPT-4 days, it's obvious that it will not be a “fast takeoff”, bu21×390k1.4k164682
14Jan 15, '25Magic is what happens when an unstoppable RL optimization algorithm powered by sufficient compute meets an unhackable RL environment20.9×388k1.6k125400
15Mar 16, '23I’m hearing chatter of PhD students not knowing what to work on. My take: as LLMs are deployed IRL, the importance of studying how to use them will increase. Some good directions I20.7×386k1.6k2741k
16Oct 2, '23One pattern I noticed is that great AI researchers are willing to manually inspect lots of data. And more than that, they build infrastructure that allows them to manually inspect 20.6×382k1.7k195648
17Aug 12, '24Inspiring words from a young OpenAI engineer: “Why have I done well so far? I don’t think I’m smarter or more experienced than other people. But my competitive advantage is that I 20.3×377k2.2k163985
18May 16, '24One liberating thing about OpenAI (and presumably other small companies) is that there are no expectations of project scope being tied to your level. What I mean is that an ambitio20.2×375k60147221
19Jul 10, '23Moving from Google Brain to OpenAI, one of the biggest changes for me was the shift from doing individual/small-group research to working on a team with several dozen people. Speci19.9×369k1.5k144692
20Aug 5, '23Yann LeCun is obviously a legend but I found this tweet to be quite misinformed. The whole point of "emergent abilities" such as few-shot prompting and chain-of-thought prompting, 19.5×363k1.3k141550
21Jan 25, '23Yesterday I gave a lecture at @Stanford's CS25 class on Transformers! The lecture was on how “emergent abilities” are unlocked by scaling up language models. Emergence is one of th19.4×360k2.1k2941.3k
22Nov 20, '23OpenAI is nothing without its people18.7×348k1.3k6516
23Mar 1, '23Best AI skillset in 2018: PhD + long publication record in a specific area Best AI skillset in 2023: strong engineering abilities + adapting quickly to new directions without sunk 18.7×348k1.9k167408
24May 5, '23Turns out the ancient Chinese knew a lot about modern neural networks17.1×318k2.7k450390
25Sep 5, '23Hot take: what if Google Scholar reported two new metrics: (1) median citations per paper and (2) *percent* of papers with 100+ citations? I computed these metrics for some ~200 se17×316k50562322
26May 2, '25The longest chain-of-thought / reasoning trace I have witnessed was almost twenty minutes long and involved crazy backtracking, constraint verification, and tool use. But in the en16.6×309k6.4k250588
27Nov 10, '24There is a nuanced but important difference between chain-of-thought before and after o1. Before the o1 paradigm (i.e., chain-of-thought prompting), there was a mismatch between wh16.5×307k1.5k169861
28Sep 27, '23Large language models are notoriously hard to evaluate because (1) they are highly multi-task, (2) they generate long completions, and (3) grading is subjective. After spending ~5 16.2×301k1.6k159291
29Nov 28, '23It was an honor to give a guest lecture yesterday at Stanford’s CS330 class, "Deep Multi-Task and Meta-Learning"! I discussed a few very simple intuitions for how I personally thin14.9×278k1.5k2651.6k
30Sep 12, '23It seems to be not a coincidence that some of the strongest leaders in AI who manage large teams frequently do very low-level technical work. Jeff Dean doing weekly IC (individual 14.7×273k1.2k143469
31Dec 1, '23Today I am pleased to announce the new board of directors for my relationship. The new board of directors will be: 1. My mom 2. My girlfriend’s sister 3. @hwchung27, who I pair pro14.6×271k97023143
32May 15, '25AlphaEvolve is deeply disturbing for RL diehards like yours truly Maybe midtrain + good search is all you need for AI for scientific innovation And what an alpha move to keep it se14.4×268k1.8k94618
33Feb 28, '24My mental model of Sora is that it is the “GPT-2 moment” for video generation. GPT-2, which came out in 2018, could generate paragraphs of text that are coherent and grammatically 13.9×258k783116249
34Feb 10, '25Made this plot for an upcoming talk---crazy how quickly benchmarks get saturated these days. Looking forward to seeing how things play out for Humanity's Last Exam!13.5×251k966118445
35Jun 21, '25My favorite thing an old OpenAI buddy of mine told me is, whenever he hears that someone is a “great AI researcher”, he just directly spends 5 minutes looking at that person‘s PRs 13×241k97760459
36Sep 14, '23Pair programming isn’t standard at most companies and basically non-existent in academia, but I’ve been doing it with @hwchung27 for almost a year now. While it naively seems slowe13×241k934107368
37Oct 21, '23I tried to give this talk in the spirit of "a college soccer player watches videos of Messi and analyzes what makes him such a great soccer player." IMO it's great for people to as12.8×238k5281788
38Sep 12, '24o1-mini is the most surprising research result i've seen in the past year obviously i cannot spill the secret, but a small model getting >60% on AIME math competition is so good12.5×233k1.6k86271
39Nov 26, '24Prediction: within the next year there will be a pretty sharp transition of focus in AI from general user adoption to the ability to accelerate science and engineering. For the pas12.5×232k1.4k167496
40Feb 3, '25Very excited to finally share OpenAI's "deep research" model, which achieves twice the score of o3-mini on Humanity's Last Exam, and can even perform some tasks that would take PhD12.4×230k1.3k178447
41Sep 25, '23I recently dug up my work log from when I was an AI Resident at Google Brain and found some amusing nuggets: - I used to block out specific periods of time to code while wearing a 11.9×221k58020291
42Sep 16, '23Lost my waymo virginity last night; it was the first time I felt like SF is a futuristic city (though also dystopian, with all the homelessness around). I don’t think I’ll take hum11.8×220k5323293
43Nov 13, '23One lesson that I learned from moving to OpenAI (which is applicable to changing companies generally) is that the opportunity to reinvent myself and adapt to a new optimization lan11.5×213k82655350
44Nov 24, '23Like the International Math Olympiad or Spelling Bee, there should be a “language modeling competition” where humans compete to predict the next word in a sequence. The best humans11.3×210k4502737
45Aug 14, '25Old friends, new lab10.9×203k9692996
46Feb 7, '23I sometimes get questions on how to do good AI research, so I wrote a blog post about it: My personal take is that you can decompose research into four skills, each of which can be10.6×198k1.1k160861
47Jan 23, '24For most companies, hiring more people is strictly better. However, this is often not true in AI research. AI research is often bottlenecked by compute, and when this is the case, 10.3×191k53238229
48Mar 14, '23IMO GPT-4 is a bigger leap than GPT-3 was. - GPT-3 advanced AI from task-specific models to a single prompted model that is task-general - GPT-4 is human-level on many hard tasks, 10.3×191k4487480
49Dec 21, '24Biggest unsolved problem in artificial intelligence is what to do in the 3-5 minutes when o3 is thinking10.1×189k1.9k54122
50Oct 4, '23I reached 10k citations recently, a goal of mine for many years. It’s a nice moment to reflect back, and I mostly feel bittersweet: (1) When I joined Google Brain back in 2020, I t10.1×188k92368236
51Apr 13, '23I gave an invited lecture at New York University for @hhexiy's class! I covered three ideas driving the LLM revolution: scaling, emergence, and reasoning. I tried to frame them in 9.6×179k561147395
52Dec 8, '242022: I never wrote a RL paper or worked with a RL researcher. I didn’t think RL was crucial for AGI Now: I think about RL every day. My code is optimized for RL. The data I create9.2×170k1.3k83406
53Sep 19, '23An open question these days is what the role of task-specific / finetuned models will be. I can only think of three scenarios where it makes sense to work on task-specific models. 9.1×170k42267366
54Mar 22, '23AI moves fast, which means incumbents don't have a big advantage over new-joiners. For example, no one has >4 years of experience at prompting. Even 1k hours of practice makes y9.0×167k67279133
55Dec 31, '22My 2023 goals: - Spend 1,000 hours writing code (context: 377 hrs in 2022, 886 hrs in 2021) - Publish 0-2 first-author papers, but not more than 2 - Write 50 thoughtful tweets - Do7.9×147k57616114
56Apr 25, '24In AI research there is tremendous value in intuitions on what makes things work. In fact, this skill is what makes “yolo runs” successful, and can accelerate your team tremendousl7.8×146k45836212
57Feb 20, '25Seems to be not a coincidence that what AI is good at is correlated with the backgrounds of AI researchers. Demis was a chess prodigy; Jakub (Chief Scientist) and Mark (CRO) of Ope6.8×127k91263219
58Sep 6, '23My girlfriend doesn’t like the weekend plans I make for us, but she also doesn’t want to make plans herself. Instead, I should propose multiple schedules and then she picks one she6.7×125k65114103
59Jul 17, '23In the past weeks I received many questions (from undergrads especially) about AI research, so I'm putting together a "Ask Me Anything" doc. Add any questions to the doc, I'll answ5.9×110k715140635
60Dec 22, '24Doing a hyperparameter sweep tonight on wagyu (thread)5.9×110k1.2k23117
61Mar 10, '24Had a bit of a fanboy moment today meeting @bryan_johnson, who has been super inspirational to me in prioritizing my health. I asked him about the best way to balance career and sp5.8×107k55834139
62Oct 30, '24Excited to open-source a new hallucinations eval called SimpleQA! For a while it felt like there was no great benchmark for factuality, and so we created an eval that was simple, r5.7×107k854119543
63Oct 12, '23For any new prompting technique (e.g., tree-of-thought, least-to-most, graph-of-thoughts prompting), I consider four things to decide if it will become widely adopted: 1. How easy 5.7×106k47960458
64May 24, '24New blog post where I discuss what makes an language model evaluation successful, and the "seven sins" that make hinder an eval from gaining traction in the community: Had fun pres5.7×106k51874483
65May 3, '23Since GPT-4, some have argued that emergence in LLMs is overstated, or even a "mirage". I don't think these arguments debunk emergence, but they warrant discussion (it's generally 5.7×106k38887242
66Jun 28, '25I would say that we are undoubtedly at AGI when AI can create a real, living unicorn. And no I don’t mean a $1B company you nerds, I mean a literal pink horse with a spiral horn. A5.6×105k72341105
67May 18, '25Discriminator-generator gap seems to be the most important idea in AI for scientific innovation. With compute + clever search, anything that we can measure will be optimized. First5.2×97k61053366
68Jun 29, '25The most rewarding thing about working in the office on nights and weekends is not the actual work you get done, but the spontaneous conversations with other people who are always 5.1×95k80429160
69Apr 19, '24nothing gets my heart rate up like waiting for eval results on new models to come in4.9×92k4052332
70Dec 23, '24Realization: the old style of “hallucinations research” via self-calibration is probably going to die down. I used to be very excited about it but now I am skeptical because giving4.6×86k51348261
71Jan 3, '25An underrated but occasionally make-or-break skill in AI research (that didn’t really exist ten years ago) is the ability to find a dataset that actually exercises a new method you4.5×84k69769448
72Dec 30, '24Reflecting back, these were the biggest technical lessons for me in AI in the past five years: 2020: you can cast any language task as sequence prediction and learn it via pretrain4.4×83k54150286
73Aug 9, '23A somewhat amusing personal revelation: a friend recently asked me what were the best skills I had. I think he expected me to say something like "prompt engineering", "writing pape4.4×82k40746186
74Dec 5, '24Very excited for o1 to come out from preview! Main takeaways: - o1 thinks harder and is more performant on tough problems - o1 thinks faster on easy problems - o1 reasons over imag4.3×80k6045274
75Jun 2, '25There are traditionally two types of research: problem-driven research and method-driven research. As we’ve seen with large language models and now AlphaEvolve, it should be very c4.3×79k70592448
76Nov 29, '24Andrej’s tweet is the right way to think about it right now but I totally believe that in one or two years we will start relying on AI for very challenging decisions like diagnosin4.2×79k41348222
77Jun 12, '24As a kid I loved whiteboard lectures way more than slides, so for Stanford’s CS25 class I gave a whiteboard lecture! My goal was to simply and clearly explain why language models w4.1×76k64396673
78Apr 2, '25I am super excited for AI for scientific innovation, a direction that will certainly grow in the next five years. I think there will be two flavors of it. The first is “deepmind st3.9×72k62070283
79Jun 24, '25AI research is strange in that you spend a massive amount of compute on experiments to learn simple ideas that can be expressed in just a few sentences. Literally things like “trai3.8×70k59137226
80May 29, '25A recent clarity that I gained is viewing AI research as a “max-performance domain”, which means that you can be world-class by being very good at only one part of your job. As lon3.6×66k65653380
81Jan 7, '25Nice paper from Deepmind takes a fresh angle on factuality: While most existing factuality datasets focus on public world knowledge, this paper evaluates whether responses are cons3.6×66k708107551
82Oct 16, '23The highest possible compliment in our field is people not being on their laptops while youre giving a talk.3.4×64k4081321
83May 20, '24Enjoyed this extremely comprehensive study on predicting language model performance . Found many insightful nuggets: - In a single model family there usually aren't that many model3.4×63k40767406
84Jun 21, '22Today I acquired the "Research Scientist" title at Google Brain! (previously: software/research engineer) At 23 years old, does this make me the youngest Research Scientist at Goog3.4×01.6k22115
85Dec 19, '24Law of reinforcement learning: if you can measure it, you can optimize it3.3×61k73147136
86Jan 31, '24One thing in AI research that I have finally recognized with clarity is the idea of “inertia bias”: continuing to do something when it’s not the best option. The most basic instanc3.0×56k48767205
87Jun 11, '25RL environment specs are among the most consequential things we can write as AI researchers. A relatively short spec (e.g., <1000 words of instructions saying what problems to crea2.9×53k43331276
88Apr 10, '25New benchmark for deep research agents! An agent that is creative and persistent should be able to find any piece of information on the open web, even if it requires browsing hundr2.9×53k62564303
89Jun 6, '25Was attending a talk in a big lecture hall and the guy in front of me had the craziest conversation with ChatGPT for the whole hour about how to get his girlfriend back. Dozens of 2.8×52k413797
90Jan 21, '25Somewhat meta but there is a dopamine cycle in doing AI research that is pretty interesting Every day you wake up and you think about what experiment to run. You think thing X matt2.4×44k47533174
91Apr 20, '22progression of prompt engineering research2.4×01k11158
92Feb 1, '24One of the great pleasures in life is waking up and immediately going to my computer to check the results of experiments I launched last night2.2×42k4262728
93Jun 16, '22New survey paper! We discuss “emergent abilities” of large language models. Emergent abilities are only present in sufficiently large models, and thus they would not have been pred2.1×0560115182
94May 4, '22There's a skill I acquired early in my career that I wish I could unlearn. The skill is being able to package mediocre or uninteresting results into a paper that could be accepted 1.8×07274794
95Jul 8, '22I asked GPT-3 what are "fun things to do in Mountain View" and it returned nothing. I thought the Open API must be broken but then i realized that is the correct answer1.7×08473427
96Nov 21, '22Paper writing bro tip: I like to include a "Frequently Asked Questions" (FAQ) section in the Appendix. I always get positive feedback on it, but don’t see anyone else doing it. Gre1.6×053846147
97May 3, '22Meta's open source OPT-175 is comparable to GPT-3 175B. This is a massive step forward for bringing big LM research to academia. Expected nothing less from @LukeZettlemoyer and cre1.2×04267179
98Aug 30, '22"So do you speak Chinese?" Normal person: "I can understand but I don't speak well" @YiTayML: "my encoder is OK but my decoder is broken" 🤦1.0×05062015
100Jul 12, '22The year is 2017. I am training deep neural nets. I do hyperparameter tuning - it is partially science, but mostly black magic. The year is 2022. I am prompting large language mode0.9×04023422

Scan another creator

@
@karpathy@levelsio@paulg@naval@marclou@illyism

Run this on any account.

XOutlier builds this table for every public profile you open on x.com, with live data and every post, not just the top 50.

Add to Chrome Download .zip