XOutlier
Download .zipAdd to Chrome

Jeremy Howard

@jeremyphoward331k followers

🇦🇺 Co-founder: @AnswerDotAI/@FastDotAI ; Prev: Professor@UQ; @kaggle founding president; founder @fastmail/@enlitic/…

Open on X
Top outlier
54.4×
Best organic post
Peak views
2.5M
Single post
Bookmarks
102k
Across 93 organic posts
Last scan
Oct 16
Community cache
Posts 93 posts
#DatePost↓ OutlierViewsLikesRepostsBookmarks
1Nov 18, '23OK everyone's asking me for my take on the OpenAI stuff, so here it is. I have a strong feeling about what's going on, but no internal info so this is just me talking. The first po54.4×2.5M4.8k5202.7k
2Aug 10, '23Now that ChatGPT has rolled out custom instructions to most users, try out this instruction -- it makes GPT 4 far more accurate for me: (Concat the rest of this 🧵 together and put44.1×2M6.4k69611k
3Jul 10, '25I replicated this result, that Grok focuses nearly entirely on finding out what Elon thinks in order to align with that, on a fresh Grok 4 chat with no custom instructions.37.3×1.7M4.6k662979
4Nov 20, '24Wow! What on earth happened here? (I don't mean "why did so many people move to a different social network," but "why did they all go to 🦋")29.3×1.4M5.8k4651.1k
5Apr 15, '24"transformers or diffusion" Ummm.... who's gonna tell him that that makes no sense at all? (Transformers is an architecture, and it's quite common to use it *with* diffusion.)25.9×1.2M3.6k131567
6Jul 19, '24Wtf? The CrowdStrike file that broke everything was entirely full of null characters. How does this happen?25.7×1.2M6.6k4981.1k
7Oct 13, '23If you're like me and find it easier to read *code* than *math*, and you have access to @OpenAI GPT 4V (or use @bing or @google Bard), try pasting a image of an equation you wanna 24.7×1.1M1.8k245828
8Jul 10, '25Here's a complete unedited video of asking Grok for its views on the Israel/Palestine situation. It first searches twitter for what Elon thinks. Then it searches the web for Elon's23.8×1.1M7.1k8101.4k
9Apr 9, '23Lots of folks misunderstanding how GPT 3.5 & 4 work, such as in the take below. "Predict text" is the *pre-training* objective, but it's not what's used for the final model. Th23.5×1.1M1.7k208877
10May 4, '23There's a new programming language in town - it's Mojo! I'm more than a little excited about it. It's Python, but with none of Python's problems. You can write code as fast as C, a23.3×1.1M4.6k8642.5k
11Apr 5, '23Our new course, "From Deep Learning Foundations to Stable Diffusion", is finally done after 8 months of work!!! With >30 hours of video content (all free, no ads!), you'll learn22.5×1M4.6k1.1k3k
12Jun 29, '24For those that hope (or worry) that LLMs will do breakthrough scientific research, I've got good (or bad) news: LLMs are particularly, exceedingly, marvellously ill-suited to this 21.8×1M3.9k5683.1k
13Sep 24, '23I just uploaded a 90 minute tutorial, which is designed to be the one place I point coders at when they ask "hey, tell me everything I need to know about LLMs!" It starts at the ba20.7×958k6.2k1.1k7.4k
14Sep 28, '23Here, in full directly on Twitter, is "A Hackers' Guide to Language Models". This 90 minute tutorial is designed to be the one place I point coders at when they ask "hey, tell me e16.8×777k5.9k1k8k
15Apr 10, '23Don't listen to me. I don't understand language model fine tuning. I'm merely the 1st author of the paper "Universal Language Model Fine Tuning", which explained 5 years ago how to16.7×775k2.5k169339
16Jul 11, '23I've spent the last few months interviewing >60 experts in law, economics, AI, alignment, etc, on the impacts of AI, and safety interventions. Today I'm publishing my first arti16.4×761k2.2k5221.3k
17Jul 7, '25Study by my genius friend to CURE DEAFNESS cancelled because, I shit you not: "Research programs based primarily on artificial and non-scientific categories, including amorphous eq16.4×759k3.6k204487
18Nov 18, '23Screw this shitty credentialist arrogant take. Helen Toner is one of the most brilliant people I've ever chatted to. I'm not sure I've ever dealt with anyone else who was so inform15.2×706k1.6k65278
19Jun 19, '20So do these non-medical masks we're all meant to be wearing actually do anything? Well yes, yes they almost certainly do. Here's a thread about the current science. Let's start by 15.1×014k8.1k1.6k
20Jul 29, '24Announcing FastHTML. A new way to create modern interactive web apps. Scales down to a 6-line python file; scales up to complex production apps. Auth, DBs, caching, styling, etc bu14.4×667k5k6554.4k
21Feb 12, '24Also unfortunately not satire13.4×623k5.2k120255
22Apr 3, '23There's a lot of folks under the misunderstanding that it's now possible to run a 30B param LLM in <6GB, based on this GitHub discussion. This is not the case. Understanding why13.3×614k2.4k3901.7k
23Apr 24, '24In Australia we have a computational thinking competition for kids each year called Bebras. You get a certificate if you get all of them right. Here's one of the "easy" level quest13.2×610k3.9k207815
24Aug 1, '24I spent quite a lot of time and energy building something and releasing it for free as open source software. Someone posted a link to it on reddit. This is how some guy on reddit d12.5×579k4.2k57265
25Sep 7, '23I was chatting on Discord with an anon account ,who had just come up with a really smart new approach to model merging and had some encouraging results to show. But then he had to 12.1×562k3.9k304483
26Sep 29, '23Wow somehow I've been completely erased from the history of Kaggle. For the record, the US company that was funded, Kaggle Inc, was owned equally by Anthony Goldbloom and I with a 10.9×503k1.9k128154
27Oct 13, '23If you're like me and find it easier to read math than code, and you have access to @OpenAI GPT 4V, try pasting a image of an equation you wanna understand in there. It might just 10×463k2.4k305764
28Mar 7, '24Today, with @Tim_Dettmers, @huggingface, & @mobius_labs, we're releasing FSDP/QLoRA, a new project that lets you efficiently train very large (70b) models on a home computer wi9.6×445k3.4k6232.2k
29Mar 25, '20The Czech Republic went from zero mask usage to 100% in 10 days, and in the process they halted the growth of new covid-19 cases. How? They made their own! They didn't need governm9.5×08k4.4k1.3k
30Dec 19, '24I'll get straight to the point. We trained 2 new models. Like BERT, but modern. ModernBERT. Not some hypey GenAI thing, but a proper workhorse model, for retrieval, classification,9.5×438k4.7k6503.1k
31Aug 9, '25For many years, I've been looking for the perfect replacement for @jQuery. Modern vanilla js has many of the features that made jQuery popular, but often with worse ergonomics. Tod9.2×426k4.3k213665
32Jan 17, '25We tried really really hard to make Devin (the coding agent) work for us. But it didn't. Check out Hamel's detailed writeup blog linked below, describing the many tasks of many typ9.0×418k2.3k2031.4k
34Jul 8, '24Many PhD students have told me most earnestly that they could win Kaggle competitions any time they wanted. On the vanishingly rare occasions they've put their money where their mo8.3×383k1.8k89931
36Aug 8, '23I was curious about what @OpenAI GPT 3.5 would do if I used `function_call` to force it to call a function, but passed in a prompt that didn't make sense for that function. As it t7.7×357k1.6k111210
37Feb 6, '24GitHub banned their own co-founder, because something "went wrong with our automations". Please folks, do NOT use automated decision making tools to make decisions that can harm pe7.7×355k2.2k212214
38Nov 23, '23Bitcoin-turned-AI bros updating their profiles right now. ❤️ b̵i̵t̵c̵o̵i̵n̵ n̵f̵t̵ w̵e̵b̵3̵ A̵I̵ g̵e̵n̵e̵r̵a̵t̵i̵v̵e̵ RL/q-learning7.6×353k2.1k174103
39Sep 6, '23It looks like @johnowhitaker & I may have found something crazy: LLMs can nearly perfectly memorise from just 1-2 examples! We're written up a post explaining what we've seen, 7.5×348k1.7k2901k
40Dec 30, '24SW eng manager: No real work gets done in Jupyter notebooks. Alex Radford: I invented GPT and CLIP in Jupyter notebooks.7.2×333k3k176733
41Aug 10, '25Now that the era of the scaling "law" is coming to a close, I guess every lab will have their Llama 4 moment. Grok had theirs. OpenAI just had theirs too.7.0×325k2.2k97326
42Jan 16, '25Lots of OpenAI employees have been dropping hints recently that they've developed ASI internally. But OpenAI's latest product is a buggy mess. This strongly suggests to me that Ope6.8×316k3.4k157683
43Apr 16, '23The @karpathy approach to LLMops is the correct way6.7×311k1.6k77231
44Aug 14, '25The vast gulf between all of our actual daily experience with LLM improvements plateauing, vs what some researchers claim is going on, is pretty interesting. You can't properly und6.6×304k1.9k112394
45Apr 5, '25Just read through the Llama 4 release announcement. I'm really grateful they've released this with open weights. But tbh I'm also pretty disappointed. The models are both giant MoE6.5×299k1.9k97546
46Apr 23, '24Today at @answerdotai we've got something new for you: FSDP/QDoRA. We've tested it with @AIatMeta Llama3 and the results blow away anything we've seen before. I believe that this c6.4×296k1.8k2791.5k
47Sep 24, '25Chris stopped coding for a few weeks to raise $250m, then straight back to coding!🚀6.3×293k1.9k77326
48Nov 9, '25Side effect of blocking Chinese firms from buying the best NVIDIA cards: top models are now explicitly being trained to work well on older/cheaper GPUs. The new SoTA model from @Ki6.3×290k2.8k286733
49Mar 9, '20As data scientists, we (with @math_rachel) have done our best to look at the data around covid-19, and what it means to you and your community. Our view: it is appropriate to be ve6.2×05.8k3.1k790
50Feb 5, '25Wow, actual grown men are still doing the "I asked the LLM about itself and it said" thing. In 2025. Folks, LLMs don't know anything about how they themselves are built or deployed5.9×272k4.3k232253
51Feb 3, '25Imagine if this was a Chinese company.5.6×259k2.3k117228
52Jan 29, '24I used to find writing CUDA code rather terrifying. But then I discovered a couple of tricks that actually make it quite accessible. In this video I introduce CUDA in a way that wi5.3×245k2.5k3472.7k
53Jun 8, '20New analysis. Texas re-opened early, without requiring masks. Things are looking bad. NB: chart shows hospitalizations, not cases, so this is not a side-effect of increased testing4.9×04.4k2.6k228
54Jul 8, '22I led the team that studied mask efficacy in early 2020 and published our results in the Proceedings of the National Academy of Science. I spent three months earlier this year revi4.8×03.9k1.7k1.6k
55Jul 21, '22After 2 years, Practical Deep Learning for Coders v5 is finally ready! 🎊 This is a from-scratch rewrite of our most popular course. It has a focus on interactive explorations, &am4.6×04.6k1k2k
56Jan 28, '25Y’all do understand, I hope, that the reason for the NVIDIA collapse is that the US announced massive tariffs on imported chips, which would destroy NVIDIA’s business entirely, if 4.6×212k1.9k89247
57Jan 30, '24I used to find writing CUDA code rather terrifying. But then I discovered a couple of tricks that actually make it quite accessible. In this video I introduce CUDA in a way that wi4.6×211k2.4k3782.6k
58Oct 19, '25til some folks nowadays consider running a normal linux web server like we've all done for decades to be a "non-standard stack" feels like some kind of learned helplessness or some4.5×206k2.7k142314
59Oct 14, '25Interesting experience from @karpathy on this new LLM training repo.3.9×178k2.7k101560
60Jan 21, '24If you're a Python programmer looking to get started with CUDA, this weekend I'll be doing a free 1 hour tutorial on the absolute basics. Thanks to @neurosp1ke, @marksaroufim, and 3.7×169k2k2611.4k
61Feb 20, '20This is huge. The creator of the YOLO algorithms, which (along with SSD) set much of the path of modern object detection, has stopped doing any computer vision research due to ethi3.6×04.1k1.9k254
62Oct 16, '2518 months ago, @karpathy set a challenge: "Can you take my 2h13m tokenizer video and translate [into] a book chapter". We've done it! It includes prose, code & key images. It's3.5×161k3.1k3312.5k
63Jan 25, '25R1 reasoning stronger than many humans on this one:3.3×151k1.6k84252
64Jan 11, '21Our paper, "An evidence review of face masks against COVID-19", written by a cross-disciplinary team of 19 international experts, was published in the Proceedings of the National A3.2×03.1k1.3k585
65Oct 20, '22I got a special surprise for you all... We just released the first 5.5 hours of our new course "From Deep Learning Foundations to Stable Diffusion", for free!2.9×02.8k5561.4k
66Jan 7, '20I completed my 1st data science project ~30 years ago. Since then I've been continuously developing a questionnaire I use for all new data projects, to ensure the right info is ava2.9×03.2k891867
67Aug 21, '20Today is ’s biggest day in our four year history. We are releasing our new course, fastai v2, fastcore, fastscript, fastgpu, and the official launch of the book. 1/2.9×03.7k927499
68Oct 22, '24CEO of Microsoft getting ratioed by an anime pfp account...2.7×127k1.9k73185
69Sep 16, '22Big news: we're launching a new course in <4 weeks. "From Deep Learning Foundations to Stable Diffusion". Bigger news: for this course, we're teaming up with @StabilityAI! AFAIK2.7×03.1k5971.1k
70Jul 24, '24Woohoo gpu.cpp has now been officially released! This is a really exciting project which provides a device-independent way to use GPU compute. No more writing separate code for CUD2.6×121k1.6k220614
71Feb 28, '20We've made a draft of our deep learning book available as @ProjectJupyter notebooks. Important: This is for personal use, and not licensed for any redistribution or change of forma2.5×02.8k835688
72Mar 24, '20This is one of the most amazing and hopeful things I've read. How the Czech Republic, through community effort, got to 100% mask usage in 10 days, nearly all thru home DIY effort. 2.4×02.5k1.1k143
73Jan 5, '21Wow. @OpenAI has trained a model that can create new images - you just describe them in English. 🤯2.3×03.1k778279
74Jul 23, '22I've just published over 20 hours of tutorials and live coding showing how to: install python the right way; set up a terminal; write shell scripts; use vim; use a remote Jupyter s2.2×02.1k4041.1k
75Jan 27, '19This single line of code solves a years-old bug that has caused problems for thousands of people. No one had figured it out before. It took @StasBekman, one of the best programmers2.1×02.9k90783
76Jan 25, '19Practical Deep Learning for Coders, 2019 edition, is now available. With a shiny new video player with searchable transcripts. This course is 100% new material, including some new 2.0×02.4k852258
77Feb 27, '20In our book we always show code equivalents for any math notation. So often, the Python & @PyTorch code is clearer and simpler. And you can actually execute it and try it! For 2.0×03k454314
78Aug 15, '19"New State of the Art AI Optimizer: Rectified Adam (RAdam). Improve your AI accuracy instantly versus Adam, & why it works" It's been a long time since we've seen a new optimiz2.0×02.3k797354
79Apr 13, '20We've just completed a 19-author analysis of the effectiveness of mask wearing, with 84 references. To explain what the science shows, I teamed up with the wonderful Prof @trishgre2.0×01.6k988180
82Sep 20, '21I'm the lead author of the first and largest academic evidence review of masks, published in the Proceedings of the National Academy of Sciences, with 141 references. The claim tha1.9×02.1k877188
83Jan 26, '19Wow wow wow! @GoogleColab now supports the fastai library directly :D That means you can run all the lessons at , and do your own research and development with fastai, all for free1.9×02.6k664202
84Feb 24, '20We're launching `fastpages`, a platform which allows you to host a blog for free, with no ads. You can blog with @ProjectJupyter notebooks, @office Word, directly from @github's ma1.8×02.1k606395
85Feb 3, '25There's only 5 grok icons on my home screen? C'mon folks you can fit on *way* more than that.1.8×83k2.4k5651
86Sep 26, '18Proud to be launching "Introduction to Machine Learning for Coders" today. If you're looking for a modern, code-first, opinionated introduction to machine learning, I think you mig1.7×02.2k697240
87Oct 12, '18Learn: - Intro Machine Learning - Practical Deep Learning - Cutting Edge Deep Learning - Computational Linear Algebra1.7×02.1k603325
88Aug 1, '24My recording studio for the FastHTML video :D1.7×78k1.9k1970
89May 12, '22TIL it's this easy to plot math functions in Python, thanks to @SymPy. No more mucking about manually with np.linspace and matplotlib and friends...1.6×02.2k305399
90Jun 28, '19Learn: - Practical Deep Learning - Deep Learning Foundations - Computational Linear Algebra - Intro Machine Learning1.4×01.7k423403
91Mar 4, '22I am over the moon to announce: 1) I'm now a professor at University of Queensland (UQ), the top institute in my home state! 2) I'll be teaching a brand new deep learning course at1.4×02.6k169137
92Jul 18, '20I'm happy this little-known data scientist helped make a difference! 😊1.4×02.5k23583
93Feb 24, '21I'm really excited to say that I'm sending this tweet from Australia - our new home!😃 I learned lots in SF and had a great experience, but we've decided that long-term the US isn'1.4×02.8k6530
95Oct 11, '18Every week I get asked "what's 's business model?" Our business model is: we spend our personal money to do stuff that helps people use deep learning. That's it. We're totally inde1.1×01.9k31416
97Jan 25, '19Practical Deep Learning for Coders, 2019 edition, will be released tomorrow. It's looking amazing.1.1×01.7k30484
98Feb 21, '20We have a *lot* of really great news to share with you regarding our forthcoming book, "Deep Learning for Coders", thanks to the amazing @OReillyMedia content team. 1: We're expand1.1×01.5k287216
99Dec 5, '22I remember well when Google was first released. I felt confident that it would become a critical tool that we'd all rely on. I haven't felt that way again, until now: ChatGPT. In j1.0×01.8k145111

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