Jeremy Howard @jeremyphoward
🇦🇺 Co-founder: @AnswerDotAI & @FastDotAI ; Hon Professor: @UQSchoolITEE ; Digital Fellow: @Stanford answer.ai Brisbane/Queensland, Australia Joined August 2010-
Tweets55K
-
Followers220K
-
Following5K
-
Likes10K
Boeing exists in one of the most heavily regulated industries on Earth. In no way do they "regulate themselves." Here's Perplexity's rundown of all the regulation Boeing faces all over the world: perplexity.ai/search/how-man… I mean, surely all these safety boards should have…
I keep thinking back to my thesis - an optimal computer, using a 3D grid of a minimal number of carbon atoms with a doping Nitrogen or something - a dream if, ya know, manufacturability wasn't a concern. She would be fast... repository.rit.edu/theses/8080/
Just read @jeremyphoward's personal response to bill SB-1047. I'm sharing some insights because the aftermath of this bill will affect other regions and countries, too. "It could reduce AI safety, through reducing transparency, collaboration, diversity, and resilience." He's…
We are already seeing an explosion of AI regulation that is designed to ban open source while claiming to be neutral. SB 1047 designates a "hazardous capability" to include what a third party can show with infinite fine-tuning and re-training. Meanwhile, closed models get points…
We are already seeing an explosion of AI regulation that is designed to ban open source while claiming to be neutral. SB 1047 designates a "hazardous capability" to include what a third party can show with infinite fine-tuning and re-training. Meanwhile, closed models get points… https://t.co/xzTMMBePaK
We're all incredibly excited about MAX launching on @nvidia GPU's this summer 🔥 It's going to beyond epic - get ready! 🙌🏼🚀
There’s an art to distilling these to the absolute minimal necessary text. The human brain can’t comprehend how stupid these things are without practice.
Llama 3 degrades more than Llama 2 when quantized. Probably because Llama 3, trained on a record 15T tokens, captures extremely nuanced data relationships, utilizing even the minutest decimals in BF16 precision fully. Making it more sensitive to quantization degradation.…
Open letter to @Scott_Wiener re: SB-1047. A Safe Harbor for Independent AI Evaluation? Hi Scott, Just a personal thought from the investing perspective, 1047 seems likely to be about 1 week - 6 months away from an SBF 2.0-style scandal. The bill sponsors likely aren't being…
Perfect example of the statist mindset. Use an example of an incredibly mature public company and tech. Built in a highly regulated market. Where the defense contractors are an effectively oligopoly. And draw the conclusion that you should regulate startups. It's ludicrous.
Perfect example of the statist mindset. Use an example of an incredibly mature public company and tech. Built in a highly regulated market. Where the defense contractors are an effectively oligopoly. And draw the conclusion that you should regulate startups. It's ludicrous.
🚨 Effective Altruism's Bait-and-Switch: From Global Poverty to AI Doomerism 🚨 The Effective Altruism founders planned – from day one – to mislead donors and new members in order to build the movement's brand and community. aipanic.news/p/effective-al…
What a fucking disaster. California Bill 1047 is an attack on AI innovation. It has much of the bullshit often found in these bills (kill switch, certification, criminal penalties). It will hurt researchers, academics, & startups. legiscan.com/CA/text/SB1047…
How important is sample packing with 4d attention (multipack), training on inputs and naive (greedy) sample packing? Is there an study on this? My experiments on Alpaca fine-tunes yield to very similar models. cc @jeremyphoward @Teknium1
Zhaodong Chen is going to present his CUTLASS paper - EVT: Accelerating Deep Learning Training with Epilogue Visitor Tree in ASPLOS'24 on May 1. EVT is a framework to fuse almost any combination in the epilogue. dl.acm.org/doi/10.1145/36…
Be warned: I will continue the Cohere-command-r+ propaganda until the whole world has switched.
You can try out the mysterious gpt2-chatbot at chat.lmsys.org (select "Direct Chat" and pick it from the menu) Initial impressions: I'm very impressed. It gave me a better answer for an ego search ("Who is Simon Willison?") than any other model I've tried
You can try out the mysterious gpt2-chatbot at chat.lmsys.org (select "Direct Chat" and pick it from the menu) Initial impressions: I'm very impressed. It gave me a better answer for an ego search ("Who is Simon Willison?") than any other model I've tried
never thought i'd be featured in the financial times, but if i am going to be featured, i'm glad it's because of my crazy commute! 😂 ft.com/content/26b552…
Does anyone else feel like diffusion models have a hard time generating high frequency details? Any experience / thoughts / ideas / pointers on that manner? We have observed often that our models don't generate high frequencies in images and have found it hard improving it.
The best CUDA intro course by @nvidia with 460 bite sized videos. It was the course released with Udacity 9 yrs ago. It is kinda old, but you can grasp core ideas around it. youtube.com/playlist?list=…
This remains one of the best scifi stories I read.
This remains one of the best scifi stories I read.
Just stopped @AnthropicAI's Claudebot from completely ddos'ing a site with a large phpbb forum. Seriously uncool to be that aggressive with data-collection for your LLM.
François Chollet @fchollet
470K Followers 770 Following Deep learning @google. Creator of Keras. Author of 'Deep Learning with Python'. Opinions are my own.Yann LeCun @ylecun
711K Followers 718 Following Professor at NYU. Chief AI Scientist at Meta. Researcher in AI, Machine Learning, Robotics, etc. ACM Turing Award Laureate.Sebastian Raschka @rasbt
267K Followers 906 Following Machine learning & AI researcher writing at https://t.co/A0tXWzG1p5. LLM research engineer @LightningAI. Previously stats professor at UW-Madison.AK @_akhaliq
310K Followers 3K Following AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo follow on Hugging Face: https://t.co/q2Qoey80GxMark Tenenholtz @marktenenholtz
114K Followers 543 Following Head of AI @PredeloHQ. XGBoost peddler, transformer purveyor.elvis @omarsar0
189K Followers 486 Following Building with LLMs @dair_ai • Prev: Meta AI, Galactica LLM, PapersWithCode, Elastic, PhD • Creator of the Prompting Guide (~4M learners)Soumith Chintala @soumithchintala
186K Followers 883 Following Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.Harrison Kinsley @Sentdex
71K Followers 200 Following Neural networks from Scratch book: https://t.co/MWlYbXicwc YouTube: https://t.co/5osPue5EW9 @skunkworks_aiabhishek @abhi1thakur
81K Followers 662 Following 🤗 I build AutoTrain @huggingface 👨🏽💻 World's First 4x Grand Master @kaggle 🎥 YouTube 100k+: https://t.co/BHnem8fTu5 ⭐ GitHub StarChris Albon @chrisalbon
86K Followers 2K Following Director of Machine Learning at the Wikimedia Foundation. We host Wikipedia.Jean de Nyandwi @Jeande_d
38K Followers 773 Following Deep Learning, Vision 🤍 Language, Multimodal LLMs • AI Education • CMU Research blog: https://t.co/1BEFLZAqe7 ML Pack: https://t.co/7PkTyDvuriRadek Osmulski 🇺�.. @radekosmulski
25K Followers 555 Following Resources to take your Machine Learning skills to the next level 🧪 Senior Data Scientist, RecSys @NVIDIAAI 🏫 @fastdotai trained DL Eng 📝 https://t.co/By87iXx5Pu(((ل()(ل() 'yoav))).. @yoavgo
46K Followers 2K Followingmerve @mervenoyann
56K Followers 4K Following open-sourceress at @huggingface 🧙🏻♀️ proud mediterrenean 🍋 I do TL;DR on ML papersTanishq Mathew Abraha.. @iScienceLuvr
54K Followers 1K Following PhD at 19 | Founder and CEO at @MedARC_AI | Research Director at @StabilityAI | @kaggle Notebooks GM | Biomed. engineer @ 14 | TEDx talk➡https://t.co/xPxwKTq6QbSanyam Bhutani @bhutanisanyam1
35K Followers 994 Following 👨💻 Sr Data Scientist @h2oai | Previously: @weights_biases 🎙 Podcast Host @ctdsshow 👨🎓 International Fellow @fastdotai 🎲 Grandmaster @KaggleLior⚡ @AlphaSignalAI
84K Followers 897 Following Covering the latest in AI R&D • ML Engineer • Ex-Mila researcher • MIT Lecturer • Building AlphaSignal, a technical newsletter read by 180,000+ ML experts.Lucas Beyer (bl16) @giffmana
56K Followers 447 Following Researcher (Google DeepMind/Brain in Zürich, ex-RWTH Aachen), Gamer, Hacker, Belgian. Mostly gave up trying mastodon as [email protected]TuringPost @TheTuringPost
62K Followers 16K Following Newsletter exploring AI & ML - Weekly trends - LLM/FM insights - Unicorn spotlights - Global dynamics - History Led by @kseniase_ Elevate your AI game 👇🏼tanhuang @tanhuang520915
2 Followers 47 FollowingFaisal @faisal_aldarees
16 Followers 401 Following Data Scientist 👨💻, Computer vision. Classical music takes a big chunk of my life🎶. Pianist/Oudist. #ChopinDaniel Silva @danieljclsilva
19 Followers 451 FollowingPrashanth Koripalli @prashanthko7
4 Followers 130 FollowingNathan Hodgson @nathanhodgson_
573 Followers 312 Following Increase revenue, save time and reduce costs with AI & Automations.Lim Hoang @limdauto
41 Followers 346 Following SWE @Meta. ex @InflectionAI, QBLabs, @DeliverooEng, @Memrise | Engineering, edtech, AI, memory, standup, fiction. Also 🦮🐣. Here for the memesMrstevew @Mrwell732Mrwell
99 Followers 545 FollowingM Scott Maxwell @mscottmaxwell
1K Followers 68 Following Royal Navy Engineer turned Tech Production GuyM S @URfree4ever
11 Followers 228 FollowingAmonRA @AmonRA93
306 Followers 3K Following We're all mad here... "All the world's a stage, and all the men and women merely players."🇮🇳V143 @P143143143v
36 Followers 4K Following I am not into any political, personal, global wars , conflicts with anyone ❤️❤️❤️❤️❤️❤️ Indian 🇮🇳fan of modiji 😊😊SSR @Nature2tech
27 Followers 89 Following 🎆 Tech enthusiast 🤓. 🎆 Committed to nature 💐. 🎆 📖 + 🧠 + ☕ + 🧑💻 +👨🍳. 🎇Learning from everyday experience 🪄Ravi Yenugula @yenugularavi
3 Followers 24 FollowingAnBa Ca @ca_anba72855
13 Followers 84 FollowingUtopia DS @utopia_ds
11 Followers 70 FollowingReshmanth @Reshmanth1
23 Followers 37 FollowingWilliam Stephen Jones @WilliamSte80710
2 Followers 27 FollowingBulePanda @AIDreamfinder
219 Followers 1K Following AI creator, digital artist, world traveller and readerPaylz @paylza
129 Followers 2K Following The best online market for digital downloads with best prices.Nathan Shan @NathanShan5
0 Followers 34 FollowingKareem @kero_qm
0 Followers 10 FollowingHamza Essahbaoui @HamzaEssahbaoui
81 Followers 502 Following I build software, I sell and write. From a human to humains. ☪️LilyBoo___xh00 @LilyBoo___xh00
2K Followers 127 Following Travel ✈️ |Beach 🌊|Yoga 🧘♀️|Hardworking person|Like to explore and experience new things|#windenergy|#travelHariprasath @Hariprasath____
544 Followers 395 Following Eras 07/29 ✨It’s been a long time coming🫶🏽 || 23Educarte IA @EducarteIa
264 Followers 3K Following Desarrollador de soluciones con inteligencia artificial / Consultor Bussines IA / Researcher IA / especialista en SEO / Formulador de proyectosrich pavlovskiy @leg0m4n
214 Followers 989 Following co-founder @useBlanc // mathematics dropout @UofT // 𝘱𝘳𝘰 𝘓𝘌𝘎𝘖 𝘣𝘶𝘪𝘭𝘥𝘦𝘳safora jolfaie @sajolfaei
3 Followers 18 FollowingMichael Lai 赖天宸 @Mtclai
3K Followers 1K Following Running for the SF Board of Supervisors District 11 | SF DCCC | community first | before: reimagining early education at Tinycare @MinervaUni @HarvardAlejandro D. @windsof_fortune
22 Followers 176 Following Endless curiosity met with the fierce determination to create a lasting legacy.gQ - e/acc @caprands
32 Followers 90 Followingape ♨️ @apestrats
141 Followers 1K FollowingGopi Palamalai @gopipalamalai
0 Followers 92 FollowingRazorbill @wapooka
150 Followers 2K FollowingJesse White @byjlw_
42 Followers 258 Following Taking things apart to figure out how everything worksFuzionTech @FuzionTech
483 Followers 3K Following Full time data geek @ Posthog. When not in front of a computer you can usually find me on 2 wheels. Uber, SVB, Disqus, GroovesharkKK @CatharsistK
11 Followers 15 FollowingIlan @IlanShesh
0 Followers 67 FollowingBill W @BillW31501900
438 Followers 477 Following EE retired from the medical device industry. Tech geek. Amateur Radio Operator - WE5P , TSLA and 2022 MYLR owner. Tesla Referral link: https://t.co/lGR7TZwMtbFrançois Chollet @fchollet
470K Followers 770 Following Deep learning @google. Creator of Keras. Author of 'Deep Learning with Python'. Opinions are my own.Yann LeCun @ylecun
711K Followers 718 Following Professor at NYU. Chief AI Scientist at Meta. Researcher in AI, Machine Learning, Robotics, etc. ACM Turing Award Laureate.AK @_akhaliq
310K Followers 3K Following AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo follow on Hugging Face: https://t.co/q2Qoey80Gxelvis @omarsar0
189K Followers 486 Following Building with LLMs @dair_ai • Prev: Meta AI, Galactica LLM, PapersWithCode, Elastic, PhD • Creator of the Prompting Guide (~4M learners)AI at Meta @AIatMeta
532K Followers 255 Following Together with the AI community, we are pushing the boundaries of what’s possible through open science to create a more connected world.Soumith Chintala @soumithchintala
186K Followers 883 Following Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.PyTorch @PyTorch
379K Followers 77 Following Tensors and neural networks in Python with strong hardware acceleration. PyTorch is an open source project at the Linux Foundation. #PyTorchFoundationHarrison Kinsley @Sentdex
71K Followers 200 Following Neural networks from Scratch book: https://t.co/MWlYbXicwc YouTube: https://t.co/5osPue5EW9 @skunkworks_aiabhishek @abhi1thakur
81K Followers 662 Following 🤗 I build AutoTrain @huggingface 👨🏽💻 World's First 4x Grand Master @kaggle 🎥 YouTube 100k+: https://t.co/BHnem8fTu5 ⭐ GitHub StarPeyman Milanfar @docmilanfar
67K Followers 264 Following Distinguished Scientist at Google Research. Computational Imaging, Machine Learning, and Vision. Tweets = personal opinions. May change or disappear over time.Chris Albon @chrisalbon
86K Followers 2K Following Director of Machine Learning at the Wikimedia Foundation. We host Wikipedia.Jürgen Schmidhuber @SchmidhuberAI
107K Followers 0 Following Invented principles of meta-learning (1987), GANs (1990), Transformers (1991), very deep learning (1991), etc. Our AI is used many billions of times every day.Yi Ma @YiMaTweets
71K Followers 123 Following Chair Professor in AI, Director of IDS, Head of CS, HKU; Professor of EECS, Berkeley; Author of Book: High-Dim Data Analysis, https://t.co/gwaqMJp8av.Radek Osmulski 🇺�.. @radekosmulski
25K Followers 555 Following Resources to take your Machine Learning skills to the next level 🧪 Senior Data Scientist, RecSys @NVIDIAAI 🏫 @fastdotai trained DL Eng 📝 https://t.co/By87iXx5Pu(((ل()(ل() 'yoav))).. @yoavgo
46K Followers 2K Followingmerve @mervenoyann
56K Followers 4K Following open-sourceress at @huggingface 🧙🏻♀️ proud mediterrenean 🍋 I do TL;DR on ML papersDeepLearning.AI @DeepLearningAI
221K Followers 30 Following We are an education technology company with the mission to grow and connect the global AI community.Tanishq Mathew Abraha.. @iScienceLuvr
54K Followers 1K Following PhD at 19 | Founder and CEO at @MedARC_AI | Research Director at @StabilityAI | @kaggle Notebooks GM | Biomed. engineer @ 14 | TEDx talk➡https://t.co/xPxwKTq6QbMihai Chirculescu @m_chirculescu
128 Followers 288 FollowingEthan @3thanPetersen
323 Followers 2K Following DevEx and OSS AI at @CrusoeEnergy - talk data to me. ex-@Teslaifioravanti @ivanfioravanti
5K Followers 1K Following Co-founder and CTO of @CoreViewHQ GenAI/LLM addicted, Apple MLX, Ollama, Microsoft 365, Azure, Kubernetes, Investor in innovationTomas Petricek @tomaspetricek
10K Followers 971 Following Assistant prof @matfyz. Interested in new ways of thinking about programming and history & philosophy of computing. Previously at @UniKentComp and @Cambridge_CLRosio @learningrosio
22 Followers 702 FollowingGradient @Gradient_AI_
2K Followers 42 Following Accelerate AI transformation with Gradient AI Foundry, the most comprehensive solution to deploy autonomous assistants.Crusoe Energy @CrusoeEnergy
5K Followers 130 Following Aligning the future of computing with the future of the climate. We eliminate the environmental impact of energy intensive computing applicationsSteven Walton @WaltonStevenj
283 Followers 529 Following Ph.D. Candidate @ University of Oregon | Visiting Scholar @ Georgia Tech | Studying Computer Vision | SHI LabPengcheng Yin @pengchengyin
577 Followers 123 Following @GoogleDeepMind. Formerly a Neulab member @LTIatCMU. Interested in machine learning for NLP and code, dog training and aviation.Arman Cohan @armancohan
2K Followers 728 Following Assistant Professor of CS @Yale Research Scientist at AI2 @allen_ai NLP/AI ResearchAnsong Ni @AnsongNi
1K Followers 384 Following Final-year PhD student @Yale, #NLProc, LLM for Code. (ex-)intern @GoogleDeepMind, @MetaAI, @MSFTResearch, @allen_ai. MS from @SCSatCMU. Opinions are my own.Miltos Allamanis 🇪.. @miltos1
1K Followers 338 Following Researching deep learning for generating and understanding programs. Research Scientist @GoogleAI Also at @[email protected] (Opinions are my own.)Vivek Raghunathan @vivek7ue
4K Followers 2K Following * AI + search at @snowflakedb. * Co-founder @Neeva (acquired by @snowflakedb). #NeevaAI = AI search engine with LLMs. * Ex-VP of Engineering @Googlechrisrohlf @chrisrohlf
11K Followers 783 Following 🇺🇸 Waging algorithmic warfare since 2003. Software and Security Engineer. Non-Resident Research Fellow @CSETGeorgetown CyberAILaura Edelson @LauraEdelson2
12K Followers 442 Following Assistant Professor of Computer Science, Northeastern University. Co-Director, Cybersecurity for Democracy. Formerly: Chief Technologist, DoJ Antitrust DivisionOmar Kilani @omarkilani
367 Followers 704 Following eng @groqinc, co-founder @rememberthemilk. you can just do things enthusiast. @waymo fan account.Mo Tiwari @mo_tiwari
451 Followers 798 Following Computer Science PhD student at Stanford UniversityLuke DH Lee @luke_lee_ai
87 Followers 105 Following CS undergrad at UCL. Researching MAB & LLMs at Thrun Lab (SAIL)Jeyong Lee @vxbrandon00
95 Followers 14 FollowingSumit @_reachsumit
1K Followers 389 Following Senior ML Engineer @Meta | prev: @TikTok_us, @Amazon, @Samsung | UChicago Alum https://t.co/hcCJ2n979W 🇮🇳→🇰🇷→🇦🇺→🇨🇦→🇺🇲Ahmad Al-Dahle @Ahmad_Al_Dahle
4K Followers 53 Following #Girldad of twins. Leading GenAI @ Meta (llama, imagine, meta ai and more)Run-Ze Fan @Vfrz525_
362 Followers 656 Following Research Assistant@GAIR Lab @sjtu1896. NLP/LLMs/Alignment/Instruction Tuning. Looking for a Ph.D. in the 2025 fall (US)Aart Bik @AartBik
1K Followers 800 Following 🇳🇱🇺🇲 Dutch-American computer scientist Utrecht (MSc), Leiden (PhD) @Google @Intel #MLIR #LLVM #astronomy #chess #compilers #simd #sparse #vectorizationAna Brandusescu @anabrandusescu
6K Followers 2K Following Researcher. Into policy. AI governance, "the economy", sectoral power, privatization. She/herKieran Healy @kjhealy
41K Followers 97 Following Still just about here. Tweets self-destruct quickly. Web: https://t.co/sCKaQj48iX Elephant: https://t.co/Okfy3dsFtw Preorder: https://t.co/GFpkgfGBv6Eric Auld @AuldEric
308 Followers 691 Following AI, math, CS. Former @uclamath. I’ll let you be in my dream if I can be in yoursTim Paul @timpaul
5K Followers 900 Following Head of Interaction Design @GDSTeam (he/him). Views his own, not his employers.Tamay Besiroglu @tamaybes
3K Followers 720 Following Thinking about economics, computing and machine learning @EpochAIResearch @MIT_CSAILLech Mazur @LechMazur
32K Followers 392 Following CEO, Advameg, Inc. https://t.co/iLf8qsp4Qz founder Author: Local COVID-19 machine learning case prediction model. Author: https://t.co/tnwqrUXTsY. AI assistant for melody compositionmichael @mkwng
3K Followers 1K FollowingThomas Dohmke @ashtom
26K Followers 380 Following Building GitHub Copilot for the sake of developer happiness. CEO @GitHubsearch founder @n0riskn0r3ward
374 Followers 900 Following Solo entrepreneur passionate about search tech. Self-taught dev building a niche search product and sharing what I learn along the way.Scott Kennedy ⠕ @stkenned
1K Followers 306 Following VP of Engineering @Replit | Always hiring great people: https://t.co/aRv9pxW0z5lhchavez ⠕ @lhchavez
2K Followers 521 Following i'm lhchavez. i solve problems. cto @ https://t.co/SEPPbKpdq2Rob Reich @robreich
15K Followers 2K Following Professor, Stanford University Co-Author, SYSTEM ERROR: Where Big Tech Went Wrong Author, JUST GIVING: How Philanthropy is Undermining DemocracyMax Bain @maxhbain
2K Followers 498 Following multimodal @RekaAILabs | prev: phd @Oxford_VGG hardwork-pilledNathan Godey @nthngdy
535 Followers 841 Following 3rd year PhD student @InriaParisNLP Working on the representations of language models, architectures, and pretraining methods https://t.co/CTHFx1ZqPoTom Dörr @tom_doerr
198 Followers 416 FollowingDaniel Jeffries @Dan_Jeffries1
22K Followers 1K Following I'm an author, futurist, thinker and systems architect.Yatharth Gupta @_yatharthg
66 Followers 79 Following Facing the Exploration - Exploitation Dilemma........ CSE Undergrad@IITIndoreOrion Reed @OrionReedOne
2K Followers 107 Following advocating widespread dissatisfaction with computing. @[email protected]Boeing exists in one of the most heavily regulated industries on Earth. In no way do they "regulate themselves." Here's Perplexity's rundown of all the regulation Boeing faces all over the world: perplexity.ai/search/how-man… I mean, surely all these safety boards should have…
There’s an art to distilling these to the absolute minimal necessary text. The human brain can’t comprehend how stupid these things are without practice.
If you use a good mask like the Aura and you're not a healthcare worker you don't need a fit-test. Non-experts get an average fit factor of 88, well over the recommended goal of 10. (In healthcare the goal is 100, to provide a 10x safety margin.) tandfonline.com/doi/abs/10.108…
@rohanpaul_ai @HlibIvanov @Yampeleg Crucial T705 has 14Gb/s for 250$ that's ~6s Could be used in raid for 28Gb ~3s or 56Gs (1000$?) ~1s With 8b quantization it's more interesting It also suggests that GPT4 lvl LLM could run on ~1k$ PC with ~1 tokens/s That changes use cases types
@hyhieu226 Back to back matmul fusion does not have advantages unless you can keep X*W_1 in memory, which you cannot unless the reduction dim of W_1 is sufficiently small (i.e. <512 probably). It's a good thought though! From the draft of a post I'm writing.
I looked at that AI model legislation from a few weeks back -- it looks like it would establish an org authorized to regulate literally all ML training. In general -- I think gov orgs should have more limits than this: 1a3orn.com/sub/essays-lim…
"The amount of serendipity that will occur in your life is directly proportional to the degree to which you do something you're passionate about combined with the total number of people to whom this is effectively communicated."
hmm apparently if you ask "heyu" to phi-3 it regurgitates part of the synthetic data generate prompt and data it was likely trained on
Because AI is still only a pattern machine (an incredibly powerful one!). Here's how to see. Take a common trick question or puzzle, change it to not be a trick, ask GPT (here's Claude since I'm on Mobile, I use GPT4 by API and it does the exact same thing).
Twitter folks who agree with this claim that AI does not “reason”: what does this even mean? I’m befuddled
The vast majority of people expressing concern over AI + cyber have no experience or background in cyber security. If you’re in this camp I’ve got some sobering news for you, sophisticated and low skill attackers alike are already compromising “critical infrastructure” and thats…
Geoffrey Hinton is right. So-called open sourcing of the biggest models is completely crazy. As AI models become more capable they should become increasingly useful in bioweapons production and for use in large-scale cyber attacks that could cripple critical infrastructure.…
SoTA LLMs typically exhibit 99%+ non-zero activations, but it turns out that they are still intrinsically quite sparse! We introduce CATS, a simple post-training technique that achieves 50% activation sparsity for MLP layers with almost no drop in downstream evals, while…
@jeremyphoward @benjamin_warner I would like to thank you @benjamin_warner for your amazing work! 👏 🚀
I pulled together notes on all of the LLM plugins that have worked for me for Llama 3 - both for hosting locally (I've run 8B and 70B on my 64GB M2) and access via APIs (Groq is SO FAST for that) Options for accessing Llama 3 from the terminal using LLM simonwillison.net/2024/Apr/22/ll…
Links: - Code github.com/lm-sys/arena-h… - Arena-Hard Prompts huggingface.co/spaces/lmsys/a… - Blog post lmsys.org/blog/2024-04-1… Authors: @LiTianleli @infwinston @evan_a_frick @lisabdunlap @BanghuaZ
Finally found a single actual screenshot of the DARPA Digital Tutor (sort of—a later commercial adaptation). Crazy-making that there were zero figures in any of the papers about its design, and not enough details to imagine one. Some observations: * An instructional interface is…
Quick follow up thread on my "learning to code with AI when you suck at coding" thread. I continue to code with AI, doing more and more complex projects, and unexpectedly, it's teaching me to code. It's also giving me fresh insights into LLMs that many folks miss coming at it…
The proper method for estimating π is to toss sausages (from week 4 of my stats course)
How can you statistically estimate π?