Monicaxie (@Monica_XieY) started following @mathemagic1an on Sep 16, 2024

Code is a (graph) data structure. Storing it in text files is convenient for IDEs, but has many downsides: ❌ Import hell ❌ Large files ❌ Poor tree shaking ... etc. I think we'll soon eliminate the "file" abstraction and instead operate directly on the symbol graph.

Terence Tao is giving commentary on o1's math capabilities on Mastodon, and has mixed but overall optimistic takeaways World's foremost expert is seriously evaluating a model vs. his PhD students w/ mixed results. Rate of improvement is incredible. https://t.co/CsGf5weNX2 https://t.co/tjqJVT8dxt

RT @RWKV_AI: RWKV.cpp - is now being deployed to half a billion systems worldwide Making it one of the world's most widely deployed, truly…

RT @jandotai: Integration to run a GGUF model in Jan? Oh, not needed! - Find the GGUF model on @huggingface - Copy the link & paste it in…

Something I don’t understand about test time compute: why steer MCTS with RM vs. steering MCTS with logprobs on a PPO’d model?

I’m enjoying Wayward Pines atm https://t.co/V1h3OwuseG

world, meet world labs! language is only one facet of human intelligence -- even more visceral is is spatial intelligence, or understanding and reasoning about the physical world. we @radicalvcfund are thrilled to partner with this truly world-class team - @drfeifei @jcjohnss

can't say enough about this team and the opportunity they're tackling. go @Nominal_io ! 🚀

RT @molwelch: world, meet world labs! language is only one facet of human intelligence -- even more visceral is is spatial intelligence, o…

RT @annbordetsky: I remember going to an AI talk by @drfeifei about 7 years ago, long before this GenAI wave... it was mind blowing Super…

5/ All evals are here: https://t.co/NlMKVjYx7s

6/ Read our full blog post for more! https://t.co/pDDMp8uv8v

RT @LeptonAI: Here is the fastest diffuser in the west https://t.co/MlGIBicC6k by collaboration with @hippoml_com 😝 #sdxl #diffusion #GPU

RT @KubernetesPod: What did you do for your internship? Xiang Li (@xiangli0227) created the distributed key-value store @etcdio. #etcd n…

RT @gravicle: #LumaDreamMachine 1.5 is here and it's a great leap in text-to-video quality and intelligence! This new model understands tex…

RT @PatronusAI: 1/ Introducing Lynx v1.1: an 8B State-of-the-Art RAG hallucination detection model 🚀 - Beats Claude-3.5-Sonnet on HaluBenc…

too many people let their first 25 years define them. you can get cracked, master an instrument, land your first dunk, publish a paper, whatever. 10k hours is only 2 years of grinding.

altera beta players have now created thousands of agent friends to play games with. here i create, play with, and get killed by ironman within two minutes. https://t.co/2dMxLy1HuQ

Great work resonating our recent research on agent learning (https://t.co/7fe7yOYYBn). Love to see more inspiring works on this new research direction!

RT @rongwu_xu: ☕️New paper 👉Our latest paper delves into LLMs' ability to perform safety self-correction, namely COURSE-CORRECTION. In th…

RT @neuralmagic: Last week's vLLM office hours recording is ready! 🎥 @tms_jr showed how to use NVIDIA CUTLASS for high-performance inferenc…

RT @liger_kernel: 🎉 We are excited to have surpassed 50,000 downloads (stable+nightly) in less than 3 weeks! We've added many new features…

NEW Call - EARLY ENTRY 🙉 🙈 $APES - Meme coin with moonshot potential (Solid narrative + farming game + viral-worthy meme) 🚀 🍌🍌 LP burned 100% | Ownership renounced CA: 984GBL7PhceChtN64NWLdBb49rSQXX7ozpdkEbR1pump TG, TW, INST: @Apesol_gang Website: https://t.co/ppeeHHkSbT https://t.co/lTOccRMTj3

RT @physicaldoge: Meets second hamster kombat !! 🔥🔥🔥 Train and nurture your PDOGE to become the ultimate battle king!⚔️⚔️⚔️ ⭐ Official web…

We are looking for mathematicians to contribute original and difficult mathematics problems to a novel math benchmark for AI systems. This is a full-time, temporary, remote contract. Applications are rolling, so apply soon!

15/ See the thread below for a summary of the whole report! https://t.co/1gaR6zHggd

RT @cognitivecompai: @TheEricHartford and Cognitive Computations present dolphin-2.9.4-llama3.1-8b Now with extra SystemChatMultilingual,…

You can now finetune the new Llama 3.1 models by @AIatMeta on custom synthetic datasets on Glaive. Performance on our internal finetuning benchmarks looks extremely promising, go ahead and try it out. https://t.co/AXZlZLNYwi

RT @awong_SF: Solid quote from the new owner of Anchor Steam https://t.co/N4v0gcRiXR https://t.co/CdHQwWm05x

RT @lulumeservey: SF is an amazing place for women, and it’s surprising there are so few. Job opportunities, potential cofounders, eager i…

RT @meliarobin: Signature Block is one of my favorite newsletters for early to mid-career founders and investors. You know @rrhoover and @v…

RT @rrhoover: Investing can take you down some wild rabbit holes. @Julian, thanks for introducing us to @GillVerd. 🙏🏼