Erfan Al-Hossami

Super excited to continue working on AI4Code with all the awesome folk @carperai! So many brilliant ideas and directions to pursue!

The price of knowing is a question.

@lexfridman What advice does the cursor team have to future tech founders?

I'm doing a podcast with the Cursor team. If you have questions / feature requests to discuss (including super-technical topics) let me know! For those not familiar, Cursor is a code editor based on VSCode that adds a lot of powerful features for AI-assisted coding. I've been using Cursor and love it, so I thought why not talk to the team. The founders are a bunch of brilliant folks from MIT. This conversation will be bigger than just about Cursor, but more generally about the future of programming with AI.

Try huggingface.co/stabilityai/st…, A 3B model nearly competitive to Llama-7B model.

Amazing work! Super proud of everyone in the Code.AI team!

Introducing FreeWilly1 and FreeWilly2 - The latest groundbreaking LLMs from Stability AI's and @carperai lab! ⭐Open access and remarkable versatility. #AI #LLMs #OpenAccess #StabilityAI bit.ly/3q1cCx2

Over the past year, Stability AI has grown rapidly, dedicated to our goal of building the foundation to activate humanity's potential. Here are some of the milestones and accomplishments we've had over the past year ↓

- LLMs were worse at causal discovery (discovery of what actions cause what outcome) compared to children. 3/n

- Needing to do RLHF indicates that we may not know what cognitively is going on with LLMs. - Cog Science debunks general intelligence. Is a 5-year-old smarter than a 4-year-old? 2/n

My main takeaways from the keynote by Alison Gopnik at #ACL2023. Main point: LLMs are cultural technologies. LLMs are tools that transmit culture/info just like the printed press. - LLMs are not about modeling/generating "the truth" but rather imitating human behavior. 1/n

@ncooper57 Congratulations Dr. Cooper!! 🎉🎉 It has been a long (but worthwhile) journey :'). Don't forget to celebrate your accomplishments!

I passed my Ph.D. defense! As one friend put it: I am now able to use Dr. Nerd 🤓🤓

Announcing StableLM❗ We’re releasing the first of our large language models, starting with 3B and 7B param models, with 15-65B to follow. Our LLMs are released under CC BY-SA license. We’re also releasing RLHF-tuned models for research use. Read more→ stability.ai/blog/stability…

.@carperai just passed a major milestone a few days ago, its now a whole year old. I thought we could take this time to recap the achievements we've accomplished over the last year as well as discuss potential directions. of the lab. A thread 🧵1/N

@ncooper57 @NathanThinks @StabilityAI @carperai @CodeDotAI I'm dead 😂😂

@NathanThinks @StabilityAI @carperai @CodeDotAI it's now O(n^2) compute!

Started my first day at @StabilityAI in the @carperai department 🤓 Can't wait to do all the @CodeDotAI things 😎😎😎

@ncooper57 @StabilityAI @carperai @CodeDotAI 🚀🚀🚀!!!!! Let's change the world together one "Compile and Run" at a time 🤓

@ncooper57 @beeonaposy @PhungVanDuy1 @reshinth_ FACTS🤣

@beeonaposy This resonates soooo hard... @PhungVanDuy1 @ealhossa @reshinth_

Me when the massive dataset I've been building all week finally comes together

@ncooper57 @langchain @code SWEET! I'mma fork -- we can do so much with this LOL

well would ya look at that? I did manage to whip something together with @langchain after all 🤓 Checkout my VSCode Extension "VSCode Chat" a chatbot built right into @code! marketplace.visualstudio.com/items?itemName…

Anyone know how to converting TikToken to Huggingface Tokenizer? discuss.huggingface.co/t/converting-t…

@ncooper57 It was used in CodeBERT: arxiv.org/abs/2002.08155 I thought it was really intuitive perhaps we can revisit it :')

Why isn't ELECTRA pretraining not talked about more? It proposes a way more sample efficient training scheme to BERT, yet I have not seen anyone using it tbh. What's up with that? arxiv.org/pdf/2003.10555…

Ok friends, it's weekend, we've some time for us. So let me tell you about a possible future for AI where the largest AI spendings, in billions of dollars, in 2-3 years would be on... ... antitrust legal fees A quick 🧵1/8

@ncooper57 @reshinth_ @PhungVanDuy1 TRUEEEE!!! 😂😂😂🤣

@ealhossa @reshinth_ @PhungVanDuy1 trying to use argoflow...