You might like
what changed is that evals are now RL training data
apropos of the LM arena raise — I'm skeptical of eval startups, because so many failed in previous waves. why? 1) great eval ppl should probably work on post-training instead 2) customers graduate quickly to doing their own evals 3) labs try rly hard to goodhart eval startups
I’m stoked to share our new paper: “Harnessing the Universal Geometry of Embeddings” with @jxmnop, Collin Zhang, and @shmatikov. We present the first method to translate text embeddings across different spaces without any paired data or encoders. Here's why we're excited: 🧵👇🏾
It's deeply concerning that one of the best AI researchers I've worked with, @kaicathyc, was denied a U.S. green card today. A Canadian who's lived and contributed here for 12 years now has to leave. We’re risking America’s AI leadership when we turn away talent like this.
Starting salaries at Cambridge are less than $70k for most fields. Stanford? Double that. As someone trying to hire the best global talent to work on UK and EU projects, I can't compete with the American universities. One election won't change that.
Talent and academics will leave the US. The EU (Member States) should initiate (tech) talent visas that can be expedited swiftly
a reason why it’s painful for academics to be on X is the default academic epistemic humility mode we like to talk and write in does not work on this platform where shitpoasters speak/write as if they dug up a tablet and god spoke to them about everything.
We should call models like Llama 3, Mixtral, etc. “open-weight models”, not “open-source models”. For a model to be open-source, the code and training data need to be public (good examples: GPT-J, OLMo, RedPajama, StarCoder, K2, etc.). Weights are like an exe file, which would be…
That AAAI felt compelled to put an upper bound on the number of papers (ten!) from an author should tell you something about the state of AI publishing today.
This would essentially make it illegal for a foreigner to do a postdoc in the UK.
🚨 BREAKING: Rishi Sunak will increase the salary threshold so that EVERY migrant has to earn over £38k before coming to Britain in a major plan to slash migration He will also announce new limits on them bringing family [@JackElsom]
We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett…
Launching today: AI-powered candidate feedback summaries in Rippling Recruiting 🚀 This new feature auto-synthesizes interview feedback into a digestible, helpful snapshot of each job candidate. Stay tuned—more AI-driven features are coming soon🛠️ rippling.com/blog/bringing-…
We investigate what features of tabular data explain the difference, for this, we modify tabular data to narrow the gap. Smoothing the outcome in feature space narrows the gap: deep architectures struggle with irregular patterns. Tree models do not care about smoothness 6/9
What are the latest research trends in AI? Explore all NeurIPS submissions from 1987 to 2022 in Atlas. atlas.nomic.ai/map/neurips
A lot has been speculated about TikTok's recommendations. This is the first paper I've read by the team, and it has many interesting details: expirable embeddings, parameter server, online training... Good #recsys stuff arxiv.org/abs/2209.07663
If you don’t think @MetaAI is not going to do something amazing with all those GPUs, think twice.
What I learned over the years is that people outside of the machine learning research community are so much better at UI/UX, design, web-dev, creating apps and plugins for all sorts of platforms, and making ML actually useful to a large number of real users.
I turned stablediffusion-infinity (#outpainting with #stablediffusion on an infinite canvas) into a web app using @pyscript_dev and @Gradio: PyScript for the frontend and Gradio for the backend. Source code: github.com/lkwq007/stable… Also available on Colab
📢Announcement📢: After careful consideration, we have decided to host AISTATS 2023 in Valencia (Spain), rather than a location in the US. Valencia was the original planned location for the 2022 conference, which had to be moved online. See you in Valencia!
Neural rendering takes its next step with DLSS 3.0 on Ada! In addition to DL-powered superresolution, it uses optical flow, motion vectors, and DL to generate entire frames. 7 out of 8 pixels being rendered with DLSS3 come from Neural rendering. #GTC22
I made an accessibility map of #Neurips2022. The country is marked in red if its resident can not get appointment in time from getting accepted to the conference (76 days).
This seems worthy of a first tweet! BB3 is the first 175B param, publicly available chatbot, with model weights, code, datasets, model cards, and focus on safety & explainability. See the Meta AI blog for more ai.facebook.com/blog/blenderbo…. And try out the demo blenderbot.ai
Claude Shannon is the founder of Information Theory His essay on creative thinking is criminally underrated I've unpacked 2 of Shannon's ideas below THREAD...
United States Trends
- 1. GTA 6 43.8K posts
- 2. GTA VI 15.3K posts
- 3. Rockstar 44.4K posts
- 4. #LOUDERTHANEVER 1,493 posts
- 5. GTA 5 7,029 posts
- 6. Nancy Pelosi 114K posts
- 7. Rockies 3,699 posts
- 8. Paul DePodesta 1,739 posts
- 9. Antonio Brown 3,467 posts
- 10. Ozempic 15.7K posts
- 11. Grand Theft Auto VI 35.7K posts
- 12. GTA 7 N/A
- 13. Justin Dean N/A
- 14. Luke Fickell N/A
- 15. $TSLA 54.4K posts
- 16. Elon Musk 222K posts
- 17. RFK Jr 27.3K posts
- 18. Pete Fairbanks N/A
- 19. Jonah Hill 1,454 posts
- 20. Michael Jackson 88.9K posts
Something went wrong.
Something went wrong.