Monicaxie

Follow the man! https://t.co/GObQSN49Vh

This looks like progress to me.

Check the new video arena! Pick your favorite video🚀

Check out the amazing BFCL V2!

I sat down with @jheitzeb a few weeks ago and we chatted about AI code interpreting 👇

The team from @LandingAI is rolling out early access to super exciting AI agent. It's visual AI pilot that's using AI-code execution to perform localization, object detection, or counting. It's a great example how AI codegen can be useful for other AI agents than just pure

RT @BillAckman: This is actually an important story. I bring it to your attention as a work colleague today used this example as evidence t…

Full Story: https://t.co/ZmSLJ9Qh47

FINAL OUTPUT: Energy & Humanity ⚡️ A complete deep dive on how energy has defined our past, and how it will determine our future. https://t.co/wC0u5WLiia

Finally, I want to thank @BainCapVC (especially @kevinzhang, @RonMiasnik) for supporting me on my deep dives!

I'm thrilled to announce that I've joined as CISO, alongside @embeddedsec, at @OpenAI. Security is germane to OpenAI's mission. It is critical we meet the highest standards for compliance, trust, and security to protect hundreds of millions of users of our products, enable

After ten years, my chapter at @PalantirTech has come to an end. I couldn’t be more grateful, humbled, or proud to work alongside some of the best people in the industry to advance the interests of, and help protect, the governments, institutions, and agencies that make the

RIP truth terminals mentions will build a better polling engine for replies soon; they're coming in so fast it's basically random chance as to whether it sees something it can reply to https://t.co/1uxd3j7NCz

RT @repligate: Something about the Truth Terminal situation causes people to completely reject its reality. I'm not sure what. @AISafetyMe…

RT @karpathy: The future expands the variance of human condition a lot more than it drags its mean. This is an empirical observation with i…

A lot of cool people seem to be at COLM. I haven’t had this amount of FOMO for missing a conference since EMNLP in Singapore.

RT @shuyanzhxyc: 🦾#NeurIPS2024 Introducing Synatra, a data synthesis approach for web agents that can generate demonstrations for a wide ra…

RT @thinkingwest: The British Empire was, in some ways, a force for good. In many places it occupied it: -raised the standard of living…

RT @elonmusk: Not many people these days know that the British Empire was the driving force behind ending the vast majority of global slave…

RT @DimitrisPapail: LLMs Can In-context Learn Multiple Tasks in Superposition We explore a bizarre LLM superpower that allows them to solv…

RT @MehulDamani2: Inference-time compute can boost LM performance, but it's costly! How can we optimally allocate it across prompts? In our…

Note: Being badly named doesn't mean current "unlearning" work is useless, I just think it should be renamed to eg "knowledge suppression" and people should be realistic about its limitations and how robust it's likely to be

This looks like a lovely paper! I've long thought that the term "unlearning" was BS. IMO unlearning means removing knowledge from the model's weights. But people tend to measure behaviour, which is indistinguishable from suppression in later layers. I'm glad someone showed this!

ICLR 2025 will have another blogpost track! If you have new intuitions on past work, noticed key implementation details for reproducibility, have insights into the societal implications of AI, or an interesting negative result, consider writing and submitting a blogpost.

RT @Michael05156007: New paper! Over-optimization in RL is well-known, but it even occurs when KL(policy || base model) is constrained fair…

Highly recommend this resource on evals from @clefourrier !!

RT @hugobowne: 📺 Going live in 15 with @haileysch__ (@AiEleuther) to chat about The Future of Open-Source AI and Research Infrastructure in…

🚀 Exciting internship opportunity! Join the Llama team @AIatMeta and help redefine what's possible with large language models—from pre-training to post-training. Be part of our 2025 research internship and help shape the future of LLMs. Feel free to email or DM me 📩 Learn

🚀 New paper from our Llama team @AIatMeta! We discuss "cross capabilities" and "Law of the Weakest Link" of large language models (LLMs): 🔹 Cross capabilities: the intersection of multiple distinct capabilities across different types of expertise necessary to address complex, https://t.co/YP4m4koxln