Twitter Beğeni Hilesi

Got invited to a pre-New Years Eve party in SF… But instead of partying, everyone was just hacking on startups and side projects. They were intense. Here are some of the demos we saw from the party animals at AGI House SF (🧵):

Dans son discours du Nouvel An, la reine régnante du Danemark, Margrethe II, a annoncé son abdication du trône danois, prévue pour le 14 janvier 2024. Pour les amateurs de Tolkien, Margrethe est surtout connue en tant 👇

Why is Madrid the capital of Spain? Madrid is quite unique: • No river • Very recent • Tiny when it became a capital So why?

The "generative" part of the brain is what turns ideas and plans into actions, including spoken and written words. But the hard part of intelligence is to come up with those ideas and plans. In the brain, ideas are formed in the prefrontal cortex. Turning them into actions is

The perfect quote to describe LLMs can be found in a 1946 Jean Cocteau movie -- "Réfléchissez pour moi, je réfléchirai pour vous" (think for me, I will reflect for you). What you get from the model is always a reflection of the training data you put in -- itself a by-product of

When you hear about "human-level AI" being just "one LLM training run away", it's as if people building 3D printers (the kind that makes small, flimsy plastic objects) were telling you that they've nearly built an infinite matter replicator, capable of scanning and cloning any

Oh oh c'est drôle ! Ou comment tirer le meilleur parti de ChatGPT...

I have been working on vision+language models (VLMs) for a decade. And every few years, this community re-discovers the same lesson -- that on difficult tasks, VLMs regress to being nearly blind! Visual content provides minor improvement to a VLM over an LLM, even when these

🥁 Llama3 is out 🥁 8B and 70B models available today. 8k context length. Trained with 15 trillion tokens on a custom-built 24k GPU cluster. Great performance on various benchmarks, with Llam3-8B doing better than Llama2-70B in some cases. More versions are coming over the next

As long as AI systems are trained to reproduce human-generated data (e.g. text) and have no search/planning/reasoning capability, performance will saturate below or around human level. Furthermore, the amount of trials needed to reach that level will be far larger than the

We trained a robot dog to balance and walk on top of a yoga ball purely in simulation, and then transfer zero-shot to the real world. No fine-tuning. Just works. I’m excited to announce DrEureka, an LLM agent that writes code to train robot skills in simulation, and writes more

One thing that even relatively senior ML people often fail to grasp is that deep learning models are curves fitted to a data distribution. You cannot expect them to solve tasks outside of their training distribution (which is the sort of thing that you need intelligence for).

The "fun" part is that such good model actually exists, it's MiniCPM-Llama3-V 2.5. But, MiniCPM got a lot less attention, incl from me. Given the similar results, the main reason seems to be not coming from an Ivy-league Uni, but from a Chinese lab? Bad look for all of us...

Le drame HTTPS du jour : academie-francaise.fr a un certificat correct mais académie-française.fr a un auto-signé. #cybersécuritay

We are announcing €600M in Series B funding for our first anniversary.  We are grateful to our new and existing investors for their continued confidence and support for our global expansion. This will accelerate our roadmap as we continue to bring frontier AI into everyone’s

Today is a good day for open science. As part of our continued commitment to the growth and development of an open ecosystem, today at Meta FAIR we’re announcing four new publicly available AI models and additional research artifacts to inspire innovation in the community and

Intelligence Artificielle : on se posait déjà les mêmes questions métaphysiques il y a 65 ans… #LLM #perceptron

Je crois que c’est le meilleur titre (et le plus juste) qu’on lira sur ces Jeux Olympiques. #Paris2024

Je déteste en arriver là et c'est même la première fois que je fais ça mais au bout d'un moment, il faut pas déconner. Ce n'est d'ailleurs pas pour moi car cela concerne les deux plus belles personnes de notre monde. CANAL+, HONTE A VOUS ! INFO ABONNE CANAL+ CANAL+