About LLMs export control. A few interesting things that people right now may not yet focus since they are "far" problems. 1. It is impossible to really take LLMs safe. They are a a few TB files worth of weights. People will find ways to copy them, there will be leaks. #LLMs #ExportControl #AILeaks #Therearenospecificcountrymentionsinthetweetprovided.
Ambiance magnifique ! Les supporters des Bleus 🇫🇷 ont été escortés jusqu’au stade avec la police américaine 🇺🇸👮♂️ https://t.co/LlmSir5ky7 #Supporters #Stade #Bleus
Modern AI resulted from research made also by many non-US scientists (Hinton, the French folks, Linnainmaa, many others). The pre-training corpus was produced worldwide with massive code contribution from Europe OSS. What is happening with frontier LLMs is unacceptable. #AI #LLMs #OpenSource #FR #EU
Prediction: now that with GLM / DeepSeek & more we will see very good LLMs available open weight, and even more when the GPU/RAM insanity will be contained, the open source agents landscape will create *much* better tools that we are used to use today. For instance: #LLM #OpenSource #AI #Nospecificcountrycodesarementionedinthetweet.
While Fable is an amazing model, don't get too excited: it is great, but still has the usual failure models of the other good LLMs we saw in the past, including GPT 5.5. If you look at Anthropic, Opus -> Fable was a huge jump. If you look at the field, GPT 5.5 -> Fable is #Fable #LLM #AI #Nospecificcountrycodesarementionedinthetweet.
Lo más importante de la salida de Fable / Mythos es la demostración, una vez más, de que no hay muro. El paradigma de los LLMs y el test-time-compute lejos de frenarse sigue trepando la colina trayendo modelos cada vez más inteligentes y capaces, que están transformando el mundo. Da igual los análisis coyunturales de qué lab se pone por encima, a qué coste, cuánto es subvencionado, o con qué limitaciones. Nada de esto es tan importante como que la humanidad sigue encontrando rendimientos de una tecnología que cada vez nos aporta más inteligencia en la resolución de problemas cada vez más complejos. La frontera se sigue moviendo y el muro nunca existió. #InteligenciaArtificial #TransformaciónDigital #Tecnología
The side of evil - "AI-Worms": Virus and Computer Worms "Powered with LLMs" infecting networks (translated)
🤝REENCUENTRO MACRON - MBAPPÉ. 🇫🇷 El presidente de Francia visitó a los jugadores de la selección antes de viajar al Mundial. https://t.co/LlMS7SlynN #Francia #Mundial #Fútbol
It's been shown that LLMs make 90% fewer typos than humans. To make the web a safer place, you will soon need a special endorsement on your internet license if you wish to enable manual typing in textboxes and text editors #LLMs #SaferWeb #TypingPolicy #Thetweetdoesnotmentionanyspecificcountries.Therefore #therearenocountrycodestoreturn.
It's been shown that LLMs make 90% fewer typos than humans. To make the web a safer place, users will soon need a special endorsement on their internet license if they wish to enable manual typing in textboxes and text editors #LLMs #SaferWeb #TypingEndorsement #Nospecificcountrycodesarementionedinthetweet.
Before we claim that LLMs generate bad frontend code, we should establish the frame of reference. Bad compared to what? Many of these discussions compare LLMs against some idealized standard of “perfect” frontend code, forgetting the realities of most human-authored code that #LLMs #FrontendDevelopment #CodeQuality #Nospecificcountrycodescanbeidentifiedinthetweetprovided.
Yes: “LLMs do not and cannot achieve the purpose of education, which is to encounter greatness (per Allan Bloom).” #Education #LLMs #Greatness #Nocountrycodesarementionedinthetweet.
RT @haider1: Yann LeCun says LLMs are strongest in domains where language itself is the substrate of reasoning, like math and code They ca… #LLMs #ArtificialIntelligence #MachineLearning #Thetweetdoesnotmentionanyspecificcountries.Therefore #therearenocountrycodestoreturn.
Phenomenal post, and one of the best uses of LLMs I've seen: "I wanted a better sense of how long it takes for some technology to appear once its necessary predecessors are in place. So I used AI to try and find out. I used a list of 190 major inventions that I’ve used for #AI #Innovation #Technology #Therearenospecificcountrymentionsintheprovidedtweet.
RT @haider1: Yann LeCun says you cannot build a reliable agentic system without a world model LLMs don't have world models. They can't pre… #AI #MachineLearning #WorldModel #Thetweetdoesnotmentionanyspecificcountries.Therefore #therearenocountrycodestoreturn.
RT @bergerthor: A lot of the debate on AI and science focuses on whether LLMs can generate *new* ideas. Yet 20–80% of existing papers are n… #AI #Research #Innovation #Nospecificcountrycodesarementionedinthetweet.
RT @OpenRouter: `stripe projects add openrouter/api` An easy way to accept @stripe payments for 400+ LLMs and image/video/audio models, no… #AI #Payments #MachineLearning
What if AI helped dismantle inequality? 👊 Our free research series returns with: "Advancing AI for Society via Race-conscious Algorithmic Approaches" Exploring: - Bias in LLMs - The myth of neutral data - Designing AI that centers lived experience https://t.co/VoqpEAmM3L https://t.co/jqcgBBnzzk -
Congrats on the launch @simile_ai ! (and I am excited to be involved as a small angel.) Simile is working on a really interesting, imo under-explored dimension of LLMs. Usually, the LLMs you talk to have a single, specific, crafted personality. But in principle, the native, primordial form of a pretrained LLM is that it is a simulation engine trained over the text of a highly diverse population of people on the internet. Why not lean into that statistical power: Why simulate one "person" when you could try to simulate a population? How do you build such a simulator? How do you manage its entropy? How faithful is it? How can it be useful? What emergent properties might arise of similes in loops? Imo these are very interesting, promising and under-explored topics and the team here is great. All the best! -
