Non. C'était vrai aux débuts de Chatgpt. C'était d'ailleurs l'argument de tous les vendeurs de formation IA : formez vous pour prompter ou vous serez remplacés. Ce n'est plus vrai. L'IA décèle notre intention et s'ajuste seule pour obtenir un résultat de qualité. https://
x.com/obelix_idefix/
/obelix_idefix/status/2035641941698777338
…
LLMS
-
Prompt Engineering No Longer Needed as AI Detects Intent
By
–
-
NVIDIA CEO Jensen Huang interviews 10 AI CEOs on open models
By
–
Panel discussions usually suck.
— Robert Scoble (@Scobleizer) 22 mars 2026
This one didn't (for the AI community).
It is of NVIDIA CEO/founder Jensen Huang interviewing 10 CEOs from frontier AI labs about what's happening in Open Models. In audience was thousands of the world's top AI thinkers.
CEOs on stage in the… pic.twitter.com/0SJBJCT7fHPanel discussions usually suck. This one didn't (for the AI community). It is of NVIDIA CEO/founder Jensen Huang interviewing 10 CEOs from frontier AI labs about what's happening in Open Models. In audience was thousands of the world's top AI thinkers. CEOs on stage in the
-
STEM: Scaling Transformers Efficiently with Static Token Embeddings
By
–
How can we scale AI models to be smarter and more efficient without the instability and immense compute? CMU & Meta researchers introduce STEM! Their new method replaces complex Transformer up-projections with a static, token-indexed "embedding lookup system." Think of it as
-
Weekly Space Planned for Local AI Hardware Discussions
By
–
I am thinking of hosting a weekly space to discuss and answer questions on hardware options for local ai / llm If you’re interested please interact w/ this tweet I am open to suggestions for other topics as well, but I am hoping this will lead to active discussions in the space
-
Local LLM Bottlenecks: Bandwidth, Interconnect, and Inference Engines
By
–
When running LLMs locally,
the bottleneck isn’t just “VRAM size” It’s: – memory bandwidth
– interconnect (PCIe vs NVLink vs RDMA)
– inference engine (vLLM, TensorRT-LLM, SGLang) Unified Memory is way slower than VRAM btw -
MiniMax-M2.7 Weights to Be Open-Sourced Within Weeks
By
–
MiniMax-M2.7 weights will be opensourced within the next couple of weeks
-
Microsoft’s MAI 2 Model Shocks OpenAI, Reaches Top 3
By
–
Microsoft's New MAI 2 Shocks OpenAI and Hits Top 3 https://
youtu.be/LBZc0uHtepk?si
=WZSqjrLMBlWcgGrb
… via @YouTube #microsoft #MAI2 #OpenAI #GenAI #AI #GenerativeAI #LLMs #LLM #artificialintelligence @PawlowskiMario @chidambara09 @Ym78200 @CurieuxExplorer @efipm @bigfundu @sayedflah @Ronald_vanLoon -
OpenClaw Security: How NVIDIA Addressed AI Safety Risks
By
–
OpenClaw Was Dangerous… Until NVIDIA Stepped In? https://
youtu.be/h5S2dBFLQHQ?si
=6sOExrLWxBHDj1Sn
… via @YouTube #OpenClaw #NVIDIA #generativeAI #AI #GenAI #LLMs #LLM #OpenSource @SpirosMargaris @PawlowskiMario @mvollmer1 @gvalan @ipfconline1 @LaurentAlaus @Shi4Tech @Fisher85M @kalydeoo @Ym78200 -
NVIDIA NemoClaw, OpenAI’s Pivot and Shopify Agents
By
–
NVIDIA NemoClaw, OpenAI’s pivot and Shopify agents https://
youtu.be/Ce_p69dV1jw?si
=Oa1ltOQbe6wzw5E5
… via @YouTube #opensource #openai #nvidia #nemoclaw #artificialintelligence #AI @AlbertoEMachado @Eli_Krumova @postoff25 @Khulood_Almani @anand_narang @NutritiousMind @baski_LA @TanyaSinha_ @devaang
