Test-time compute scaling is simple but revolutionary: Instead of making models bigger, you make them think harder during inference. The model generates multiple reasoning paths, verifies answers, backtracks when wrong, and improves solutions in real-time. It's thinking, not
TECHNOLOGY
-
OpenAI’s o1 demonstrates smarter models via inference-time compute
By
–
OpenAI's o1 proves you can make models smarter by making them "think longer" at inference not training bigger models. DeepSeek, Google, Anthropic all pivoting to test-time compute. Training wars are over. The inference wars just started. Here's the paradigm shift happening
-
OpenAI Releases GPT-5.2-Codex Model
By
–
OpenAI released GPT-5.2-Codex model on the Responses API, and it is now available on OpenAI's Platform.
-
OpenAI partners with Cerebras for AI compute boost
By
–
BREAKING : OpenAI is partnering with Cerebras to add 750MW of ultra low-latency AI compute to its platform. “That means faster responses, more natural interactions, and a stronger foundation to scale real-time AI to many more people”
-
OpenAI Partners with Cerebras for Advanced AI Computing
By
–
OpenAICerebras https://
openai.com/index/cerebras
-partnership/
… -
Software-based optical correction reduces hardware complexity and costs
By
–
This is incredible engineering and makes a lot of sense. Autorefractor machines can already get your prescription. Oculus similarly got a leg up on other HMDs at the time by correcting distortions in software instead of hardware. Lowers the complexity and cost by a lot.
-
New Operating System for Autonomous Space Operations
By
–
There is a new operating system for autonomous space operations:
-
FiloBot: Revolutionary Robot That Grows Like Plants
By
–
What if robots didn’t move through the world, but grew into it? That’s what makes FiloBot so intriguing. Inspired by climbing plants, it doesn’t navigate by walking or rolling. Instead, it extends its body forward, growing where it needs to go. Here’s the key idea:
FiloBot -
Jensen Huang Bets Everything on Physical AI Next Decade
By
–
Suite à la dernière interview de Jensen Huang il est maintenant clair que NVIDIA mise TOUT sur l'IA physique pour la prochaine décennie.
— VISION IA (@vision_ia) 14 janvier 2026
Interview commentée à retrouver ici : https://t.co/rmVWFzaxic https://t.co/BYOk7C77dC pic.twitter.com/kk484JUqc2Suite à la dernière interview de Jensen Huang il est maintenant clair que NVIDIA mise TOUT sur l'IA physique pour la prochaine décennie. Interview commentée à retrouver ici : https://
youtu.be/RfhFro35kTk -
Google Tests Gemini’s Fast Path to Production
By
–
Google is testing "The fastest path from prompt to production with Gemini" on AI Studio. A new homepage UI to select from various options like Chat, Build and API Key creation.