Excited to try GPT-5.2-Codex in Droid!
LLMS
-
Top Trending Model on Hugging Face Platform
By
–
Top trending model on HF for the past few days! https://t.co/nXICZFlqbe
— clem π€ (@ClementDelangue) 14 janvier 2026Top trending model on HF for the past few days!
-
LLMs for Revenue Prediction: Confidence Without Mathematical Accuracy
By
–
https://
linkedin.com/pulse/beyond-h
ype-gartners-ai-heat-map-reveals-where-falls-short-moioli-ndixe
β¦ Using an LLM to predict next quarter's revenue is like asking a poet to do your accounting. It sounds confident, but the math is a hallucination -

Jamba2 Achieves Competitive Hallucination Rate on HHEM Leaderboard
By
–
New benchmark signal: Jamba2 places high on @vectara
's HHEM leaderboard for hallucination rate competitive with open models in its size class. Check out the leaderboard: https://
huggingface.co/spaces/vectara
/leaderboard
β¦ -
World Models and the Future of AI-Driven Glasses
By
–
I see it coming back eventually, particularly as world models get more and more developed. However, we're still one to two years away from really seeing amazing world models that will bring utility to AI-driven glasses. The next couple of years are sure going to be interesting!
-
Google Updates Veo 3.1 Across Gemini and More
By
–
Google is rolling Veo 3.1 updates across Gemini, Flow, AI Studio and APIs.
— π¨ AI News | TestingCatalog (@testingcatalog) 13 janvier 2026
– Vertical formats support
– Veo 3.1 Ingredients to Video
– Improved ingredients to video consistency
– Upscaling to 1080p and 4K across all Veo models
– Verification of AI-generated videos in Gemini https://t.co/FMGClzXPxb pic.twitter.com/Wc3VI0rIS7Google is rolling Veo 3.1 updates across Gemini, Flow, AI Studio and APIs. – Vertical formats support
– Veo 3.1 Ingredients to – Improved ingredients to video consistency
– Upscaling to 1080p and 4K across all Veo models
– Verification of AI-generated videos in Gemini -
Efficiency challenges in AI world models
By
–
I see some breakthroughs coming on compute, so we'll see. That's the big question: can they get the world models to be very, very efficient?
-

Technical Writing Framework for Large Language Models Era
By
–
A framework for technical writing in the age of LLMs https://
buff.ly/ToYl7uD
#AI #MachineLearning #DeepLearning #LLMs #DataScience -

Improving Self-Evolving LLM Agents Learning Capabilities
By
–
On building more powerful self-evolving agents. LLM agents struggle to learn from experience after deployment. Fine-tuning is expensive and causes catastrophic forgetting. RAG retrieves based on semantic similarity alone, often pulling noise instead of what actually works.
-

Google’s Auto Browse for Gemini
By
–
Google is working on a new tool for Gemini called "Auto Browse". Auto Browse will Chrome into a proper agentic browser, where Gemini will serve as a control interface. This addition is a continuation of agentic functionalities announced in September.