yes soon on open router.
heard on prompt caching
INNOVATION
-
Open Router Deployment and Prompt Caching Soon
By
–
-
Codex Improves Long Context Management Capabilities
By
–
codex has gotten very good at long context management:
-
Wafer Memory Optimization for LLM Inference Parallelization
By
–
lots of wafers = lots of mem for kv cache
multiple users at a time overlapping the pipeline -
Google Integrates Gemini 3 into Search with Generative UI
By
–
In case you missed it, our latest Google AI: Release Notes episode is out! Google Search VPs Robby Stein and Rhiannon Bell join @OfficialLoganK to deep dive into the integration of Gemini 3 into Search.
— Google AI (@GoogleAI) 9 janvier 2026
Chapters:
– Intro
– What is Generative UI?
– From static to generative… pic.twitter.com/XuNh8apRt7In case you missed it, our latest Google AI: Release Notes episode is out! Google Search VPs Robby Stein and Rhiannon Bell join @OfficialLoganK to deep dive into the integration of Gemini 3 into Search. Chapters:
– Intro
– What is Generative UI?
– From static to generative -

AI Foundation Model Understands 3D Depth in 360° Panoramas
By
–
What if a single AI could understand the 3D depth of any 360° panorama? Insta360 Research & UCSD present Depth Any Panoramas (DAP), a new foundation model. They built a massive, diverse training set and designed a model that learns robust, real-world distance cues without
-

Datadog Uses OpenAI Codex for Incident Prevention
By
–
How @datadoghq uses Codex for preventing incidents: https://
openai.com/index/datadog/ -
Claude Code changes perspective on small tasks worth automating
By
–
One of the bigger shifts for me with Claude Code over the past few months has been shutting down that initial dismissal I have when a task feels "not worth my time" Like I'll think "it would be nice to rename all my screenshots with what's actually in them" and immediately move
-

Visual Tokenizers Enhanced: Semantic Objectives Beyond VAE Pixel-Level Accuracy
By
–
VAE-only visual tokenizers has hit a wall for diffusion, as better pixel-level accuracy does NOT lead to higher-quality generation To be truly effective, this MiniMax paper adds semantic objectives via CLIP + self-supervision during tokenizer pretraining, having the latent space
-
Reachy Mini Robot Featured at NVIDIA Jensen Huang CES Keynote
By
–
was fun to see reachy mini closing @nvidia Jensen's keynote at CES this week! pic.twitter.com/xgHKqwp1Qu
— clem 🤗 (@ClementDelangue) 9 janvier 2026was fun to see reachy mini closing @nvidia Jensen's keynote at CES this week!
-
Understanding Composite AI: Advances in Machine Learning
By
–
What Is Composite AI?
#AI #AIio #AIInnovation #ML #DataScience #Futureofwork @timnitgebru @oriolvinyalsml @ceobillionaire @soumithchintala @waitin4agi_ @sallyeaves @bernardmarr