Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Kunlun Tech

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 2

  1. Kunlun open-sources Matrix-Game 4.0, real-time game world model on one RTX 5090

    Kunlun Tech's Skywork AI released and open-sourced Matrix-Game 4.0 on October 15, describing it as the world's first game world model that runs in real time on a single consumer GPU. The system splits the work: a coding agent generates the game code, the game engine maintains world state and interaction rules and emits a structural frame as a constraint, and a video model renders the final image in real time — a loop the company says runs from "code-generated game logic to structural constraints to live video rendering." Kunlun says it delivers 480p at 60 FPS on one NVIDIA RTX 5090 with under 150 ms input-to-frame latency and about 18.5GiB peak VRAM; planned work includes a refiner for higher resolution, consistent multi-player viewpoints and more game genres. Code and models are published on GitHub, Hugging Face and a project site.

    IT之家 AI · 🔥 0
  2. Alibaba Cloud unveils Context Engine, ApsaraLakebase and Agent Context

    At the 2026 Apsara Conference main technical forum, Alibaba Cloud proposed supporting agent deployment across three dimensions — Model, Harness and Context — and introduced a Context Engine that turns enterprise-wide data into real-time context agents can use. It also launched or upgraded the ApsaraLakebase open lakehouse storage, the AI-native big data service ADA, the enterprise agent-oriented Agent Context, plus PAI-RLS for reinforcement-learning training, the PAI-InferX inference platform and the AIDBS data-agent platform. Alibaba Cloud said new RDS PostgreSQL instances grew more than 400% quarter over quarter, with about 80% created by agents and roughly 50% of input/output traffic agent-driven; it also claims Agent Context can improve token efficiency by more than 3x and raise accuracy on complex knowledge tasks by up to 50 percentage points — all vendor-reported figures.

    InfoQ 中文 AI&大模型 · 🔥 0

Experience and discussion from the community

Share my Kunlun Tech experienceAsk about Kunlun Tech

Nobody has shared their experience with Kunlun Tech yet.