Akamai Launches Edge AI Inference and Agent Sandbox — WebAssembly Isolation for Agent Code
Akamai shipped a managed inference service and a WebAssembly-based agent sandbox in beta. Here is what the two-tier isolation model means for developers ...
LiteLLM Shipped Malware on October 21 — 49K Downloads Before Anyone Noticed
On October 21, 2026, someone with maintainer access to LiteLLM pushed a poisoned release — ...
Mistral Large 4: 1T Open Weights Drop — Free to Use
Mistral dropped open weights for Large 4 today. 1T params, free to deploy. Here is ...
tstack: 9 Agent Skills That Automate Your Entire PR Workflow
Theo Browne open-sourced tstack - 9 MIT-licensed agent skills for Claude Code and Codex that ...
Anthropic’s sandbox-runtime: Sandbox AI Agents Without Containers
Anthropic open-sourced the OS-level sandbox behind Claude Code as an npm package. Here’s how to ...
Anthropic Replaced Git with Scion: 30,000 Merges a Day
Anthropic built Scion to replace Git after AI agents drove merges from 2,500 to 30,000 per day. Seven engineers, five months, 200x smaller ...
Qwenputer: A RISC-V CPU Now Lives Inside LLM Weights
A developer compiled a full RISC-V CPU into an LLM's weights. Qwenputer outperforms tool calls ...
Claude Mythos Built Working Exploits for Chrome, Safari, Windows, and Linux
Anthropic's Claude Mythos Preview autonomously built five working exploit chains across Chrome, Safari, Android, Windows, ...
Rime Prism: The First Fine-Tunable Speech-to-Speech Model
Rime Prism is the first speech-to-speech model developers can fine-tune, run on-prem, and audit via ...
GitHub Octoverse 2026: Agents Now Open 40% of Pull Requests
GitHub Octoverse 2026 report reveals agents started 40% of pull requests in September — up ...