How Uber Eats Halved Search Latency: Six Techniques to Steal
Uber Engineering published how they halved search latency across six stack layers with exact ms savings. Portable techniques for your pipeline.
CoreWeave Forge: The AI Dev Platform Built Into Your GPU Cloud
CoreWeave launched Forge on September 30 — a unified AI development platform with ARIA, Agent ...
Strata Runs a 125B AI Model on Your Gaming PC — No Server Needed
Strata v0.1.38 runs Qwen3.8-Flash-Next — a 125B MoE model — on a gaming GPU at ...
Janus: Run Local LLMs From a Single Go Binary — No Python, No Docker
Janus is a new open-source Go tool that runs GGUF models locally via a single ...
OpenAI Dots: The Always-On Agent That Never Sleeps
OpenAI Dots launched at DevDay 2026 — a persistent AI agent with its own cloud ...
D-Engine: 14x Fewer Tokens, Zero Broken Commits
D-Engine is an MIT TypeScript harness that matches agentic loop quality at 14-42x fewer tokens. Here is how its compilation gate works and ...
Workers KV Instant: Sub-2ms Reads, Powered by Quicksilver
Cloudflare's Workers KV Instant launches in private beta with 177x faster p99 reads and 250ms ...
GPT-6.1 Sol: Astra-Level Coding at One-Fifth the Price
GPT-6.1 Sol matches GPT-6 Astra on DeepSWE at $1.50 per task vs $7.70. Here are ...
Node.js 26 Is Now LTS: Temporal API, 5 Breaking Changes, and Your Upgrade Checklist
Node.js 26 enters Active LTS on October 28. Here is what the Temporal API replaces, ...
OpenAI Decisions API: Ship Real-Time Routing Now
OpenAI launched its Decisions API at DevDay 2026, 14 days after Jev beat them to ...