
OpenAI Jalapeño: First Benchmark Results Are In
OpenAI's first custom inference chip Jalapeño delivers 1.5–1.9x more work per watt and up to 3.6x lower latency than Nvidia Blackwell in early benchmarks.
41 transmissions

OpenAI's first custom inference chip Jalapeño delivers 1.5–1.9x more work per watt and up to 3.6x lower latency than Nvidia Blackwell in early benchmarks.

Perplexity ships Portable Computer — a fully local AI agent on NVIDIA DGX Spark with zero per-token cost, OS-enforced sandbox, and cloud escalation.

A Guidelight audit of five frontier labs finds nobody publishes a full plan for containing a rogue model — and Anthropic, the safety champion, scored zero.

Mudah.my car listings now appear automatically on Carousell Malaysia — up to 35,000 cars by end September. The one-listing, two-marketplace playbook.

A new independent study of real AI conversations shows that work-focused company reports filter out nearly half of what people actually do with AI.

YTL Power and JLand Group are building a gigawatt-scale data center campus in Sedenak, Johor. Here is what the deal actually covers and why it matters.

Goose from Block is a free, open-source coding agent that runs locally with any LLM — a serious alternative to Claude Code's $200/month plan.

Google's Sovereign Agent Mesh is a zero-config, zero-trust P2P overlay that lets AI agents share tools across networks without exposing private endpoints.

Cartesia released Sonic-3.6 in beta, claiming #1 on both Artificial Analysis speech leaderboards with sub-90ms latency across 44 languages.

Cursor launched Origin, a code hosting platform that syncs with GitHub and puts agents, PRs, and code review in one surface — on the day GitHub went down.

Anthropic's annualized revenue hit $65B by July 2026 — a 7x jump in one year. The AI lab is now racing toward a $2T IPO this fall.

DeepSeek released Harness under MIT license — an agent framework where models, tools, sessions, and even the agent loop itself are swappable plugins.