Google unveiled the training-focused TPU 8t and the inference-focused TPU 8i.
GM-SEUS v2 has released 3.42 million panels and a new rooftop array.
Google Cloud and NVIDIA are expanding AI Hypercomputer to grow agentic and physical AI.
TPU 8i and 8t for agents were unveiled, strengthening inference and training performance respectively.
GitHub CLI now collects pseudonymous telemetry to improve usability.
MIT and IMO have released MathNet, spanning 40 countries and 4 decades.
The little-coder and Qwen3.6-35B-A3B combo scored 78.7% on Aider Polyglot.
OpenAI briefed US and Five Eyes governments on its new cybersecurity product.
WebSocket mode made agentic workflows on the Responses API up to 40% faster.
ChatGPT's workspace agents are trigger- and tool-based AI for repetitive tasks.
A bug that caused Opus 4.7's 1M context to be truncated to 200K has been fixed.
Introduces the "Skillify" methodology, which permanently fixes AI agent mistakes by converting them into markdown skills and deterministic scripts.
WAAP unifies WAF, API security, and bot management at the edge.
Miricanvas shares the index optimization problems and solutions it encountered during its transition to Amazon DocumentDB.
Miricanvas dropped MongoDB Atlas for DocumentDB, cutting p50 latency in half and reducing costs by 30%.
Anthropic has recruited the former Novartis CEO, moving to secure expertise in AI regulation and biotech.
A Metal kernel implementation combining INT3 model compression with INT2 KV cache has been released.
Meta collects input and screen data from employee PCs to use in AI training.
Google Cloud has selected ElevenLabs as its 2026 Applied AI Partner.
By tying together product, coupon, gift, and promotion benefit data through Kafka, delays of up to 45 minutes were cut down to a matter of seconds.
Firefox 150 fixed 271 vulnerabilities found through early AI evaluation.
44% of Deezer's daily uploads were tallied as AI-generated tracks.
There are indications that Anthropic's Mythos was accessed without authorization in a 3rd-party environment.
Maestro v1.6.3 expands specialized agents and Qwen Code support.