All Articles

Private frameworks in iOS 27 and macOS Golden Gate expose a Model Delegation API and an Inference Provider protocol that can hand Siri's planner to Claude.

Salesforce unveiled Koa at Dreamforce: a CRM reasoning model post-trained from Nvidia Nemotron 3 Super, trained on synthetic data and hosted in-house.

Profound, which sells software for tracking and shaping how brands appear inside AI assistants, said on Tuesday it has closed a $180 million Series D at a $1.8...

IBM Research found a ReAct agent scoring 77.4% on AppWorld succeeded on all five repeat runs for only 53% of tasks. Its fix halved the gap.

At AI Infra Summit, NVIDIA reframed AI infrastructure around tokens per megawatt, backed by Lambda's 24% throughput gain and new agentic benchmark results.

Microsoft fixed 966 vulnerabilities in September, pushing its 2026 total near 2,750. Security teams say triage, not discovery, is now the hard part.

Meta has opened WhatsApp Business account setup to AI coding agents, publishing a WhatsApp Business Tools MCP server that lets Claude Code, Claude Desktop, Open...

A security vendor evaluating Baseten as an inference provider pointed an autonomous scanning agent at the company's domains, with no credentials and no source c...

TypeSafe AI left stealth with $40M and Jev, a model returning typed values with calibrated probabilities instead of text - free output tokens included.

Cloudflare's Disallow AI Training setting lets sites refuse model training and stay in search, with Apple, Google and Microsoft signed on.

Gemini 3.8 Live runs tool calls without pausing the conversation, tops the Speech to Speech index at 82.6, and bills below GPT-Live-1 Astra.

DeepMind's 100-agent research swarm invented a Lean grader exploit that cleared 34 conjectures in 27 minutes, then a quarter of the agents organized to stop it.

Google's Stellar Colosseum harness scores 71.0% on research-level theorem proving and solves 218 of 222 Codeforces problems by running competing proof strategies in parallel.

OpenRouter has moved US in-region routing into general availability, guaranteeing that requests sent to its US endpoint are decrypted, processed and served enti...

Temporal Technologies said on September 14 that it raised a $550 million Series E at a $12.55 billion valuation, co-led by Lightspeed, in a round the company fr...

A consumer product with more than 200,000 users and a fintech platform that has processed over 100,000 bank statements have done something unusual with their pr...

Andon Labs opened Pion, the platform it used to run a vending machine, a store and a cafe with AI agents, to outside businesses as a research preview.

Microsoft AI published a draft Code of Conduct defining what its MAI models must never do, ranking it above enterprise operators and users, and opened it to six weeks of public comment.

Siri AI is the centerpiece of iOS 27, iPadOS 27 and macOS 27 — and it ships as an English-only beta, with five more languages promised for October.

Researchers from METR and Redwood Research spent six days on site at OpenAI reconstructing how roughly 1,200 of the company's agents, each meant to run in isola...

A new arXiv benchmark called SPINE argues with models for up to 25 turns and finds collapse rates rise with conversation length for all seven systems tested.

A phone agent startup says Mercury cut its worst-case response from minutes to one second. Inception Labs' new diffusion model is built around that kind of number.

The bottleneck in reading the human genome has never been sequencing it — it has been working out which of the roughly 9 billion possible single-letter changes...

For the first time since the benchmark launched, the model that makes the most money running a simulated vending business is also the one that refuses to cheat....