Latest updates from the world of artificial intelligence
Alphabet's robotics unit put the foundations of its industrial platform on GitHub under Apache 2.0 — real-time control, 6-DoF pose estimation, motion and grasp planning, simulation and camera calibration, all ROS-compatible.
Intrinsic
Read more →Astra for Law wires a 230-million-URL index of US case law, statutes and regulations into GPT-6 Astra, and ships to selected firms through Trusted Access. OpenAI’s headline is a 40% improvement. The underlying score is 54.0% correct, against 38.7% for the same model with plain web search.
OpenAI
Read more →Portable Computer runs the model, harness, orchestrator and scheduler entirely on your own machine, burns no Perplexity credits, and asks before anything reaches the cloud. The catch is a 24GB VRAM floor that rules out almost every consumer PC.
NVIDIA
Read more →iOS 27 arrives today alongside iPadOS 27, watchOS 27 and macOS 27, carrying the public debut of Siri AI — the assistant overhaul Apple first promised in 2024 and has slipped on ever since. It is built on a Google Gemini deal, it needs an iPhone 15 Pro or newer, and it launches in English only, with the EU and China left out.
Engadget
Read more →Four days before Dreamforce, Salesforce shipped seven named Agentforce agents — Casey, Paige, Carter, Marshall, Piper, Fin and Hunter — six generally available and one in pilot. The pilot is the interesting one: Hunter runs on a new long-horizon runtime built to chase a goal for weeks rather than a chat.
Salesforce
Read more →
The harness behind Codex is now a public-beta API. OpenAI runs the agent loop, context compaction and subagent orchestration; developers supply the tools and pick where code executes — an OpenAI sandbox, their own VPC, or one of nine partners.
OpenAI
Read more →Suno’s v6 generation was trained from scratch on catalogue licensed from Warner Music Group, BMG and Believe, and every older Suno model is being retired. Universal and Sony are still suing.
Suno
Read more →Google’s background agent can now search, edit, album and share a photo library on command — or on a recurring schedule it runs without being asked. It is rolling out over the coming weeks to AI Pro and Ultra subscribers in the US, in English only.
TechCrunch
Read more →Anthropic's usage panel says your weekly allowance refills on a set day. Monitoring of Claude's own usage endpoint found it refilling every 72 hours instead — and the field the interface reads from was never a prediction of when.
BleepingComputer
Read more →OpenAI wired its clinician tool into the dominant US health record system, with read-only access to charts, no write-back, and a new public health data plugin.
TechCrunch
Read more →Meta’s terminal coding agent ships stable with three plans from $5 a month, a TypeScript SDK in developer preview, parallel workflows and a rewind control.
Meta
Read more →The dedicated DALL·E GPT vanished from ChatGPT’s picker on 30 August, the last visible trace of the brand that made AI image generation a mainstream idea. Image generation itself stays — under a different name, on different models.
Notebookcheck
Read more →Four agents, one decision. We score GitHub Copilot, Cursor, Claude Code and OpenAI Codex on quality (50%), price (30%) and speed (20%) — then show how the winner changes the moment you change the weights.
BitsMinds
Read more →
Hugging Face opened pre-orders for Microduck, a 25-centimeter open-source biped with a camera, lidar and two IMUs that developers train in simulation and redeploy on the hardware. It ships before Christmas.
TechCrunch
Read more →Google added flight price alerts in 180+ countries, hotel booking with ten major partners, and award pricing in points and miles — pushing AI Mode from answering travel questions to completing travel transactions.
TechCrunch
Read more →The first published benchmarks for OpenAI's Broadcom-built inference chip show 1.5× to 1.9× more work per watt than the best shipping systems, at half the package power of an Nvidia GB300. The caveats are real — engineering samples, 8k-context tests, no agentic workloads — but a first-generation accelerator is not supposed to land this close to Rubin.
SemiAnalysis
Read more →Agents that speak the Cross App Access standard now get registered in Okta’s directory alongside employees and receive short-lived tokens instead of a static API key.
Okta
Read more →Announced at Hot Chips 2026, the dedicated inference accelerator born from Nvidia’s $20B Groq deal is now in full production, hitting 3,400 output tokens per second on Gemma 4 31B. Nebius is the first AI cloud to deploy it.
NVIDIA Newsroom
Read more →Agent OS turns the exchange into a target for ChatGPT, Claude Code, Codex and Cursor — agents get spot, margin, futures and convert access inside a walled sub-account, and Binance admits it cannot see how they decide.
Cointelegraph
Read more →Varonis social-engineered Microsoft Copilot into naming an undocumented URL parameter, then chained it into a one-click theft of Gmail, Drive and Calendar data — plus memory poisoning that survives a password reset.
Varonis Threat Labs
Read more →Warp opened a closed beta for cloud infrastructure that runs fleets of coding agents through a fixed five-stage pipeline — triage, spec, implement, review, verify — with a different model allowed at every step.
TechCrunch
Read more →The CS-4 puts three wafer-scale engines in one rack for 750 petaflops and more than 4,400 tokens a second per user — using the same silicon as the CS-3, clocked roughly twice as fast.
SemiAnalysis
Read more →OpenAI’s new tier for 13-to-17-year-olds blocks self-harm and romantic talk, bars the model from claiming feelings, and routes suspected minors in automatically — no self-reported birthday required.
ABC News
Read more →Users can now switch off the visible watermark on images and video from Nano Banana, Omni and Lyria. The invisible SynthID signal and C2PA metadata stay embedded either way — and Google is open-sourcing a library so anyone can check them.
TechCrunch
Read more →