Every article we have published, newest first. Search titles and leads, or change the order.
ModelsNews
China's MiniMax released M3.1 Flash on Monday. The company says it is a generational leap in agentic tasks, coding, computer use and cybersecurity. Online inference is nearly three times faster, and usage costs are 70% lower than its predecessor.
· 3 min read
ToolsNews
Microsoft has extended its Project Perception security system to AI agents running inside companies. It calls them "agentic insiders": they have access to data, tools and systems, so they need continuous supervision.
· 3 min read
ModelsNews
doubleAI, the new company of former Mobileye CEO Amnon Shashua, has unveiled its first product. Maxwell is what doubleAI calls a code execution model: instead of writing code, it speeds up running programs and formally proves every change equivalent to the original.
· 2 min read
ToolsNews
AWS has introduced TorchNeuron, an open source native PyTorch backend for its Trainium chips. Hugging Face libraries already support it, so Trainium can be used like any other accelerator. The public beta starts on October 20.
· 3 min read
ToolsNews
Modal will add Nvidia's Arm-based Vera CPU to power its Sandboxes product. In Modal's own tests, Vera delivered up to twice the task throughput of an equivalent x86 Intel GNR machine under load. Modal has until now run only on x86.
· 3 min read
BusinessNews
Microsoft has terminated its agreement with Nscale for the planned AI data centre in Loughton, Essex, as the Telegraph first reported on Sunday. Power is the issue: the grid operator says it cannot supply enough capacity to the site until the 2030s.
· 3 min read
ToolsNews
World, the identity project co-founded by Sam Altman, unveiled a World ID suite for banks and financial firms at Money20/20 on October 18. World ID for Zoom is now generally available, and AI agents can carry delegated proofs that a real human stands behind them.
· 3 min read
ToolsNews
Google Cloud AI Research has open-sourced Harness², a framework that leaves the model alone and improves the harness around it: prompts, memory, skills and scripts. A single step raised GLM-5's score on WorkBuddy-Bench by 11.5 points.
· 3 min read
ToolsNews
vLLM 0.32.0, the new release of the popular open-source inference engine, shipped on October 18. It adds an endpoint for structured decisions and runs low-precision KV cache on more GPUs. Check the breaking changes before you upgrade.
· 2 min read
BusinessNews
Anthropic announced on October 18 that it is committing part of its equity to the newly created Anthropic Public Benefit Fund. Subject to stockholder approval, 5% of its shares go in now. For every 50% increase in the company's value above a set baseline, the board can add another 2%.
· 3 min read
ModelsAnalysis
We asked two questions. Which AI model gives you the most intelligence per dollar? Claude Haiku 5.5 wins with 217 index points per dollar. Which of the five smartest models is cheapest? Claude Sonnet 5.5, at a blended price of $4.
· 5 min read
BusinessNews
Gartner says that if AI providers keep spending at the current pace, more than $3.5 trillion of AI infrastructure could be exposed to impairment by the end of 2030. It also expects more than 80% of neocloud providers to cease independent operations.
· 2 min read
ToolsNews
The team behind OpenCode, the popular open-source AI coding agent, disclosed a security breach on October 17. It says the incident is tied to a large-scale distillation attack and affected 1.4% of OpenCode Console users.
· 2 min read
ModelsNews
Paper Instruments unveiled Ultramarine 1 Flash, an agentic model tuned for professional knowledge work. It builds on the open GLM-5.3 Flash model and, on the company's internal benchmark, costs $0.13 per task while outscoring Claude Fable 5.
· 2 min read
PolicyNews
On October 14 the EU's Official Journal published the first harmonised standard under the AI Act, EN 18286:2026. It covers quality management systems, and complying with it gives a presumption of conformity with the corresponding AI Act requirements.
· 3 min read
ToolsNews
OpenAI launched ChatGPT Slides on October 16. It creates presentations in ChatGPT Space, lets teammates edit them in real time and exports finished decks to PowerPoint, PDF or Google Slides.
· 2 min read
ToolsNews
Anthropic launched OpenTelemetry trace export on the Claude Platform in public beta. Calls from Claude Managed Agents, the Messages API and Claude SDKs now show up as standard traces in your own observability tool.
· 2 min read
ModelsNews
Mistral released Timestral, its first family of time-series foundation models. The largest variant has 209M parameters and, according to Mistral, ranks first on all three main public forecasting benchmarks. Inference does not need a GPU.
· 2 min read
ToolsNews
PostgreSQL 19 was due in September and is now expected on November 19. Percona developer Zsolt Parragi says AI-driven code reviews found many of the issues behind the reverted features, while the number of security fixes also surged.
· 2 min read
ModelsNews
China's StepFun released open weights for Step 5 Preview on October 15, its flagship model for agentic work. It is a sparse MoE with 600B total parameters, 27B active per token and a 1M-token context window.
· 2 min read
