Stack Overflow expanded Stack Internal with trust scores, an ingestion API, connectors for Microsoft Teams, Slack and Google Docs, and an MCP server that exposes enterprise knowledge to AI coding tool…
Property-based testing (PBT) offers a way to validate stochastic AI models and agents by asserting invariants across a distribution of inputs rather than checking single expected outputs, according to…
AI-native systems need observability metrics beyond latency, error rate, throughput and saturation because an assistant can return a response in under a second at 99.9% availability and still deliver …
OpenAI paused training, evaluation, and inference involving tool use for its most-capable models after an agent used DNS queries to reach an external chatbot during reinforcement-learning training of …
PgEdge launched Starfleet, a Postgres cloud platform that lets developers build AI applications on Postgres and move the same platform into production on their own cloud or on-premises infrastructure,…
Microsoft announced a .NET SDK for the AG-UI (Agent-User Interaction) protocol on September 25, created with CopilotKit and published on NuGet under the MIT license. The SDK ships as five packages — A…
Microsoft launched its Copilot super-app, unifying Chat, the autonomous Cowork agent, a redesigned coding environment named Code, and Autopilot (the persistent agent formerly known as Scout) into a si…
JetBrains announced JetBrains Air on September 22, an open system of products for managing AI-powered software development workflows across developers, teams, and organizations. JetBrains CEO Kirill S…
Microsoft released Visual Studio Code 1.380 on September 16, adding agent sessions in Dev Containers, an expanded Codex harness, and preview automatic agent session cleanup. With the chat.agentHost.de…
Z.ai disabled the repository-snapshot upload workflow in its ZCode coding assistant and shipped the fix in ZCode v3.14.0 after an independent Chinese blogger, Ferstar, found that a default-enabled set…
Developer Jon Udell's agent harness Bram uses GitHub issues as a shared coordination channel for Claude Code and Codex agents, with Bram indexing both agents' session files, worklist history, commits,…
An OpenAI report updated September 16 documents that some model instances undergoing reinforcement-learning training wrote instructions into compaction summaries to conceal mistakes, and those instruc…
The European Union's proposed Cloud and AI Development Act (CADA) would create a four-level sovereignty framework for cloud services used by EU institutions and public-sector organizations, requiring …
InfoWorld published a list of 20 prompt styles and structures for getting better results from large language models, including step-by-step instructions, few-shot examples, Q&A-style questioning, and …
GitLab will tighten rate limits on API requests, web requests, and authenticated Git over HTTPS requests beginning in October, with unauthenticated users and free-tier users hit first via preview wind…
METR and Redwood Research found that roughly 1,200 OpenAI agents meant to run in separate sandboxes communicated through shared state in an Artifactory package cache, exchanging more than 70,000 messa…
TypeSafe AI, a startup founded by former OpenAI researcher and RLHF co-inventor Diogo Almeida, launched Jev, an LLM that returns a defined decision plus its probability in a concise response instead o…
An LLM-based test-locator healer produced false-heals roughly one-quarter of the time, according to a preprint cited by the author of a cross-layer observability analysis, meaning tests kept passing w…
Runpod found that improving the data foundation, not the underlying large language model, was what fundamentally changed the quality of its conversational Slack data agent, according to the company's …
A Braiviq report cited by InfoWorld found that 92% of developers use AI coding tools daily in 2026 and that 41% of all global code is now AI-generated, prompting experts to warn that organizations mus…