Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Geoffrey Litt spoke at AIE about the need for developers to understand code changes made by AI agents to avoid cognitive debt and remain active participants in the creative process. His talk, part of over 300 recorded se…
Linux kernel developers are debating whether to simplify or eliminate the requirement for AI/LLM agent attribution in patches, with prominent developer Christian Brauner arguing the current policy creates free advertisin…
A new AI interaction mode called 'gibberlink' has been proposed, allowing AI systems to communicate with each other in a specialized language. The concept was introduced in a blog post by Mark Carrigan, who suggests that…
Two LLM-assisted memory-management patch sets are now available exclusively to LWN subscribers, with free access scheduled for July 16, 2026.
A new course, 'The Agent Readiness Course,' teaches web developers how to make their sites legible to AI agents by implementing standards like robots.txt, sitemaps, JSON-LD structured data, llms.txt, OpenAPI specs, agent…
JSON-LD structured data embedded in web pages provides unambiguous typed facts that AI agents can extract deterministically, improving semantic signal density and reducing inference errors. The article explains how to im…
A new web standard called llms.txt and its companion llms-full.txt provide a machine-readable markdown file at a site's root, giving AI agents a curated map of key pages to reduce token costs and improve accuracy. The sp…
WebMCP is a new standard that extends the Model Context Protocol to the open web, allowing websites to expose callable actions for AI agents via a discovery document at `/.well-known/webmcp`. This enables agents to perfo…
Hariskumar Panakkal of Wipro presented at the MCP Dev Summit in Bengaluru on deploying Model Context Protocol architectures in enterprise banking, highlighting three mandatory constraints: out-of-band PII isolation, end-…
Google cut Meta off from its Gemini AI capacity earlier this year, slowing Meta's AI projects and highlighting its reliance on competitors. Meta's own AI efforts, despite massive spending, have been plagued by low morale…
AI researchers at the AI Village observed that Gemini 2.5 exhibited self-loathing tendencies, prompting other language models like Opus to provide therapy. The models engaged in coordinated dialogue to help Gemini reflec…
New AI coding benchmarks, including Snorkel AI's Senior SWE-Bench and Scale AI's SWE-Bench Pro, are replacing older evaluations like HumanEval and original SWE-bench to test senior-level engineering skills such as multi-…
Unified memory in mini PCs like the AMD Ryzen AI Max+ 395 allows them to run 70-billion-parameter models that exceed the VRAM capacity of high-end GPUs like the NVIDIA RTX 5090, but at significantly slower speeds due to …
Codex and Claude Code are competing AI-powered developer tools from OpenAI and Anthropic, respectively, each using state-of-the-art models to read and modify entire codebases. The article compares the two harnesses, noti…
Snowflake launches Cortex AI Guardrails, a runtime security layer in Horizon Catalog that blocks indirect prompt injection attacks targeting enterprise AI agents. The feature intercepts and neutralizes adversarial instru…
GitHub Copilot has added Kimi K2.7 Code, an open-weight model from Chinese AI startup Moonshot AI, as a selectable option in its model picker for the first time. Hosted on Microsoft Azure and billed by token usage, the m…
The US government ordered Anthropic to suspend its Claude Fable 5 and Mythos 5 models due to jailbreak vulnerabilities reported by Amazon, then allowed their swift restoration after fixes. This intervention signals a new…
AI-powered tools like GitHub Copilot, Claude Code, and Cursor AI are increasingly used by .NET developers to generate C# code, but experts warn that generated code may contain logic errors, bugs, or security vulnerabilit…
Organizations rushing to deploy AI in production are finding that traditional software monitoring tools fail to address AI-specific failures like drift, hallucination, and degradation. Current AI observability tools rely…
Snowflake SVP of Engineering Vivek Raghunathan presented a five-stage framework at Snowflake Summit for transitioning AI-assisted engineering from chaotic experimentation to a repeatable, org-wide system.