Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Sigilix argues that larger context windows are not the solution for coding agents, advocating instead for memory-native agents that maintain a persistent, selective memory of codebase decisions and corrections to avoid r…
A developer benchmarked four AI gateways—GoModel, LiteLLM, Portkey, and Bifrost—on runtime and deployment overhead. GoModel, a small open-source gateway written in Go, showed significantly lower memory usage, faster cold…
OpenAI released GPT-5.6 Sol, Terra and Luna on June 26, placing the frontier model family into a limited preview after a U.S. government request. The controlled launch gives government visibility and selected customer ac…
American Productivity Company's agent Righthand, powered by Claude Sonnet 4.6, achieved an 84.0% score on the Human Bench benchmark, which evaluates AI agents on realistic professional tasks requiring long-lived memory, …
OpenAI has begun a limited rollout of its GPT-5.6 Sol model to government-approved users at the request of the Trump administration, marking a shift from traditional software launches to a defense-contractor-style cleara…
A developer used OpenAI's Codex AI assistant to port the williamcotton.com website from a custom language called Web Pipe to Rust, with the AI successfully recreating the site and integrating Contentful for dynamic conte…
A writer urges people to stop using large language models to draft communications, arguing that the practice strips messages of personal style and human connection. The author emphasizes that imperfect, authentic writing…
Apple's iOS 27 beta introduces a smarter Siri AI, but the upgrade will not be available on older devices like the HomePod and Apple TV 4K, forcing users to live with two different Siri versions. Apple's recent price incr…
OpenAI released the GPT-5.6 System Card, a technical document detailing the model's capabilities, safety evaluations, and limitations. The report outlines performance benchmarks and risk mitigation strategies for the lat…
Google Cloud introduces the Gemini Enterprise Agent Platform, an agentic stack designed to help enterprises move from building single AI agents to managing thousands of them with complete governance. The platform enables…
A developer built a zero-leak Postgres MCP gateway in Go that enables LLMs to query databases without exposing raw schema or data. The gateway uses dynamic schema reflection to auto-generate tool manifests from whitelist…
OpenAI announced a limited preview release of GPT-5.6, claiming it surpasses Anthropic's Claude Mythos. The announcement, reported by @WatcherGuru, aligns with market predictions of an 83-89% probability of a late June r…
OpenAI launched a limited, US-only preview of its GPT-5.6 model after the U.S. government requested early access due to national security concerns over cybersecurity capabilities. The Office of the National Cyber Directo…
Epoch AI's new MirrorCode benchmark tests AI models on recreating complete programs without access to original code. Claude Opus 4.7 leads with a 56 percent solve rate, rebuilding a 16,000-line toolkit in 14 hours, but a…
A developer's colleague built a production app using AI-generated code without understanding its stack, highlighting the risks of 'vibe coding'—a term coined by Andrej Karpathy for accepting AI code wholesale. The practi…
Mercury Agent, a soul-driven AI agent with second brain memory, a full skill system, and multi-agent orchestration, has been released. It runs 24/7 across terminal, browser, Telegram, Discord, Slack, and Signal, featurin…
A new MCP server called Ebook-MCP supports mainstream eBook formats including ePub and PDF, enabling natural language interactions with digital books through AI-powered IDEs like Cursor and Claude. The tool allows users …
OpenAI is releasing GPT-5.6, a new generation of models comprising three tiers—Sol, Terra, and Luna—in a limited preview due to coordination with the U.S. government. The models introduce new reasoning modes and are pric…
OpenAI released a limited preview of the GPT-5.6 series with three tiers—Sol, Terra, and Luna—after the White House requested restricted access. The flagship Sol tier features new reasoning modes and is priced at $5 inpu…
OpenAI released a limited preview of GPT-5.6 on June 26, 2026, introducing a three-model suite (Sol, Terra, Luna) with pricing roughly half that of Anthropic's Claude Fable 5. The launch came less than 24 hours after the…