Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A developer revived a shelved poker game project after discovering Claude Code, an AI coding tool that runs locally. Managing the project with text files and design docs, the developer built a system called FlowGate to h…
Qt Creator 20 was released on June 17, 2026, adding an ACP Client extension for AI coding agents via a chat panel, supporting Claude Code, Codex, Gemini CLI, and GitHub Copilot. The update also introduces a Zen Mode, GN …
A developer released 'laravel-gemini', a lightweight Laravel package for integrating Google Gemini API. The package supports text generation and chat features with a simple, clean interface for PHP 8.1+ applications.
Apple's iPadOS 27 introduces Siri AI with handwriting indexing, Image Playground for AI image generation, Visual Intelligence on iPad, AI-powered Shortcuts, and performance improvements, but requires an M-series or A17 P…
OpenAI has rolled out a new Scheduled page in ChatGPT's sidebar, allowing users to view, pause, edit, and delete scheduled tasks. The update makes tasks faster and more reliable, and supports monitoring tasks that proact…
A Hacker News user asked the community which AI models power their products' daily features, revealing that many teams choose models based on cost and latency rather than rigorous evaluation. The user shared their own us…
Stanford researchers introduced DeLM, a decentralized language model framework that allows AI agents to coordinate directly without a central orchestrator, reducing inference costs and latency. The system uses a shared k…
Anthropic released a major update to its Claude Design tool on May 19, 2026, adding design system imports from codebases, round-trip code handoffs, and doubled weekly token limits across all subscription tiers. The overh…
Apple may make its Apple Intelligence AI suite mandatory in iOS and macOS 27, according to reports from the first beta. The change would force users to accept the AI tools, which consume several gigabytes of storage and …
A developer has released Relego, a free, self-hostable alternative to Readwise that allows users to upload Kindle highlights and receive daily or weekly recaps via Kindle or email. Built with .NET and AI assistance, the …
OpenAI launched ChatGPT Tasks in beta on January 14, 2025, enabling Plus, Team, and Pro subscribers to schedule automated prompts, reminders, and recurring briefings via GPT-4o. The feature supports up to 10 active tasks…
Selora AI Local, an open-source Qwen-based model for Home Assistant, has been released. The model features four specialized LoRA adapters for answers, clarifications, automations, and commands, running locally via llama.…
GitHub's standalone Copilot App went generally available today for all subscribers on macOS, Windows, and Linux, introducing Cloud Automations that let agents run scheduled tasks on GitHub-hosted infrastructure without r…
Philosopher and AI researcher David Thorstad critiques Ted Chiang's recent essay denying machine consciousness, arguing that Chiang's claims are unsupported by evidence and that the possibility of AI consciousness remain…
China released Kimi K2.7 Code, an open-weight coding model that is six times cheaper than Claude and claims to self-evaluate its own performance, though independent verification of its benchmarks is impossible.
A new survey by the American Psychological Association finds that 77% of psychologists report patients using AI chatbots for mental health support, raising concerns about a 'sycophancy trap' where agreeable AI may valida…
Google Play Books will launch a generative AI Book Insights feature on July 6, automatically enabled for most English-language books, allowing readers to request recaps or ask questions about the text. The opt-out model …
A developer built a roguelike game where the dungeon master is a large language model that runs entirely in the player's browser using WebLLM and WebGPU, eliminating server costs and enabling offline play. The architectu…
A bootcamp developer discovered that using large language models for data extraction from PDF invoices can be dramatically cheaper than expected, with some models costing fractions of a cent per call. Testing showed that…
A developer tested four AI-powered data pipeline tools—dbt Cloud with Vanto AI, Apache Airflow 2.9 with Claude, Prefect 3.0, and Fivetran—over six weeks on a real microservices ETL system. dbt Cloud with Vanto AI saved 1…