AI coding assistants and agents, testing and debugging with AI, and practical setups for developers.
GPT-5.6's threefold jump in ARC-AGI-3 scores came not from a new model but from two API settings: reasoning retention, which keeps the model's chain of thought across multi-step tasks, and compaction,…
A viral project showed a language model producing watercolor paintings by writing JavaScript with p5.brush rather than generating images directly. This article recreates Surya Narreddi's recipe fully …
Funes is an open-source Hugging Face tool that transforms coding agent session logs into a searchable, portable memory layer. With a single command, it indexes agent turns locally, provides recall and…
This tutorial shows developers how to improve schema compliance in small local language models by fine-tuning LiquidAI's LFM2.5-350M using GRPO with TRL and LoRA. Using ~500 Nemotron structured-output…
This post demonstrates a distributed asynchronous GRPO training setup using TRL v1.14's AsyncGRPOTrainer with LoRA on Hugging Face Jobs. The trainer and vLLM inference replicas run on entirely separat…
GPT-5.6, OpenAI's latest flagship model, is changing how startups build AI agents. By deliberately routing requests across reasoning effort levels and using the Responses API, teams cut costs and late…
Replit has introduced Free Mode, a token-cost-free way to use its AI development platform, powered by the GPT-5.6 Luna model. Users can describe apps in plain language, iterate, and fix errors without…
GPT-5.6 is now available in Kiro, offering developers stronger reasoning and coding capability at improved price-performance. The model supports the full software lifecycle—planning architecture, buil…
This article explores how UK-based online travel agency loveholidays adopted OpenAI Codex, an AI coding agent, to make software development accessible beyond its engineering team. Engineers delegate r…
Anthropic's redesigned Claude Code Projects feature introduces AI orchestration for cloud-based developers, enabling parallel agent threads with persistent conversations, shared memory, and cross-repo…
An MIT researcher pairs GPT-5.6 Sol with Codex to run quantum computing experiments largely autonomously. The AI agent controls lab hardware, performs repetitive qubit calibration, analyzes measuremen…
A step-by-step tutorial showing developers how to convert an existing Python script—such as a website monitor—into an AI agent. By adding the OpenAI Agents SDK and a simple @function_tool decorator, e…
AI can generate impressive code, but building great software has never been just about writing code. Drawing on Paul Ford's argument, this article explains why the flood of AI-assisted projects often …
commit-rewriter v0.1 is a web app that helps open source maintainers clean up AI-generated Git commit messages. It lets you review and edit messages through a simple web interface, automatically rewri…
Version 0.2.2 of the datasette-explain plugin adds support for displaying SQLite explain plans on read-only stored-query pages in Datasette. The update was prompted by Simon Willison upgrading his pub…
llm-keys-ui v0.1 is a Simon Willison plugin offering a web UI for storing API keys on remote machines, so coding agents like Codex can retrieve them via llm keys get without secrets appearing in chat …
Comment My comment on MCP was always a bad idea? — Hacker News This article entirely misses the value that MCP brings to
