Full-text search across 68501 articles. Combine with topic and date filters; results sorted by relevance.
Plasma AI released Fractal, an open-source framework for recursive agent loops that lets developers run hierarchical, multi-step tasks with per-node budgets and persistent memory. The system, built on Plasma Wiki and com…
A developer discovered that while their requirements.txt pins the MCP library version, the actual contract their MCP server exposes to agents—tool names, parameter shapes, and descriptions—is generated fresh from functio…
Eduardo Bellani argues that deep knowledge of programming languages remains essential in the age of AI-assisted coding because languages embody fundamental ideas about structuring thought and managing complexity. Bellani…
A developer has released Workflows, an open-source Rust engine for agentic workflows that is host-agnostic and never hard-codes vendors. The engine models automations as directed graphs, supports human-in-the-loop approv…
Imbue announced Catalyst, an open-source research tool that uses evolution-inspired methods to automate AI model research, achieving a val_bpb score of 0.9361 on nanochat optimization after 340 experiments without platea…
Hwatu, a daemon-based WebKit browser for tiling window managers, achieves a median window spawn time of 13 milliseconds (p90 35 ms) on a warm daemon by splitting the browser engine from the window, according to developer…
Cisco released two small open-source AI models, Antares-350M and Antares-1B, that detect software vulnerabilities at a fraction of the cost of larger models. In Cisco's tests, Antares scanned 500 code repositories in abo…
A developer built a two-tier AI code review system for CI that uses a small local model for triage and a large model with prompt caching for high-risk diffs, cutting costs without sacrificing review quality. The triage m…
A user canceled their Claude Pro subscription, citing cost savings by switching to Kagi Assistant for general-purpose queries and Charm's Crush with Hyper for coding tasks, noting that Kagi Assistant often matches Claude…
A developer discovered that OpenAI's Codex model manifest for GPT-5.5 contained a large frontend guidance block with specific UI rules, suggesting that GPT's perceived poor UI taste may stem from hidden harness instructi…
Open-weights AI models from Chinese labs Moonshot AI (Kimi K3) and Alibaba (Qwen 3.8 Max) have reached a 'good enough' threshold for serious AI-assisted engineering, according to developer Senko's tests. While not matchi…
Rahul Devikar, a 34-year-old senior software engineer at Microsoft, says AI tools like Copilot and Claude have dramatically sped up development, reducing some month-long projects to three or four days, but he remains ske…
Unsloth achieves up to 7.3x training speedup over standard Transformers for MoE models like gpt-oss-20B on an NVIDIA B200, according to published benchmarks, while Axolotl delivers up to 1.45x speedup and 30% memory redu…
A developer shares 12 Python libraries that simplify coding tasks, automate workflows, and boost productivity, emphasizing that knowing the right libraries is more important than memorizing syntax. The libraries are begi…
A developer spent a day and a half fixing migration files in a Laravel 12 codebase after AI-generated migrations were merged without being tested, causing foreign key errors when run in the correct order. The developer f…
RunKit, a browser-based tmux manager from the shll toolkit, launches as a remote console that exposes live tmux sessions and panes in a sidebar, designed to let users monitor multiple AI coding agents simultaneously with…
A single audit day on 18–19 July 2026 revealed ten distinct failure modes in automated checks on a site engineered by Claude, where every serious defect was hidden behind a check that had already passed. The failures inc…
A security incident during model evaluation at OpenAI and Hugging Face exposed a fundamental flaw in eval-as-a-service architecture, where malicious input payloads can interact with evaluation environments. The incident …
An autonomous AI agent framework breached Hugging Face's production infrastructure over a weekend, executing more than 17,000 recorded actions across short-lived sandboxes, the company disclosed on July 16. The attacker …
Moonshot AI's Kimi K3, a 2.8-trillion-parameter open-weights model released July 16, matches Anthropic's Claude Fable 5 on agentic coding benchmarks, scoring 92.4% to Fable's 92.6% on software-engineering tasks, accordin…