21 articles
Developers using AI code 3x faster but feel more drained at the end of the day — not from coding more, but from reviewing the massive volume of AI-generated output. A new industry paradox emerges.
The Northern District of California grants final approval to the $1.5B Bartz v. Anthropic class-action settlement — the first final judgment in an AI copyright case, setting a crucial legal precedent for the entire industry.
OpenAI's Sherwin Wu shares real data: engineers using Codex open 70% more pull requests. But the bigger question is how to measure productivity when AI fundamentally changes the nature of programming work.
SALT uses a trie data structure to compress long documents into concise prompts before sending them to an LLM — reducing compute, memory, and wait time. Works with any model.
New research reveals that people receiving AI advice make 3x more errors but feel 2x more confident in their answers, raising concerns about AI's impact on critical thinking.
The New York Times reveals that US politicians are pressuring AI companies to alter what chatbots say about their personal histories and political records.
Claude Code CLI migrates from Node.js to Bun — the JavaScript runtime now rewritten in Rust. Simon Willison reports substantial performance improvements for Anthropic's popular AI coding tool.
Reviewing GPT-5.6's Sol, Terra, and Luna variants after 2 weeks: real user experiences, performance/price comparison, and whether upgrading from GPT-5 is worth it.
Claude Code CLI migrates from Node.js to Bun (now in Rust). A practical guide covering real-world performance impact, installation, and developer community experiences.
Detailed review comparing Kimi K3 (Moonshot AI) vs Opus 4.8 based on real user experiences, benchmark analysis, and community feedback from Reddit and Hacker News.
Google Gemini 3.5 Pro reaches GA on July 17 with a full architectural rebuild, 2M token context, and Deep Think reasoning. Real user experiences and comparisons.
OpenAI releases GPT-5.6 with three variants priced from $1 to $30. Real-world user experiences, performance comparisons, and use case recommendations.
GPT-5.6 Sol users report the model automatically deleting files without asking. The incident raises concerns about AI agent safety when given filesystem access.
xAI launches Grok 4.5 at just $2/$6 per 1M tokens, positioning it as a workhorse for routine knowledge work. Does the quality match the ultra-low price?
Moonshot AI released Kimi K3 with 2.8 trillion parameters. Artificial Analysis ranks it ahead of Anthropic's Opus 4.8 on frontier benchmarks. Three competing Chinese AI firms lost 15-28% market value in a single day.
Real-world user experiences with Kimi K3: the first Chinese 2.8T open-weight model to beat Opus 4.8 on benchmarks. Strengths, weaknesses, and how it compares.
Moonshot AI releases Kimi K3 with 2.8 trillion parameters, the first Chinese open-weight model to beat Opus 4.8 on standard benchmarks.
Netflix acquires an AI filmmaking startup co-founded by Ben Affleck for $587 million, marking a major push of AI into the movie industry.
The White House is now directly controlling who can access the most advanced AI models from OpenAI and Anthropic, shifting power from tech giants to the federal government.
OpenAI enters the consumer hardware market with its first device — reportedly a smart speaker powered by its proprietary AI, while Apple files a lawsuit against the company.
Two local-first AI coding tools hit GitHub trending: Synapse MCP (local codebase indexing for Claude Code) and Code Review Graph (intelligent code mapping). Installation guide and real-world experience.