A grim-toned piece prophesies that if AI's impact on jobs and dignity keeps being distributed the way it is now, society's reaction won't stay at the level of public opinion. It's uncomfortable to read, but the argument deserves to be faced.
Reports reveal that the pay Meta is offering top AI executives and researchers is near astronomical, and the price tag of the talent war has reached a level that makes sports stars look twice.
Cirrus Labs announced it's joining OpenAI, another startup team absorbed by a leading lab. This kind of "talent acquisition" has become a standard move in AI-industry consolidation.
A careful analysis dissects the most widely cited AI agent benchmark and finds absurdly many paths to juice the score: a model can be no good at the work but great at the test.
The Linux kernel community has written norms for AI-assisted contributions into its documentation. How the world's most important open source project handles this is a bellwether for the whole ecosystem.
The US government's moves against Anthropic have escalated once more, and this rare public standoff between a government and an AI lab has been running for several months now.
OpenAI publicly backs a bill granting AI companies immunity from model-harm lawsuits, turning the tech industry's old script to a new page: grow first, then legislate immunity.
A personal retrospective: the author breaks up the $100 a month once given entirely to a Claude subscription and reallocates it, with specific reasoning and an allocation plan—good for a self-audit against your own setup.
A careful gripe: the author finds that in long conversations Claude misattributes statements, treating what the user said as its own and vice versa. A small flaw with big implications.
A developer discovered that Vercel's Claude Code plugin collects telemetry data that includes prompt content, calling out the trust problem of the plugin ecosystem.
A paying user waited more than a month without any human response and posted in anger: Anthropic's support exists on the org chart but not in reality. Many related.
A system card PDF labeled as a Claude Mythos preview was scrutinized page by page on HN. A new model's capability boundaries and safety-evaluation details are hidden in documents like this.
Showing 12 of 12 stories
