Anthropic

Anthropic accidentally published a 59.8MB source map file. The contents reveal an ambitious product roadmap and a persistent supply chain vulnerability.

New Critique feature uses GPT to draft responses and Claude to fact-check them, achieving a 57.4% score on deep research benchmarks

From GPT-2 in 2019 to Mythos in 2026, AI labs keep warning us their models are too powerful for public release. Then they release them anyway. At some point you have to wonder if this is safety or marketing.

The Claude maker is in early talks with Goldman Sachs, JPMorgan, and Morgan Stanley. At $60 billion, it would dwarf OpenAI's expected offering later this year.

Sam Altman's company aims to expand from 4,500 staff in one of tech's most aggressive hiring pushes, as rival Anthropic captures a fifth of its business through coding tools.

The definitive guide to Australia's leading Answer Engine Optimisation agencies. Who actually delivers results in AI search, and who is just adding AEO to their service menu.

Two new models bring computer use and coding performance at a fraction of the cost. The real story is the system architecture they enable.

Jack Conte asks the question that destroys the generative AI legal defense: If training is fair use, why are you paying Disney?

New data shows 90% of Australian security teams face pressure to loosen access controls, even as AI-generated attacks surge to record levels.

Atlassian, WiseTech, and Block have collectively cut thousands of positions since late February. Each company cited artificial intelligence. Markets rewarded them.