What the Emails Reveal: Inside Microsoft and OpenAI's Unlikely Partnership
The Musk trial has released a 2017 email thread showing Microsoft's senior AI researchers doubted OpenAI's tech but invested anyway to stop Amazon getting them. Combined with testimony about OpenAI's safety teams dissolving, the picture that emerges is of an industry built by competitive anxiety, not conviction.
Read full story →The Model Washington Couldn't Ignore
Anthropic's Claude Mythos, deemed too dangerous to release publicly, has triggered an IMF warning about machine-speed cyberattacks and "correlated failures" across the global financial system. The Trump administration, which explicitly rejected AI oversight just months ago, is now weighing FDA-style review of advanced models.
Read full story →Competition, Not Conviction, Built the AI We Live With
The Microsoft emails are a window into how AI really gets built: not by visionaries, but by companies too scared of competitors to stop. The same mechanism that drove Microsoft's investment explains why OpenAI's safety teams were dissolved. No malice required.
Read opinion →