LIVE DISPATCH

A new voice-cloning model just closed the gap with human speech

Independent benchmarks show the latest release matching studio-quality narration in under 30 seconds of reference audio — and it's free to try this week.

AUDIO AI · 4 min read · Updated 12m ago
Funding

Coding-agent startup raises $40M to automate PR reviews

38m ago
Launch

A new spreadsheet copilot ships with formula-explaining hovers

1h ago
Research

Paper claims a 3x jump in reasoning benchmarks from a small model

2h ago