Oct 16, 2026 · 14 min
AI spending faces scrutiny as local models and music filters advance
Apple & Meta Bring Muse to Mac, Anthropic Opens Up & Deezer Filters AI Music
The episode connects doubts about AI infrastructure returns with practical choices about model privacy, independent evaluation, and control over AI-generated music.
- 1Investor skepticism is growing around whether massive AI infrastructure spending can deliver sufficient returns.
- 2Anthropic gives METR employee-level access to models and safety tools, strengthening the case for independent evaluation.
- 3Apple and Meta’s local Mac models prioritize privacy and cost, while music platforms move toward AI labeling and filtering.
Don't miss
Anthropic’s decision to give METR researchers employee-level access turns independent model evaluation into the episode’s clearest test of research transparency.
The brief
The episode opens with a question hanging over the AI boom: can enormous infrastructure spending produce returns large enough to satisfy investors after the hype fades?
Anthropic’s decision to give METR employee-level access to its latest models and internal safety tools makes independent evaluation more practical, even as the episode questions dramatic safety messaging.
Apple and Meta are collaborating on capable local models for Mac hardware, a notable turn from their privacy and advertising disputes that could reduce costs while keeping more processing on-device.
Deezer and YouTube Music are moving toward labeling or filtering AI-generated music, but the episode notes how difficult it is to define AI involvement consistently.
The through line is listener and user choice: independent researchers should inspect powerful systems, Mac owners should run models privately, and music audiences should control what they hear.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.