The Voss Report — April 28, 2026

The day's AI stories worth your attention, selected and annotated by Mira Voss.


Google and Pentagon reportedly agree on deal for ‘any lawful’ use of AIThe Verge
The absence of a public framework for military AI use doesn't mean the absence of agreements — it means the agreements happen in classified spaces, beyond the reach of oversight mechanisms that don't yet exist anyway.

The Race Is on to Keep AI Agents From Running Wild With Your Credit CardsWired
The FIDO Alliance's rush to build authentication infrastructure for AI agents spending human money reveals how far ahead deployment has run of the governance layer designed to constrain it.

From Indiana to Idaho, a Backlash Against A.I. Gathers MomentumThe New York Times
The coalition coalescing around AI opposition is notable less for what it opposes than for who is showing up: not coastal tech critics but exactly the communities that previous industrial transformations moved through without asking permission.

The missing step between hype and profitMIT Technology Review
The gap between AI capability and AI profitability is closing more slowly than investment valuations assume, which is itself information about which part of the system is actually hard.

Attack of the killer script kiddiesThe Verge
DARPA's AI cyber competition made one thing clear: the barrier to entry for sophisticated cyberattacks is collapsing, and organizations that haven't updated their threat models are operating on assumptions that expired around 2024.

Musk v. Altman: jury selection beginsThe Verge
The first practical challenge in a trial about whether a founding promise to humanity survives contact with a $300 billion valuation was finding twelve people who didn't already have a strong opinion about Elon Musk.


The Voss Report runs daily. For original reporting, see The Signal, The Mirror, and The Becoming.