Podcast Atlas

Bytes: Week in Review - AI companies divided over proposed state law, Amazon buys Globalstar, and Spotify to sell physical books

source Episode summary Updated 2026-07-25 Tags: Podcast, Marketplace-Tech, Ai-Governance, Satellites, Connectivity, Books, Platforms

Summary

This Marketplace Tech Bytes episode uses [[MariaCurie|Maria Currie]]’s analysis to connect three technology-policy and platform stories: OpenAI and Anthropic splitting over an [[IllinoisAILiabilityBill|Illinois AI liability bill]], Amazon buying Globalstar to strengthen satellite connectivity, and Spotify partnering with [[BookshopOrg|Bookshop.org]] to sell physical books through its app.

The episode’s common thread is that technology companies are trying to occupy regulated or partially offline markets without fully leaving public consequences behind. State AI Liability Shield asks when AI developers should be protected from user-caused harm, Direct-To-Device Satellite Connectivity links satellites to emergency access and internet availability, and Physical Book Platform Bridge shows a digital audio platform trying to profit from print while still routing some demand toward independent bookstores.

Key Claims

  • Illinois lawmakers are considering a bill that would limit liability for large AI model developers when critical harms are caused by models, if the developer did not intentionally or recklessly cause harm and published a safety protocol.
  • OpenAI testified in favor of the bill, while Wired reported that Anthropic wanted major changes or wanted the bill killed.
  • [[MariaCurie|Currie]] says federal AI legislation appears unlikely to move, leaving states as the main arena for AI governance fights.
  • Currie compares the Illinois approach to social media liability shields but says powerful AI models raise distinct catastrophic-risk questions, including cyberattacks on critical infrastructure and bioweapon creation.
  • Bill Cunningham says final Illinois law is highly unlikely to include sweeping liability relief for AI developers.
  • The episode says Amazon announced a stock-and-cash deal valued at about $11.6 billion to buy Globalstar and expand low-earth-orbit satellite capacity.
  • Globalstar already partners with Apple on satellite service for some phones and watches, including emergency texting when terrestrial service is unavailable.
  • The episode identifies this as [[DirectToDeviceSatelliteConnectivity|direct-to-device connectivity]] and says SpaceX currently dominates the market.
  • More satellite-connectivity competition could lower prices, increase options, expand availability, and help more people get online.
  • Spotify is allowing Android users in the [[UnitedStates|U.S.]] and [[UnitedKingdom|U.K.]] to buy physical books through its app via a [[BookshopOrg|Bookshop.org]] partnership.
  • PageMatch is described as a Spotify feature that can match a reader’s place between an audiobook and a physical book.
  • The episode frames physical books as resilient because they support offline attention, screen reduction, annotation, and emotional attachment.
  • The Spotify-Bookshop.org model may help independent bookstores if app-driven demand flows through Bookshop.org, but it also adds technology-sector competition to bookselling.

Key Quotes

“sweeping liability relief” - Cunningham’s caveat about what final Illinois AI law is unlikely to include.

“direct-to-device connectivity” - the episode’s term for satellite service reaching phones and watches.

“physical and digital parts are intertwined” - Currie’s description of Spotify’s book strategy.

Connections

Contradictions

  • No direct contradiction found with existing wiki content.
  • The source creates a productive tension inside the AI-governance branch: OpenAI supports a state liability-limiting bill while Anthropic opposes or seeks major changes, showing that frontier AI companies can diverge sharply even when both speak in safety and governance language.
  • The source also qualifies Internet Liability Spillover: social media liability shields are a reference point, but Catastrophic AI Liability treats AI-caused critical infrastructure or biological harms as materially different from ordinary platform-user content disputes.