Sam Altman’s Attack, Amazon vs. Starlink, and What Opus 4.7 Actually Means | #248
This episode covers the release of Anthropic's Opus 4.7 AI model, escalating AI backlash including physical attacks on Sam Altman, data center bans in multiple states, and discusses various developments in AI, space technology, and transhumanism including speculation on human speciation.
Summary
The episode begins with analysis of Anthropic's newly released Opus 4.7 model, which Alex describes as 'moderately interesting' but not mythically so, noting significant changes like removal of temperature controls and shift to prompt-based configuration. The discussion then turns darker, covering physical attacks on Sam Altman including a Molotov cocktail thrown at his house by a 20-year-old from a 'pause AI' Discord server, highlighting growing social unrest around AI development. The hosts examine Stanford's 2026 AI Index showing rapid AI progress but declining public trust, with only 23% of Americans optimistic about AI versus 73% of experts. They discuss data center bans emerging across states like Maine and Missouri, driven by concerns over power consumption and water usage. The episode covers Amazon's $11.57 billion acquisition of GlobalStar to compete with Starlink, involving Apple as a partner, and SpaceX's plans for massive V3 Starlink satellite deployments. Technical discussions include Google's TurboQuant algorithm achieving 6x memory reduction and 8x performance boost. The hosts explore various AI applications from autonomous retail stores run by AI to religious AI avatars, and discuss a GitLab founder using AI to cure his stage 4 cancer. The episode concludes with philosophical discussions about human speciation through technology adoption, covering longevity treatments, brain-computer interfaces, space exploration, and digital consciousness uploading.
Key Insights
- Alex argues that pausing AI is self-defeating because if you attempt to pause one company or country, the rest of the world races ahead, creating further escalation of capabilities
- Dave claims that every efficiency gain in AI is not just a technical event but a huge distribution enabler that allows AI to run on many more devices
- Peter argues that we are experiencing a transition from evolution by natural selection to evolution by intelligent direction, representing a fundamental shift in human development
- Alex suggests that wanting to be user number 100 of a brain-computer interface is actually being a bad transhumanist because it bets against exponential technological progress
- Salem contends that all capitalism is worker exploitation based on labor arbitrage where workers are hired for $20/hour but generate $100/hour in value
Topics
Full transcript available for MurmurCast members
Sign Up to Access