Ep. 216: Google I/O, Musk v. OpenAI Verdict, Andrej Karpathy Joins Anthropic & Meta Layoffs

TL;DR

• Google is turning search and Gemini into an agent stack — At I/O, Google launched Gemini 3.5 Flash, previewed Gemini Spark as a 24/7 personal agent on a dedicated cloud VM, and expanded Search into a multimodal, conversational interface that already reaches more than 1 billion monthly users.

• Demis Hassabis intentionally put “singularity” on the table — The hosts stressed that DeepMind’s CEO chose his words carefully when he said Google is at the “foothills of the singularity,” tying that claim to working agents, code generation, and AI accelerating science and math.

• Musk lost the OpenAI case fast, but the bigger story may be xAI’s pivot — A judge dismissed Musk’s $150 billion lawsuit on statute-of-limitations grounds after less than two hours of jury deliberation, while Paul argued Musk’s real move is transforming xAI/SpaceX into AI infrastructure, including a reported $1.25 billion-per-month Anthropic compute deal.

• Andrej Karpathy joining Anthropic is a power move for recursive self-improvement — Karpathy, who coined “vibe coding” and has shaped major AI conversations from World of Bits to autonomous research, is joining Anthropic’s pretraining team to help use Claude itself to accelerate model development.

• The AI job shakeup is no longer hypothetical — Meta cut about 8,000 workers after reorganizing 7,000 more into AI-focused groups, while Cloudflare and ClickUp explicitly said AI is changing org design now, especially for “measurer” roles like middle management, ops, finance, and compliance.

• Public backlash is forming around AI’s infrastructure and social legitimacy — Gallup found 71% of Americans oppose a local AI data center, younger people show high concern about jobs, environmental impact, and institutional distrust, and the hosts warned that AI leaders increasingly risk being cast as the villains of this story.

The Breakdown