#232 - ChatGPT Ads, Thinking Machines Drama, STEM
Summary
Hosts Jeremy and Andre discuss AI safety concerns regarding authoritarian capture and OpenAI’s new ad-supported ChatGPT Go tier. They cover industry drama at Thinking Machines, where co-founders returned to OpenAI, and analyze China’s progress with Huawei-based AI models. The episode also examines global chip supply constraints, highlighting Samsung’s rising role as an alternative to TSMC.
Topics discussed
Sponsors and Intro
AI Safety, Open Source, and Authoritarian Risks
Governance, Alignment, and Grok
OpenAI Ads, Free Tier, and Education Impacts
China's AI Landscape: Ernie and Baidu
Thinking Machines Drama and Anthropic
Chinese Chips, Huawei, and Global Supply Chain
xAI's Colossus Supercomputer and Energy
New Startups and Black Forest Labs Image Model
Open Source Music Models and Swiglu Architecture
Reasoning Models and Test-Time Scaling
Why LLMs Aren't Scientists Yet
US Senate AI Bill and Safety Probing
Anthropic's Claude Constitution Release
Artist Lawsuits and Cultural Impact of AI
Listen ad-free on Castria