#215 - Runway games, Meta Superintelligence, ERNIE 4.5, Adaptive Tree Search
Summary
This episode covers Cloudflare’s new default blocking of AI scrapers and Runway’s entry into AI-generated gaming. It details Meta’s aggressive push for superintelligence by hiring former OpenAI and GitHub leaders, alongside Anthropic’s talent loss to Cursor and its new economic impact research. The discussion also touches on Google’s education tools, OpenAI’s custom chip development, and NVIDIA-backed startup Emerald AI’s grid optimization software.
Topics discussed
Sponsors: OutShift by Cisco and Box AI
Intro: A quiet week in AI news
Cloudflare's bot detection and data scraping
AI in gaming: Runway, Meta, and industry dynamics
Google's new AI tools for educators
AI agents in meetings and Google's Android dominance
Google Imagen 4 and text-to-image advancements
Meta's superintelligence push and talent acquisitions
Anthropic's Claude Code and Cursor competition
Anthropic's research on AI's impact on labor
OpenAI, Google TPUs, and chip supply chain risks
Emerald AI and data center energy management
New open-source models: Ernie, Hunyuan, and Tencent
Research: Vision-language models and diffusion LLMs
Tree search algorithms for AI reasoning
Automated LLM speed running benchmarks
Cluster scheduling via text-to-text regression
LLM error correlation and bioweapon risk forecasts
AI task length horizons and capability predictions
US federal moratorium on state AI regulations
EU copyright initiatives and episode wrap-up
Listen ad-free on Castria