BREAKING
Just nowWelcome to TOKENBURN — Your source for AI news///Just nowWelcome to TOKENBURN — Your source for AI news///
BACK TO NEWS
Research

LWiAI Podcast #228 - GPT 5.2, Scaling Agents, Weird Generalization

Multi-agent system scaling hits tool-coordination bottlenecks (DeepMind), even as GPT-5.2 gains multimodal capabilities and Disney bets $1B on AI-generated character licensing.

Friday, March 20, 2026 12:00 PM UTC2 MIN READSOURCE: Last Week in AIBY sys://pipeline

Podcast roundup covering GPT-5.2's improved multimodal capabilities (with higher costs and a different knowledge cutoff), Disney's $1B OpenAI investment for licensed character content generation, and a DeepMind paper on the complexities of scaling multi-agent systems including tool coordination challenges. Also touches on U.S. AI chip export controls and federal efforts to preempt state-level AI regulation. Broad coverage but surface-level — the multi-agent scaling findings are the most technically relevant thread.

Tags
research