#207 - GPT 4.1, Gemini 2.5 Flash, Ironwood, Claude Max
Summary
This episode covers OpenAI’s GPT-4.1 release, optimized for coding, and new ChatGPT memory features. It discusses Google’s Gemini 2.5 Flash and Ironwood TPU, alongside xAI’s Grok-3 API launch. The hosts also analyze Meta’s Llama 4 benchmark controversy, Anthropic’s $200 Max tier, and major funding rounds for Safe Superintelligence and Thinking Machines.
Topics discussed
Sponsors: Box and Outshift by Cisco
Intro and episode overview
OpenAI GPT-4.1 models and benchmarks
ChatGPT memory feature and pricing tiers
Google Gemini 2.5 Flash and Pro
xAI Grok 3 launch and pricing
Canva Visual Suite 2.0
Meta Llama 4 Maverick benchmark controversy
Google Ironwood TPU and infrastructure
Google Cloud Pro subscription
Safe Superintelligence funding and TPU access
Mira Murati's Thinking Machines startup
Hugging Face and Unitree robotics partnership
OpenAI Texas data center and tax breaks
OpenAI Deep Research benchmark
Alibaba SeedThinking reasoning model
Research: Sample Don't Search and Concise Reasoning
Research: Detecting data contamination
Grok 3 benchmark performance analysis
Elon Musk vs OpenAI legal battle
Meta China AI allegations and conclusion
Listen ad-free on Castria