#234 - Opus 4.6, GPT-5.3-codex, Seedance 2.0, GLM-5
Summary
Hosts Andrei Kurenkov and Jeremy Harris review a week of major AI model releases, including Anthropic’s Opus 4.6, OpenAI’s GPT-5.3 Codex, and Google’s Gemini 3 DeepThink. They discuss the shift toward agentic workflows, the impact of new hardware partnerships like Cerebras, and the rise of high-fidelity video generation from ByteDance. The episode also covers emerging models from Chinese firms and the evolving landscape of coding assistants.
Topics discussed
Intro, host updates, and sponsor reads
Listener comments and corrections
OpenAI Opus 4.6, Codex, and recursive self-improvement
Anthropic Claude 3.5 Sonnet and benchmark analysis
ByteDance video generation and Alibaba image models
GLM-5 release and DeepSeek engineering updates
Cursor IDE, XAI API, and SpaceX integration
ElevenLabs valuation and AI API margin economics
API gateways and Runway's video editing tools
Humanoid robotics funding and One X robot
OpenAI consumer strategy and Anthropic enterprise focus
Research: Matrix decomposition and RL training methods
Model scaling laws and performance predictions
AI alignment, misalignment, and reward misspecification
Outro and closing remarks
Listen ad-free on Castria