Descriptions:
Creators Riley Brown and software engineer Ross Mike review the first four months of 2026 in AI agent development, tracking the competitive arc from Anthropic’s dominant Q1 — marked by the Claude Co-work launch and the OpenClaw open-source surge that triggered a Mac Mini shortage — through OpenAI’s Codex release, which the hosts argue has shifted momentum back toward OpenAI. The conversation is grounded in firsthand daily use of these tools for software engineering and content production, with specific inflection points called out, including Andrej Karpathy’s viral tweet about feeling “behind as a programmer” after using Claude Opus 4.5.
The discussion covers how Claude Co-work and OpenAI Codex differ in architectural philosophy — Co-work as a broad operating system layer connecting third-party tools, Codex as a focused agentic coding environment — and where each currently has the edge for different workflows. The hosts dig into OpenClaw’s memory architecture, discussing native markdown-based memory files versus portable third-party solutions like SuperMemory and MeZero, and why memory portability matters when spinning up and destroying cloud instances.
Additional topics include computer use and browser use as the shared substrate for agentic action, SpaceX’s reported interest in acquiring Cursor, the emergence of agentic payments as a new product category, and improvements in AI image generation. For developers and builders tracking the competitive landscape between Anthropic, OpenAI, and the broader agentic tooling ecosystem in 2026, this episode offers a detailed, opinionated ground-level view.
📺 Source: Riley Brown · Published May 02, 2026
🏷️ Format: Review
