Summary
Riley Brown reviews Google’s new Gemini 4 Neon model. In some tests it matches or beats Claude Opus 5.5 and Fable 5.2 at a much lower price. The video cites benchmark data from Parallel showing Neon at the top of its leaderboard, and says it costs less than half as much as Fable 5.2 and GPT 6.2 Astra.
Brown recommends using the model through Google Antigravity, which they describe as Google’s equivalent of Claude Code and Codex, and shows a realistic smoke simulation created in Blender with Neon. They also describe a three-tier view of AI tools: chat apps, persistent cloud agent platforms and full desktop agent apps. Brown argues that users should move from the first tier toward the latter two.
The second half covers agent platform news, including Cognition’s Devonbots. The platform grew out of the company’s acquisition of Interaction, the maker of Poke, and supports iMessage and Slack, 235 connections, automations and cloud Mac sessions that can build and upload iOS apps to TestFlight. The video also notes the strong negative reaction to OpenAI’s GPT 6.2 Astra and updates to Grockbot and Dots.
📺 Source: Riley Brown · Published November 20, 2026
🏷️ Format: News Analysis







