AI browsers were supposed to be the future. But I think something bigger is happening: Codex and Claude Code are turning into the main place knowledge work gets done.
What i've been working on: https://chorus.com/agents/88be7a2a-c5...
Lenny’s Podcast: / @lennyspodcast
Dan Shipper on X: https://x.com/danshipper?s=21
In this video, I break down Dan Shipper’s argument from Lenny’s Podcast, why tools like Claude Code, Codex, Cowork, and Proof matter, and how we’re moving from browser tabs to task tabs.
The big shift is that your AI agent won’t just sit inside an app. Your apps will increasingly be built so your agent can use them with you. That means agent-native document editors, in-app browsers, generative mini apps, and workflows where Codex or Claude can control the browser, read your context, use your tools, and help complete the task directly.
I also share my predictions for where this is going next:
Multiple browser tabs per agent thread
Agents automatically opening the right apps for each task
Agent-native SaaS replacing AI buttons inside traditional software
Generative mini apps that connect to your Gmail, Notion, Docs, Typefully, and other tools
Why you should start organizing your work at the task and SOP level
What I’ve been working on:
My team and I are building an iMessage-first AI agent that runs 24/7, connects to your tools, supports automations, and lets you use models like Claude and Codex from your phone or computer. More about me: rileybrown.xyz
Sponsorships: https://rileybrown.xyz/sponsorships