What does it mean to work with AI agents in a fundamentally multiplayer way? I can review a pull request by chatting with the EXACT SAME agent that was used to develop it (which cost almost $700 lol). I can see the code diff right there, and leave comments to the agent on it. I can also easily pull in the human developer with an @-mention. And I never have to check code out locally. The live app preview is running on the agent's sandbox, so I can use it if I don't trust the agent's (or the QA agent's) screenshots/videos of the feature.

For over a year, my team has been working with AI agents in a fundamentally different way than most. All of our Claude/Codex/Pi/etc agents run in the cloud, and any session is joinable by anyone on the team. Each of our meetings has an agent that launches subagents to do research, draft posts, and implement features as we discuss things. A Chief of Staff agent lets me know what's waiting for my review, and can talk to any person or agent on the team to resolve bottlenecks. Working in this fundamentally multiplayer way has given us a preview of the way everyone will work soon, so I wrote up a short manifesto explaining • the current problems • principles for a great solution • some things that are tricky to get right Check it out, and let me know what you think! multiplayer-ai.com

Agreed. Adding an agent to a group chat is a start, but not the end. We've been thinking about multiplayer AI for a while and wrote up five principles that a great solution should have. 1. Never copy-and-paste Most work doesn't happen in chat, but in docs, emails, Salesforce, GitHub, etc. The agent has to live next to the work, and everyone involved in the work has to be able to talk to the same agent, instead of endlessly copying and pasting back and forth. 2. Work with the door open When chat sessions are private, good prompts and good habits spread slowly. When they're shared, everyone gets better fast. Shopify's internal agent is only reachable from public Slack channels (@tobi called it "learning on the shop floor".) 3. Continuously improve When someone writes a prompt that nails it on the first try, the rest of the team should learn from it. When the agent needs correcting, that correction should turn into a reusable skill. When the same workflow runs over and over, its quality should start being measured automatically. 4. People are not routers A human should never be asked a question the agent already has the answer to. Chasing status updates and relaying answers between people is work for an agent. 5. Nothing starts from scratch When sessions live in the cloud, you pick up every project exactly where you left off. Every doc, plan, and PR should have a session behind it you can resume months later. A new hire is productive on day one. There are a few more considerations, on security and build-vs-buy, on the site. Curious to get your take! multiplayer-ai.com

💪 Grok 4.6 joins the Pantheon of our benchmark, together with Opus/Fable 5 and Kimi K3. 💪 A good and fast model, and my new daily driver. Check out the video for more details, an explanation of how our benchmark works, and how you can make your own!

You can now just forward an email to an agent and get an email back when done. Here I have GPT 5.6 redline a contract, which it emails me back when done. Hard to go back from this workflow!