I’d no sooner posted my thoughts on how AI should be enabling partnered products when Simon wrote up an interesting post by Geoffrey Litt arguing that we need “HUDs not Copilots.”

A HUD feels completely different from a copilot! You don’t talk to it. It’s literally part invisible—you just become naturally aware of more things, as if you had magic eyes.

There’s a lot to like here that echoes my post and does a lovely job surfacing the many, many decades of research in efficient collaborative activity. But…

Nobody actually is a copilot anymore

As my pilot friends are quick to point out, actual pilots no longer refer to “pilot” and “copilot.” It’s Flying Pilot and Pilot Not Flying, since both have active responsibilities, both are engaged to keep the flight safe and smooth. Sure, we should have smooth UIs that don’t require us to chat to get simple things done — the spellcheck example — and it’s pretty obvious that we can do better than a conversational command line interface. I bet we all agree we need better HUDs!

But my hope is that we can go much farther. By pushing products to be partnered with us, we can create more situations that are neither copilot nor HUD, but actually people + AI deeply collaborating on solutions, sharing the right information in the right ways, catching mistakes and insights. That’s what the Flying Pilot and Pilot Not Flying do at their best.

Don’t build parallel play

After a lifetime of building social and multiplayer experiences, I find this way of thinking doubly important, because we really have to solve collaboration differently if we’re going to effectively add AI into groups. The long game isn’t about making individuals more capable, it’s about doing it for teams, groups, and organizations. Thinking in terms of partnership gives us a starting point that I can imagine scaling in interesting ways.