## Session IX: Looking forward

> [“AI didn’t break async work. It vindicated it.”](https://open-and-async.com/q/ai-vindicated/)

### Discuss

- Where would an AI agent get stuck in our workflow today — a vague issue, an undocumented decision, a thin test suite? And what does that gap reveal about how we onboard humans, too?
- Where has writing a clear request to an AI forced you to actually define the problem first — and where did it expose thinking that was fuzzier than you realized?
- The book warns: outsource your thinking and you outsource yourself. Where’s the line for us between AI sharpening judgment and quietly replacing it?

### Split the room

**For managers:** Would you hand an AI agent an issue from your backlog exactly as it’s written? If not, what’s missing — and would a brand-new human hire hit the same wall?

**For individual contributors:** What’s one task you’d delegate to an AI teammate through an issue → branch → pull request loop? What context would it need that isn’t written down yet?

### Debate

“AI changes everything about how we work.” Argue it. Then argue the book’s counter: the clarity an AI needs is the same clarity a teammate in Nairobi needs at 2 a.m.

### Try it

Take a real task and write it up as an issue clear enough that either a new teammate or an AI agent could run with it, no meeting required. If you can’t, that’s the gap to close.

---

From the [*Open & Async* book club guide](https://open-and-async.com/book-club-guide/#looking-forward).