You Don't Need to Work Faster, You Need to Work Differently

"How will people and organizations change as a result of AI adoption?"

I've been having this conversation a lot lately, and I hear many different opinions. The below is what I currently think. It'll probably be wrong in two years — and I hope it is.

Consider it loosely.

It'll be a function of how well you and your company leverage AI to achieve actual outcomes. Not busy work. Outcomes.

How well are your agentic workflows performing? How much relevant, correct work can agents do before hitting the human-in-the-loop phase? How good is each individual at directing and evaluating agent output?

The people who get great at working with and directing their agents will be the most valuable and productive people around.

Teams are the size they are today because it felt like a comfortable number to move at pace without a lot of overhead.

Agents changed that. If you keep the same team size but each person now operates alongside a team of agents, productivity increases exponentially. Unless teams shrink, you get exponential overhead to match — conflicts, lack of coordination, diminishing returns per person.

The ideal team is three. Five max. Beyond that, ROI per person drops.

I also believe in paying 40% above top of market for three exceptional people, rather than paying "just" top of market for five.

If each team is three to five people, companies will need to be more fragmented as a result. How exactly that impacts organizational structure — how things are divided, how groups relate — I don't fully know yet.

What I do know: companies who have fewer but better people will be the ones to win.

Picture this: you have three exceptional people. They use AI optimally, they ship major features daily. That creates a new problem — there's only so much you can ship for a particular product. Only so much polishing you can do.

If your best people are always in execution, who spends time understanding what should come next?

I think entire departments will form around this question. Call it the ideas department. Call it strategy. These will be mixes of people from strategy backgrounds, product managers who lean strategic, consultants, data scientists, researchers. They'll work closely with leadership to translate company vision into executable work — projects, milestones, priorities.

I truly hope this means today's product managers move toward strategy and away from sprint control. Removing tickets, cutting scope, shipping MVPs just to get something out — that doesn't matter anymore.

Products have the scope they have today because companies had to make hard prioritization calls. They'd pick things adjacent to what they already do — leveraging existing customers, reducing costs, faster time to market.

But if teams are smaller and faster, and you have entire departments thinking about what to build — product scope naturally grows.

Look at Lovable. They won't just be a platform for vibe-coding websites. They're building an entire creative suite serving different segments across business and prosumer.

Deck, which we started and are still very much focused on, is about optimizing a company's scattered feedback. But the business outcome of improving a product goes beyond processing feedback. You have to make it actionable, build it, release it, track whether it worked. If we're moving fast, there's no reason to stop at just doing a good job processing user feedback. We have to go beyond that.

When it's faster and cheaper to release software, the scope of what companies can credibly take on gets much larger.

Linear announced last week that issue tracking is dead. They're focusing on orchestrating agents rather than just tracking projects. I find that overdue, especially since I'd started using my Linear account mostly to track agent context for my Ralph Wiggum loops.

It's 2026. Coding agents do a better job than humans at implementing technical solutions given the same context. The days of going over a ticket backlog to make every ticket extremely precise so a human can implement it — those are gone. You change or you die.

Linear did the right thing by leading that change. I expect others will follow.

The hard part is what comes next. People still need to coordinate — making sure everyone works smoothly, on the right things, with minimal overlap. What does project management look like two years from now for companies spearheading this change? That's probably the one thing I still don't understand.

By now I hope we can agree that things will change. There's no doubt about that. The only doubt is how.

What will it really look like to work five years from now? What will you wake up to? What will your tasks be? What will a person who is good at their job does differently than one who isn't?

No one truly knows.

Keep an inquisitive, open mind, and we may one day find out.