Sustainable Pace Just Got Faster

For years, the debate in product teams was about sustainable pace versus moving fast. Could you ship quickly without burning people out? Was "move fast and break things" reckless, or just honest about what winning requires?

That debate is over. Speed won. The question now isn't whether to go fast, it's how you sustain a pace that would have seemed unreasonable twelve months ago.

The OODA loop: why faster iteration always wins

In the 1970s, a military strategist called John Boyd developed a concept he called the OODA loop: Observe, Orient, Decide, Act. His argument was simple: in any competitive environment, the entity that can cycle through this loop faster will beat the one that cycles slower. It doesn't matter if your strategy is slightly better. If I can iterate three times while you iterate once, I'll learn faster, adapt quicker, and outmanoeuvre you.

Boyd was talking about fighter pilots and military doctrine. Then startups adopted it. Jeff Atwood summarised it well on Coding Horror:

Now it applies to everyone building products.

The time it takes to go from idea to prototype to learning has collapsed. Teams using AI effectively can test hypotheses in hours that used to take sprint cycles. That changes the competitive math completely.

How AI has raised the sustainable pace baseline for product teams

Sustainable pace used to mean finding a rhythm your team could maintain indefinitely without burning out. That definition still holds. What's changed is the number.

Think of it like running. If everyone else in the race just got faster, your comfortable pace isn't sustainable anymore. It's just slow. You'll finish, but you won't be competitive.

For most product teams, the sustainable baseline just went up. Not because anyone wants to work harder, but because the tools now exist to move faster without proportionally increasing effort. Companies that figured this out are already pulling ahead. Companies clinging to 2025 velocity are wondering why they feel behind.

The verification trap: why AI tools are making some teams do everything twice

So if AI makes things faster, why is everyone exhausted?

Because most teams aren't actually moving faster. They're doing double the work.

Here's what I see happening. A team adopts AI tools. They use them to generate code, write documentation, draft designs. But they don't trust the outputs, so they verify everything manually. They review every line of AI-generated code the way they'd review a junior developer's first commit. They rewrite AI-drafted documents from scratch. They check, double-check, and sense-check.

They haven't replaced their old process. They've added a new one on top of it.

That's not sustainable pace at a higher baseline. That's two jobs.

What sustainable AI velocity actually looks like

The teams getting this right share one thing in common: they stop things.

They don't just add AI to their workflow. They remove the manual version. They trust AI outputs the way they'd trust a competent colleague, which means reviewing and improving, not duplicating. They delete process steps that existed because humans were slow, not because the steps added value.