Product is the new bottleneck.

Six months ago, in a planning meeting at Loyco, we'd argue about whether engineering could ship something in Q3. Today the same kind of feature lands on Friday.

That sounds great, and it is. But it's created a problem we weren't expecting.

When engineering can ship in days, the longest pole in the tent moves. It's no longer "can we build this?" It's now "should we build this, in this shape, for these users, this week?" And that question — surprise — is what product and design have always been responsible for answering.

The friction map flipped

What used to be a measured rhythm — design sprints, spec reviews, two-week iterations — now feels desperately slow compared to the velocity engineering has unlocked. We're shipping faster than we can decide. Which means, in practice, that some of the things we ship are not the things we'd ship if we'd thought about them properly.

The new bottleneck is product and design. Not engineering.

This is unsettling because it requires re-staffing in directions most companies haven't planned for. The teams that nailed engineering throughput for the last decade now find themselves staring at a queue of half-baked product decisions and a design org that wasn't sized for this velocity.

Three implications I keep returning to

Design teams need to be much larger, or much faster, or both. Most companies' design orgs were sized to keep up with the old engineering cadence. That cadence is now five-to-ten times faster. Design as a discipline has been quietly behind for nine months, and most companies haven't noticed yet because product managers are still gating the queue.

The "approval" model is breaking. When the lead time was weeks, you could afford a final sign-off step. When the lead time is hours, that step becomes the bottleneck. Decision-making has to be pushed down, distributed, and made cheaper. Trust, not process.

Product managers need to live in tooling, not in Slack. The most productive PMs I see right now spend their time in Linear, Figma, Cursor — wherever the actual product is taking shape. The ones still optimising for Slack visibility are getting left behind.

I think the next 24 months sort this out, probably noisily. Org charts haven't caught up to the production speed. The companies that resize their product and design orgs first will look like geniuses by 2027. The ones that don't will be confused about why their engineering velocity didn't translate to revenue velocity.

The bottleneck wasn't engineering. It hadn't been for a while.