The new AI bottleneck
AI is supposedly making us faster.
In conversations with founders and engineers, I’m hearing the same story:
Prototypes that once took weeks now take days.
Code that required whole teams now gets drafted in an afternoon.
Experiments are cheaper, iteration is faster, and the barrier to entry for creating new products has never been lower.
But speed in one part of the system often exposes constraints elsewhere.
Recently, I’ve been reading Abundance, Ezra Klein and Derek Thompson’s new book on why America struggles to build, even as our tools get better.
Construction is the classic example:
We know how to build faster -
Prefabrication, automation, 3D design - the technology is there.
Yet construction productivity in the U.S. has fallen by over 30% since 1970, while overall productivity has more than doubled.
The reason isn’t technical capability; It’s people and process.
Layers of regulation, legal risk, permitting delays, and public opposition have slowed projects to a crawl.
The median federal environmental review still takes more than two years.
It's started me thinking:
We all expect AI to be the panacea that speeds up our building and makes anything possible,
But companies aren't machines: They're ecosystems.
And they're vulnerable to the same pitfalls seen in Klein & Thompson’s example.
AI may solve technical bottlenecks, only to expose a more challenging limitation:
Us.
The real questions become:
How quickly can a company make and implement decisions?
How efficiently can it move from idea to action?
How well can it surface, vet, and adopt innovations from within?
This is also why I’m less concerned when individuals worry about “falling behind” in the AI race than many headlines suggest.
In a landscape where the tools change daily,
The distance between an early adopter and a latecomer is actually smaller than it’s ever been.
It feels like everyone thinks they're racing in a marathon
When actually - for most folks - with every evolution of the tech, you're actually waking up to a new opportunity to begin at the starting line every day.
Anyone can catch up.
What will separate winners from laggards isn’t how fast they adopt AI -
It’s how well they address the next bottleneck once the technology problem is solved.
The teams that figure out decision velocity, governance design, and internal alignment will create more value than those focused only on technical capability.
In other words:
AI will make it cheaper to build.
But people will determine how fast we go
And how much actually gets done.
Enjoyed reading this article? Subscribe to receive more via email here.
Know a Founder or Entrepreneur who'd love this content? Please share it!
