During a recent conversation with an investor, I was asked what sounded like a simple question:
“How fast can you deliver an app?”
Without a clear brief, there was obviously no precise answer. So I gave him a reasonable range: three to six months, depending on the size and complexity of the application. Possibly longer for something particularly demanding.
His response was immediate.
“That’s too long. With AI today, you could have an app ready in a month, tops.”
And technically, he was right.
AI can already generate functional applications remarkably quickly. In some cases, much faster than a month. But speed of production is no longer the only variable that matters.
There is another bottleneck, and it is far less scalable: us.
AI may continuously become faster, more capable and more efficient, but the human capacity to process, evaluate and understand what it produces has limits. We can only absorb so much information before our ability to judge it properly begins to decline.
Suppose AI gives you a complete, working application in two weeks. Great.
But is it actually the application your users need?
Would you deploy it immediately, or would you test every important component with real users? Observe how they interact with it? Listen to their feedback? Discover which assumptions were wrong? Refine the experience, adjust the product and then test again?
That process still takes time.
AI can produce the product you requested. But AI is not necessarily the person who will use that product every day. Your customer is.
And understanding that customer cannot simply be accelerated at the same rate as generating code.
Humans weigh alternatives. We question usefulness. We reconsider decisions. We rationalize trade-offs. Most importantly, we experience products in contexts that are messy, subjective and often impossible to capture completely in an initial prompt.
By the end of our conversation, the investor understood the point I was trying to make:
The question is gradually changing.
It is no longer simply: “How fast can AI build this?”
It is becoming: “How fast can humans confidently determine that what AI built is actually right?”
AI's capacity to generate output may continue expanding dramatically. Our capacity to carefully sift through that output will not expand at the same pace.
The next great bottleneck in technology may therefore have very little to do with machines.
It may simply be the speed of human judgment.