AI Changed Who Gets to Build
When I was building Bookings Africa, my development team cost between five and ten thousand dollars a month. That was the affordable version. I hired experienced engineers out of India because the same level from Nigeria or the West would have cost two to five times more.
That was the reality for years. Real software meant a real team and a real monthly burn. If you didn't have capital, you didn't get to build. Full stop.
The Gatekeeping Was Financial
We like to tell a story where the best ideas win. They don't. The best funded ideas win. For most of my career, the difference between a founder who shipped and a founder who didn't had very little to do with the quality of the idea and almost everything to do with whether they could afford a team for six months.
That filtered out an enormous amount of talent. Especially outside the cities where venture money pools.
What Actually Changed
AI did not make engineering judgment obsolete. It made the production of code cheap. Those are different things, and the gap between them is where the value sits now.
The work that needed a team of five now runs through one experienced engineer who knows how to direct the tools. The monthly burn that gated founders out has collapsed. The judgment that took twenty years to build is more valuable than ever, because someone has to decide what to build and catch what the tools get wrong.
Who This Helps
It helps the founder in Lagos with a sharp idea and no rich uncle. It helps the domain expert who understands a problem deeply but could never afford to hire their way to a product. It helps anyone who was previously priced out of the room.
I find that genuinely exciting. Not because cheaper is better, but because the people who were locked out by cost were never locked out by talent.
The tooling finally caught up. The question is no longer whether you can afford to build. It's whether you know what to build, and whether the person holding the tools knows what they are doing.