The AI Advantage Doesn't Belong to the Youngest Person in the Room
Stone Atwine on The Grinders Table Podcast
Everyone is talking about vibe coding. Actually, we are not just talking, we are building! I’ve spent the last few months between Claude Code and Cursor, and I have enjoyed every bit of the experience.
Founders are shipping in five days what used to take eight engineers and twelve months. The barrier to building software has essentially collapsed. And the narrative forming around all of this is straightforward: the younger you are, the more native you are to these tools, the bigger your edge.
On a recent episode of The Grinders Table podcast, I spoke with Stone Atwine, CEO and Co-Founder of Eversend, a fintech operating across 12 African markets with nearly 1.5 million users. Stone is not someone who romanticises the past or resists new tools. Two months before we sat down, he was publicly saying: Don’t vibe code fintech. He meant it.
Then he spent 14 days building something with Cursor and Claude and his view changed completely. The conclusion he reached is not the one most people would expect.
He didn’t say the youngest engineers win. He said the opposite. His words: “Some of the biggest beneficiaries of this AI revolution are going to be people who’ve been around. People in their forties. You know your industry, you know your space.”
He compared himself to his nephew, who is sharp, YC-bound, fluent in every AI tool. Then he said: “I have a massive, massive advantage because of the industry knowledge I have under my belt.”
That’s the line I keep returning to.
The bottleneck in building products has shifted. It used to be: Can you write the code? That constraint is nearly gone. The new bottleneck is: do you understand the problem well enough to build the right thing?
In financial services, Stone’s world, and one I work close to as an investor, that gap is significant. He framed it simply on the podcast: how does a card transaction actually work when you swipe? What happens in the milliseconds between tap and approval? Which players are involved, where does value move, and where does it leak? These are not questions you answer by prompting an AI well. They come from years inside the system, watching it fail in specific ways, learning which compliance rules are bureaucratic theatre and which ones exist because someone, somewhere, got badly hurt.
What AI does is hand that accumulated knowledge an execution lever it never had before. Stone can now build what he already understands. The question was never whether he understood the problem. It was whether he could move fast enough to act on it. Now he can.
There is a second layer to this that matters especially for emerging markets.
When I host founders on The Grinders Table, one of the things I probe consistently is how much of their understanding is genuinely local versus borrowed from a playbook written somewhere else. It’s a question that cuts more than most people expect.
Stone’s point on the podcast was direct: Africa is not a monolith. Eversend operates across 12 markets, 12 regulatory regimes, 12 banking partner relationships, and 12 distinct definitions of what trust looks like in a financial product. He described putting an identical product into two neighbouring countries - one generated 40,000 users in a month with zero marketing spend, the other got nothing. He still doesn’t fully know why.
No AI model can bridge that gap. You close it by being present in the market, making specific mistakes in Senegal that you then carry over to Nigeria. You develop instincts that aren’t found in any dataset because they were never documented.
That knowledge is what determines whether the speed AI enables is an asset or a liability. A founder who builds fast without that depth will hit a compliance wall, or discover their unit economics break at scale, or find that the customer behaviour they assumed from a Silicon Valley case study simply doesn’t exist in their market. Speed amplifies everything, including the gaps.
The implication Stone left hanging, and the one worth sitting with, is that the AI productivity wave may widen the gap between experienced operators and first-time founders, at least in complex, regulated industries. Not because young founders can’t learn. They can, and they will. But in sectors where the product touches regulated infrastructure and real human behaviour, experience isn’t a liability that AI neutralises. It’s the asset that AI finally lets you deploy at pace.
The full conversation is on The Grinders Table podcast. We go deeper into what capital discipline actually looks like across 12 markets, why stablecoins were always infrastructure and never a product feature, and what Stone means when he says complexity, not brand, is the only real moat.
Click to listen on the web and tell me what you think in the comments