Launching an AI Venture at 60: Why Decades of Experience Still Feel Like Beginner’s Luck

At 60, after starting 10 businesses over roughly four decades, I’ve enrolled in what feels like a fresh apprenticeship — building an AI company from the ground up. It’s the riskiest, most unnerving project I’ve taken on, even though I’ve sold firms to private investors and venture funds, endured bankruptcy once, advised public corporations, campaigned for state office (and lost), and authored two Wall Street Journal bestsellers while speaking at three TEDx events.

My career has settled into a familiar rhythm: spot an opening, pull together people, survive the brutal first year, then either grow, sell or shutter. I’ve launched restaurants, an insurance venture, media and marketing outfits, and consulting firms — industries I didn’t know at the outset. What I did bring was a grasp of unit economics, capital management and risk control. That foundation didn’t prepare me for how foreign modern AI tooling would feel. Suddenly, I’m the nontechnical founder in a room built by coders.

The difference isn’t just complexity; it’s the product itself. In restaurants you serve a plate. In insurance you underwrite a policy. In AI, the offering can be an amalgam: models that generate humanlike text, connectors to knowledge sources, and live-rendered avatars that must look and sound authentic. I don’t write code. The first quarter of this venture was spent understanding what’s technically feasible and recruiting someone who could translate that into a roadmap. I’m learning the vocabulary — enough to describe features and hand off the deep questions to engineering — but I’ll admit there’s a lot I still don’t follow, from how model-consumption is managed to how indexed knowledge stores interact with generation systems.

Still, the commercial rules haven’t changed. A valuable problem, a team that can execute, a route to paying customers and sensible unit economics remain decisive. Margins matter, the cost to acquire a buyer matters, and keeping customers is what sustains revenue. Industry observers I spoke with argue that roughly one-third of early AI projects chase clever models without a clear buyer in mind; those tend to stall. The winners mix technical chops with a tight commercial plan — tech is a means to an end, not the end itself.

There’s an unspoken preference in parts of Silicon Valley for founders who are young engineers based in the Valley, but that bias overlooks operational experience. Younger builders sometimes focus on the craft of the product and neglect distribution, pricing and cash management. That’s where seasoned operators can add real value: knowing channels, sales cycles and what customers will actually pay for. For me, the shorter climb was learning enough about AI to partner effectively with experts, rather than trying to become an engineer overnight.

My strategy now is straightforward: pair deep domain experience in running businesses with technical leaders who understand the stacks and tradeoffs. If one in three projects lacks buyer clarity, then combining practical commercial instincts with competent engineering tilts the odds. It’s humbling work, and I’m still scared most days, but I also think decades of running companies gives you a set of tools younger founders rarely have — if you’re willing to admit what you don’t know and let the engineers lead on the tech.