My First Million · Episode Brief
I run a $180M+ company...here's how I'm using AI on a daily basis
Andrew Wilkinson runs a $180M+ holding company and his most important strategic insight right now is that AI is about to make most software businesses worthless — except the ones he owns.
Andrew Wilkinson built Tiny by buying profitable, boring internet businesses at reasonable multiples and leaving them alone. That model worked for a decade. Now he's spending a significant portion of his time thinking about which of those businesses survive the AI transition and which quietly go to zero — because the answer isn't obvious, and the timeline is shorter than most operators are pricing in.
The K-shaped future framing is his organizing thesis: AI is creating a bifurcation where a small number of businesses become dramatically more valuable while the majority get commoditized or replaced. The key variable isn't size or sector — it's whether your business owns a relationship that AI can't easily replicate. Businesses that own distribution, trust, or deep workflow integration are positioned on the right side of the curve. Pure software utilities with no network effects are on the wrong side.
His personal AI usage is more granular than the usual executive talking points. He's using agents that run continuously, not just tools he queries on demand. The distinction matters: a 24/7 agent that monitors your portfolio companies for anomalies is a different kind of leverage than ChatGPT running your search. Wilkinson frames this as the difference between AI as a productivity tool and AI as a business system — and he's firmly in the second camp.
The most contrarian point in the conversation: he's hedging against his own businesses by investing in AI companies. If AI kills the value of his software portfolio, he wants to own the thing doing the killing. It's a rational hedge, but it also signals a genuine uncertainty about which side of the K-curve his current assets land on.
Key Ideas
- →The K-shaped AI future: businesses either become dramatically more valuable or get commoditized — there's no stable middle ground
- →24/7 AI agents as a fundamentally different category from on-demand AI tools — one is leverage, the other is just a faster search
- →Software as a commodity: the compression of software margins means businesses need to own relationships or distribution, not just functionality
- →Hedging against your own portfolio by investing in the AI companies most likely to disrupt it — an explicit acknowledgment that disruption risk is real
- →Wilkinson's personal daily AI workflow: what an operator actually uses versus what gets covered in press releases
Worth Remembering
Wilkinson describing his hedge investment thesis — 'I'm investing in the thing that might kill my business' — as a completely matter-of-fact statement rather than a dramatic admission
The K-shaped future framework laid out clearly for the first time: not a gradual disruption but a bifurcation where most businesses end up on the wrong side
The contrast between AI as a tool you use occasionally versus AI as a system running your business continuously — and why most people are still thinking in the first mode