Really resonates with how I’ve been thinking about this. AI doesn’t kill moats, it just makes most of them much thinner and more short-lived. We end up in HK‑telecom territory: hyper‑competitive, brutal for operators, amazing for consumers.
Foundation/model moats get commoditized fast, and the durable edge shifts to infra, regulation, data, and embedded workflows instead. Long term, it feels like the “AI surplus” mostly accrues to users, while vendors are stuck fighting in a permanently volatile market.
Doesn't the studio analogy end more hopefully than it sounds? The studios that survived stopped living on single releases, they bought distribution and franchises, and the hits became marketing for the machine.
Reminds of something you've said in a previous post: "build something valuable and charge enough for it to make a profit on each sale." Business 101. Since I started building my first SaaS from scratch last year (laand.me) I've held on to this. Anything other than that forces one into spreadsheet projections and the illusion of adding more with the hope of reaching a point of profitability that may never become reality (for example, what may be case of for most of these frontier models. But again, it seems they may exist to serve purposes other than making profit)
Really resonates with how I’ve been thinking about this. AI doesn’t kill moats, it just makes most of them much thinner and more short-lived. We end up in HK‑telecom territory: hyper‑competitive, brutal for operators, amazing for consumers.
Foundation/model moats get commoditized fast, and the durable edge shifts to infra, regulation, data, and embedded workflows instead. Long term, it feels like the “AI surplus” mostly accrues to users, while vendors are stuck fighting in a permanently volatile market.
Doesn't the studio analogy end more hopefully than it sounds? The studios that survived stopped living on single releases, they bought distribution and franchises, and the hits became marketing for the machine.
Reminds of something you've said in a previous post: "build something valuable and charge enough for it to make a profit on each sale." Business 101. Since I started building my first SaaS from scratch last year (laand.me) I've held on to this. Anything other than that forces one into spreadsheet projections and the illusion of adding more with the hope of reaching a point of profitability that may never become reality (for example, what may be case of for most of these frontier models. But again, it seems they may exist to serve purposes other than making profit)