The Distribution Problem Is Real, but It's Solvable

Obie Fernandez, arguing that your company's multi-year AI strategy is a death sentence:

Take any vertical SaaS product and strip away the frontend. Remove the dashboards, the forms, the navigation, all of it. What's left? A database. Workflows. Business rules. That's it.

He's right about the reduction. Most vertical SaaS is a data model with workflow logic and a UI that took ten years to accrete. And he's right that AI agents don't need the UI — they can talk to the logic directly. That's a genuine insight about where the vulnerability is.

But "one competent engineer over a long weekend" is doing a spectacular amount of load-bearing work. You can prototype the happy path that fast, sure. The reason enterprise software costs what it costs isn't the happy path. It's the compliance layer, the audit trail, the 4,000 edge cases nobody wrote down because they live in the heads of the operations team, and the integrations with seventeen other systems the customer can't rip out. A demo that handles the main flow is a demo. A product that handles the rest is a business. The distance between those two is where the money actually went.

The tell is near the end. After spending 1,500 words arguing that the software is the easy part and incumbents are dead, Fernandez concedes in half a sentence: "The distribution problem is real, but it's solvable." Then moves right past it. He has to, because if you dwell on that clause, the whole argument inverts. If distribution is the hard part and the product is the easy part, then the companies that already have the distribution are the ones best positioned — they just need to move faster. That's a real problem for them. But it's a different problem from "your software is worthless," and it doesn't end with a solo engineer replacing your company over a weekend.

The $2 trillion market cap line is atmosphere, not evidence. SaaS stocks dropped because the market repriced growth expectations, not because agents shipped replacement products. Conflating investor sentiment with product obsolescence is how you make a post go viral while quietly overstating what's actually happening.

The useful advice here — build something yourself, develop the domain knowledge, don't wait for a three-year roadmap to save you — is good. It just doesn't need the "enterprise software is a weekend project" framing to be true.