The feature conversation
I’m a link hoarder. Like most people, I save things I’m never going to look at again. But three links from the past few weeks kept pulling me back: a piece from Sequoia, one from Arielle Jackson in First Round Review, and one from Elan Miller. A VC firm, a positioning veteran, and a brand guy. When I actually sat down and read them, I realized I’d saved the same argument three times.
Sequoia’s piece argues that if you sell the tool, you’re in a race against the model, because every improvement in AI makes your product easier to replicate. The durable business sells the work itself, the outcome, the closed books. Arielle makes the same point one layer up: “AI-powered” stopped being a position the moment five credible companies in your category could say it, and most of them now ship the same gradients, the same screenshots, and the same copy designed to offend no one.
Elan’s piece is the one that stuck with me. His favorite question for prospective clients is what business they’re really in, and the answers come back functional every time: we make X, we help teams do Y. Then he points at Harley Davidson, a company that committed to one feeling so completely that grown adults pay a premium to wear its leather jackets. When Elan asks teams to name the feeling they want to evoke, most of them can’t, because they never picked one, and the brand ends up somewhat confident, somewhat warm, somewhat bold, and committed to none of it. He compares the result to meeting someone at a networking event and forgetting them before you’ve reached the coat check. His conclusion: “if you can’t name the feeling, you don’t have a memorable brand.”
Strip away the specifics and all three pieces arrive at the same place: when AI lets anyone build anything, the product stops being what sets you apart.
The rooms haven’t caught up
What I keep thinking about is how little the product discussions I’m part of sound like any of this. They’re about features: what to build next, what a competitor shipped last week, what’s missing before launch. I’m in those conversations several times a week across different teams and industries, and I catch myself steering them the same way, because twenty-nine years of habit doesn’t dissolve over three essays.
The habit makes sense, which is exactly why it’s so sticky. Features are concrete in a way a feeling never will be. You can put a feature on a roadmap, assign it to a sprint, and demo it on Friday, while nobody has ever closed a ticket called “make people feel less alone.” A roadmap full of features looks like progress, and often it is, because early on features genuinely separate you. The first product that does the thing wins customers from all the products that don’t.
What’s changed is the shelf life. The feature that separates you this quarter is the feature anyone can copy next quarter, and with AI in the mix, sometimes next week. The functional gap between you and a competitor used to be measured in engineering years, and now it’s measured in prompts. Competing on features still works, but only for a while, and the while keeps shrinking.
Where this bites hardest
Early-stage teams feel this most and notice it least. When the product is young everything is missing, so the feature conversation feels urgent and obvious, and closing the gap between what you have and what the demo promised is real work that has to happen. I’m not arguing anyone should stop building.
But early is also when belief is cheap. A five-person company can decide what it stands for over lunch, while a two-hundred-person company needs a committee, an agency, and a rebrand to do the same thing. The part that’s hardest to retrofit later is the one part that never makes it onto the roadmap at the start, because it doesn’t look like work.
Elan’s question turns out to be a useful test for this. Nike has had an answer for decades: they believe everyone with a body is an athlete, and the shoes are one of many ways they express it. The belief survived every product line they’ve ever shipped or killed. When I ask the question about products I’ve worked on, the honest answers were never features either. Summer Health wasn’t in the business of pediatric messaging; parents paid for the feeling of not being alone at 2am with a feverish kid, and the texting was how we delivered it. Every team I’ve seen build something that lasted could answer the question even if they’d never heard it asked, and while the features changed constantly, the answer didn’t.
So three corners of the industry, the money, the positioning people, and the brand people, have all landed on the conclusion designers have been arguing for decades and mostly losing to the roadmap. I don’t expect the conversations to change because of a workshop or an offsite. They change when someone in the room, midway through yet another feature discussion, asks what business this company is actually in, and the room realizes the question is harder than anything on the backlog.