Stop Buying Domains
The name can come later.
In 1996, two Stanford PhD students built a search engine and called it BackRub, named after the "back links" it analyzed, and about as unappealing a brand as you could invent on purpose. A year later, when they finally sat down to rename it, the word that stuck wasn't even the one they meant to register: a grad student suggested "googolplex," they shortened it to "googol," and when they checked if the domain was free, they typo'd the search and found "google.com" available instead. That typo is the actual origin of the name Google. Nobody remembers or cares, because by the time the name mattered, the product already worked.
It's not an isolated story. Airbnb spent its first year as "AirBed & Breakfast," literally an air mattress on a living room floor, rented out to cover rent during a design conference when every hotel in San Francisco was booked. The name only got shortened once they were already inside Y Combinator, weeks before demo day. Slack didn't start as a company with a name at all; it started as an internal chat tool a small team built to survive while their actual product, a video game called Glitch, was dying underneath them. The tool outlived the game. The name came after.
I keep watching people chase domains and brand names like it's still 2015. A clean .com, a punchy name, a logo before there's even a product. I get the instinct, it feels like progress, it feels like you're building something real. But almost nobody who does this actually wins. The domain sits there. The brand never gets used. Meanwhile the game underneath it has already changed, and the three stories above are what it looks like when a founder is on the right side of that change, not the wrong one.
Why this used to make sense
For years, the hard part of starting something was making it exist. Getting a working product in front of a real person took months: design, engineering, infrastructure, all gated by how fast you could type code and how much money you could burn while doing it. In that world, locking down a good name early made sense: the thing itself was scarce, so you protected it like anything scarce.
That world is over. Right now, anyone can vibe-code a prototype and have a working website or app within a day. The barrier that used to separate "idea" from "product" has basically collapsed. And when the hard part gets cheap, the value doesn't vanish, it just moves somewhere else. It goes to whatever still can't be rushed. A domain name was never that thing. It just used to feel like it, because so few people could ship fast enough to expose the difference.
The realization
Building stops being evidence of anything. "I shipped X" used to signal commitment, skill, capital, now it just signals you spent an afternoon with an AI. A one-day prototype is a demo, not a business. So the scarcity that used to justify grabbing a domain and rushing to "claim" a space isn't there anymore. You're not defending territory. You're squatting on an empty lot next to a thousand identical lots.
That doesn't mean domains and brands are worthless. It means their job changes: from "hard to build, so hard to copy" to "hard to earn, so hard to copy."
What becomes the moat instead
Once anyone can build the same thing you can, in the same afternoon, with the same tools, the competition moves to things that only accumulate through someone else's time, not yours:
Reputation. People vouching for you when you're not in the room. It's slow to build and disproportionately easy to lose: one bad incident can undo years of it. It comes from consistency: shipping, following through, not overpromising, over and over, publicly.
Community. People who show up around your thing without being pushed to. The real test is whether they'd explain or defend you to a stranger unprompted. A competitor can clone your product overnight, but they can't clone the relationships your users have with each other.
Proprietary data. Whatever you've accumulated that a same-day competitor structurally cannot get: usage patterns, labeled outcomes, edge cases you've already seen. This compounds: more usage generates more data, which improves the product, which drives more usage. It's a flywheel a fast follower can't just copy-paste.
Regulatory relationships. Licenses, certifications, audits already passed. No amount of AI-assisted coding shortens the calendar time a regulator requires. In fields like identity, biometrics, finance, or health, "we're already certified" is a real cost a new entrant has to pay that you already paid.
Integrations. Being embedded in systems other people already depend on. Not the technical connector itself, that's trivial now, but becoming the default that other things get built on top of, so you accumulate a web of dependencies instead of a single link.
Existing customer trust. The compressed, personal version of everything above. A buyer who has already seen you handle a renewal, an outage, a hard question honestly. This is why a mediocre incumbent often beats a better new entrant: the real cost to the buyer isn't the subscription, it's the risk of being wrong about a vendor a second time.
The common thread
Every one of those six requires someone else's time to accumulate: a regulator's, a user's, a partner's, a customer's. That's exactly the one input that AI-accelerated building can't compress, no matter how good the tools get.
So the priority stack is quietly rearranging:
distribution and access to real users > trust and reputation > proprietary data or workflow lock-in > taste and positioning > raw build speed.
Building the thing moves toward the bottom of that list. It's still necessary, but it's table stakes now, not the differentiator.
So stop buying domains as a first move. Stop treating a brand name as the milestone. Spend that energy where it actually compounds: in front of real users, in the relationships and track record that a same-day competitor literally cannot fast-forward through, no matter how good their AI is.
BackRub became Google. A mattress on a floor became Airbnb. A dying video game's internal tool became Slack. The name can come later. It always could.