You built it in a weekend. So did four hundred other people.
That's not a knock – it's the new baseline. AI collapsed the cost of building software to something close to zero, and everyone in tech knows it. A quarter of YC's most recent batch shipped codebases that were almost entirely AI-generated. "Vibe coding" went from a joke to a word of the year in about six months. The barrier that used to separate a real product from an idea – the months of engineering – is mostly gone.
Which means the thing you were proud of is no longer scarce. The scarce thing is attention, and specifically, the right attention.
Every channel is crowded because building is cheap
The two facts are the same fact. The reason paid ads keep getting more expensive, the reason organic reach keeps falling, the reason the app stores and the launch platforms feel like a stampede – all of it is downstream of building getting cheap. More products, same amount of human attention. a16z's Andrew Chen said the quiet part plainly: every marketing channel sucks right now. Paid buys a spike with no retention behind it. Content compounds but takes quarters you may not have. And the launch-day traffic evaporates by Thursday.
So the founders who win the next few years won't be the ones who build best. Everyone builds well now. They'll be the ones who solve distribution – which is why a16z now frames momentum itself as the moat. Getting adopted, and staying adopted, is the durable edge when the product is copyable in a weekend.
The advice that doesn't fit builders
The standard prescription is "become the marketing." Put your face on it, post daily, build a following. For some people that works. For most builders, it's a tax they'll never pay – not out of ego, but because the honest objection is "I'm not a content person," and they're right. The market decided distribution costs a performance, and a lot of excellent products are quietly refusing and dying.
There's a version of distribution that doesn't require you to become anyone. The best early growth has always come from getting a credible person inside the right room to say "this is worth your time." Not a broadcast – a vouch. The subreddits, the Discords, the WhatsApp founder groups where your exact users already gather and already trust each other. That's where launches actually convert, and it's the one channel that doesn't get more expensive every quarter, because it isn't an auction.
Distribution as a function you can call
Here's the part that should interest a builder specifically. If building became programmable, distribution should be too. That's why Torchly ships an MCP server alongside the dashboard: an agent in Claude, Cursor, or any MCP client can search communities by audience, estimate reach against a target cost per action, launch a campaign, and read back per-community attribution – without you touching a growth dashboard at all. Distribution becomes a callable function, the way Stripe made payments one.
You still don't have to become an influencer. You write the brief – or your agent does – the admin of a community that fits endorses it in their voice, and you pay per confirmed result. The product being easy to build was never the problem. Being the one product anyone hears about is. That's the moat now, and it's the only one left worth digging.