Five ideas shaping how
we build software next
On micro-teams, speed, portfolios, taste, and validation.
Big Teams Are the New Technical Debt
Until recently, building and shipping serious software required large teams — designers, researchers, PMs, engineers, QA, marketers. That era is ending. A micro-team of 2-5 people using AI agents can now outpace a team of 40. Not just on speed — on decision quality. When everyone on the team has full context, you eliminate the meetings, syncs, and alignment overhead that large teams drown in. If you have more than 4-5 decision-makers, you're already behind. The minimum viable team has collapsed, and the advantage goes to those who stay small on purpose.
Building in Stealth Is Dead
The old playbook — work in stealth for months, build a waitlist, launch with fanfare — made sense when building was slow and expensive. If it took 6 months to ship, you didn't want a competitor to out-execute you mid-build. But now even a fairly complex product can go from ideation to validation to build to public launch in low single-digit weeks. Not days (production-grade means auth edge cases, payment flows, error handling, real user testing — that's still calendar time) — but weeks, not quarters. When anyone can ship that fast, secrecy was never your moat. Speed is.
Build Like a VC — Launch 10, Keep the 3 That Hit
When launching a product costs 2 weeks instead of 6 months, you can afford to let most of them fail. This is venture capital logic applied at the individual or micro-team level. Micro-teams will build swarms of products for niche audiences — a few hundred or a few thousand users each — and make serious money from the portfolio — a single niche tool can generate $50K-$250K a year, and you only need a few to win. The ones that hit, you double down on. The ones that don't, you kill fast and move on. The economics of experimentation have fundamentally changed.
Everyone Can Build Now. Few Know What to Build.
AI has democratized execution. The gap is no longer between those who can build and those who can't — it's between those who know what to build and those who don't. Taste, judgment, and the ability to identify genuinely underserved niches are now the scarce resources. The founders who win the micro-product era won't be the best engineers. They'll be the ones with the strongest opinions about what matters, who it's for, and how it should feel. When building is free, the costly mistake isn't picking the wrong technology — it's picking the wrong problem.
The Costliest Product Is the One You Shouldn't Have Built
The costly mistake isn't building the wrong thing — it's building the wrong thing slowly. If 20 minutes of AI-powered research can tell you a market is too small, too crowded, or already well-served, you've saved weeks of wasted effort. The winning strategy isn't "build everything fast." It's validate ruthlessly, kill bad ideas early, and only commit build time to ideas that survive scrutiny. Speed without direction is just expensive wandering.
We built a tool around these ideas.
Microfactory is an AI-powered idea triage system for micro-teams. Validate fast, kill bad ideas early, ship the winners.
See how it works