Y CombinatorWhat Happens When Everyone Can Build Software
At a glance
WHAT IT’S REALLY ABOUT
Supabase bets agents will make personal databases mainstream infrastructure
- Paul Copplestone argues AI is turning far more people into “builders,” including non-engineers, which is expanding the market for easy-to-spin-up databases.
- Supabase’s product strategy has shifted across three eras—dashboard, API, and now code-first—to match how agents and modern workflows actually build software.
- The speakers describe a trend toward personal apps (finance, health, CRM, file intake) where individuals want to own their data in a personal database and attach multiple agents on top.
- Copplestone describes scaling challenges from rapidly rising database creation and explains a two-ended product approach: scale down for lightweight workloads and scale up for very large ones.
- The conversation emphasizes “agent-first” design: being a default choice in models, improving error messages for self-remediation, and building eval/testing loops to ensure agent-written code is reliable.
IDEAS WORTH REMEMBERING
5 ideasSupabase expects databases to become personal infrastructure, not just company infrastructure.
He describes a shift from a world where databases mainly served companies to one where individuals—enabled by AI and agent-driven building—create personal apps that need real storage, retrieval, and integrations.
Supabase’s roadmap is being reshaped around how agents build, not how humans click.
He explains Supabase’s product evolution as moving from a UI-centric dashboard, to an API-first platform for provisioning many databases, to a code-first and CLI-first approach because agents don’t use dashboards.
Open source is positioned as a competitive advantage for both agents and enterprise trust.
He says open source likely mattered significantly because agents can inspect source when things break, and enterprises may want to inspect source for security; he also emphasizes Supabase supporting broader Postgres ecosystem tools.
Non-engineers and domain experts are becoming meaningful builders—and they’re accelerating adoption.
He notes that growth is being driven not only by software engineers building more quickly, but by “builders” without traditional engineering backgrounds who can now create apps using AI.
Agent-oriented UX means designing error loops so agents can self-remedy instead of switching tools.
He frames “winning agents” as two steps: being chosen as the default by models, and preventing churn when agents hit errors; he gives the example of embedding remediation instructions and doc links directly in error messages.
WORDS WORTH SAVING
5 quotesI certainly think that it'll be the case that everyone in the world with access to AI will become a builder of some sort. Like you can build anything, you can customize it, you wanna make tweaks to some app, some UI, then you don't have to be beholden to, to someone.
— Paul Copplestone
All of them are kind of going up and to the right, which is a bit insane just because the models are getting so good at building, and we're starting to see people who traditionally wouldn't have a software engineering background able to come and, and build things.
— Paul Copplestone
I would never have thought about wanting a personal database like a year ago, 'cause it wouldn't have made any sense, but now I'd like, that's actually like the most valuable thing.
— Harj Taggar
No one visits dashboards anymore.
— Paul Copplestone
I'm genuinely excited about what the future could be for a lot of people.
— Paul Copplestone
High quality AI-generated summary created from speaker-labeled transcript.
