The question the video answers

  • Instagram scaled to 2 billion users on Postgres. Reddit, Notion, Heroku, Strava and Discord all run Postgres at some capacity (0:00).
  • These teams did know the alternatives and evaluated them seriously — staying on Postgres was a deliberate choice, not ignorance (0:00).
  • Two questions drive the video (0:00):
    1. Why do they all still use Postgres?
    2. What does it actually take to make Postgres work at planet scale?
  • Goal for the viewer: a clear picture of the actual bottlenecks you hit when scaling a relational database (0:33).

The thesis stated up front

Boring technology with good engineering beats new technology every single time (1:33).

The video plants this lesson early and returns to it explicitly at 15:21 — see 08-boring-technology-as-a-superpower.md.

The framing that matters

  • When one machine is no longer enough, the wrong answer is to panic and switch databases; the right answer is to find out exactly why the database is struggling, then fix that (1:59).
  • The closing restatement: the database is not the bottleneck — the architecture around it is (21:44).