Fake it before you automate it

You fake it, you do a fake door test, you do a smoke test, Wizard of Oz tests. We used a lot of those in the tabbed inbox by the way, one of the first early versions was actually we showed the tabbed inbox working to people. But it wasn't really Gmail, it was just a facade of HTML.

Instead of trying to come up with the perfect plan, the perfect PRD, the perfect pitch to executives we're like, 'Let's just build a prototype and put it in front of our colleagues and get them using it.'

She was completely manual. She was the one behind the email address posing as the AI, but doing it herself.

Don't do it for your MVP. It makes zero sense. Do not waste time of data scientists that can train models with using powerful machines that are going take weeks to train.

We were obsessed with reducing friction, this was our constant battle. And so we hired a couple of college interns and we brought them in and we were like, people are going to push this magical one button in the Anchor app and they're going to say, I want to distribute my podcast, and your job is going to be to do all that same manual stuff manually, but to them it's going to feel magical and it happened automatically.

One of our principles from the very early days of Anchor was build things that don't scale, where we were like, we're a ridiculously tiny team. We're trying to do this really big thing. We had raised very little money, and so we were always just trying to figure out how can we hack into whatever growth we're trying to make happen.