>But relying on suppliers got Tesla into trouble with the Model X, when its engineers designed an extremely difficult-to-make “monopost” seat and an impossible-to-hit timeline.

I chuckled a bit when I read this. I can relate to this a lot.

For awhile I worked at a university in the mechanical engineering department machine shop. I like to think I know when I'm talking about when I say that everyone makes this mistake. They just usually stop making it by the time they've got their degree.

Designing parts that deviate greatly from industry standards for marginal gain and then finding out it's impossible to build what you want in the time you have is rookie engineer mistake 101. It's always tempting at the time but it's basically never worth it in hindsight

If you need a widget for your application and this particular widget isn't central to your application then use an off the shelf widget or tweak it as little as possible. Don't design a bespoke widget from the ground up. Designing your own widget from scratch consumes resources you should bet spending core aspects of the thing that you're building and has a high risk of being a time and effort black hole as you work out teething issues. Reinventing the wheel is fine, you should require that people have damn good reason or every team will want to reinvent every wheel and nothing will get done on time.

edit: I want my ass to be comfortable as much as the next guy but I'd rather have an off the shelf seat and have a Tesla now than have a bespoke seat and be on a waiting list.