Thread by @skyshuford

No one wants to crash, but doing so at the edge of the design envelope, safely, is the right place to do it.

Every modern aviation startup should take this approach to RnD: 🧵

newatlas.com/aircraft/joby-…

1. Design to be remotely piloted. Why would we risk human life when we don't need to anymore?

2. Be hardware rich. Expanding the envelope requires pushing into uncertainty, and having backup parts / airframes / at-bats keeps programs moving forward quickly when uncertainties stack against you.

3. Test at the full, integrated system level (in flight ideally) as quickly as possible. Paper analyses and simulations only take you so far and break down quickly at interfaces between complex subsystems.

The real worry I have about the @jobyaviation crash is not even about the crash specifically, it's about aerospace startups being publicly traded while still in RnD.

The #newspace approach to develop (now more broadly maybe: #newflight) relies heavily on the ability to iterate and sometimes fail.

Look at @Astra, big stock hit after the recent launch failure (maybe fine depending on their warchest). Deeply disappointing, I'm sure, to the team and customer, but the next iteration will be better. We'll see what quarterly earnings and stockholder pressure does to @VirginOrbit

I can definitely see the appeal of the capital from an early aerospace SPAC, but I believe remaining private as long as possible (assuming good access to growth capital) allows an aerospace startup a greater degree of control over the culture of RnD.

Maintaining a risk-tolerant and ideally risk-forward, yet iterative approach to RnD will lead to higher performing and safer systems in production. Look at @SpaceX.

What other approaches do you think are critical to success as an aerospace startup?