Computer engineering is not substantially newer than aerospace engineering.

It's true, they just have different focuses. Aerospace (in general) has traded huge costs for safety and reliability. Computer engineering (in general) has traded correctness/reliability for low cost and rapid development.

It's a matter of tradeoffs. If your webapp is unusable 2 days a year people will just do something else in the meantime. If your flight control software fails for two days a year planes might come crashing down. The former is tolerable even if unpleasant from a customer perspective, thus they don't pay for flight control levels of reliability.

And it's not like this is unique to software. It's the difference between a disposable plastic fork and surgical instruments.

Yes, I agree. I was wildly exaggerating.

Life sustaining equipment is regulated and has a lot of resources spent on it to make it reliable. And toy software doesn't matter if it falls down.

I guess medium level services fall in between. Still causing inconvenience and monetary losses when they fail.

To be fair, a big portion of high profile computer engineering failures involve attackers actively searching for vulnerabilities, which bridges and ships don't need to deal with as much.