Kai-Fu Lee wasn't. Evan Brown wasn't (edited to clarify: https://www.theregister.co.uk/2002/08/12/alcatel_owns_us_emp...). Don't be so quick to fall prey to the Just World Fallacy, assuming that bad precedents will never be applied against good people.
Remember, too, that the founders of Fairchild and of Intel had lots of knowledge in their heads that had been discovered by their left-behind coworkers at Shockley and Fairchild. They probably had documents in their homes — by accident if nothing else, if the history of recovered source code printouts from that epoch is anything to go by — but because there was no prosecution to find out about it, we'll probably never know.
Microsoft sued Google in civil court over a non-compete agreement, not theft. The settlement allowed Lee to keep his job, so other than five months of anxiety I'm not sure how he wasn't "fine". Levandowski is being charged with 33 counts in a criminal indictment. These cases are not even in the same ballpark.
> Evan Brown wasn't.
I don't know who that is, and there are apparently too many notable people named "Evan Brown" for me to figure it out without more context.
I got this: https://www.chicagotribune.com/news/ct-xpm-2002-08-05-020805...
Alcatel's PoV: https://everything2.com/title/The+Thoughts+of+Evan+Brown
trying to clearly define what exactly are ideas owned by the employer has to fail on principle. especially when you consider that even a new grad is being hired with the expectation that they are bringing ideas and techniques to bear that they developed someplace else.
Or just leaving the industry, or lifetime employment with a single company — hardly unthinkable; in fact, it was the rule in the US within living memory, and I've worked at companies where the majority of engineers were lifers. This would be a very favorable outcome for investors in your company: they get the fruits of your million-dollar ideas for the rest of your life, and they only have to pay you a somewhat-below-market salary, since you have no negotiating power to demand a higher one. It's not such a favorable outcome for engineers, or for hacking as a vocation or programming as a profession.
It's a favorable outcome for people who want stability. There is huge value in knowing that you're not going to get fired unless you do something actively malicious.
Some people aren't interested in climbing, have poor interviewing skills, and don't want to deal with the cognitive overhead of learning a different company's tech stack every two years. Plus, most companies tie vacation time to seniority, so being able to stay at a company for 10+ years means you'll get to have four weeks of vacation every year, while people who switch jobs every two years will always be stuck at two.
If I work in medical x-ray machines, say, when I quit, I don't go to another medical x-ray company. I can still work in embedded systems, though - it's a really big world out there.
Your argument works a bit better for Shockley and Fairchild, because semiconductors was not a big world then. And it perhaps works for Levandowski today, because self-driving cars skills aren't super portable to other industries. I could see some market for his skills in more general robotics, though.
Do you have proof of this? Claiming all those companies were founded by people who stole confidential documentation from their previous employers is a pretty bold claim to make without evidence.