Yes, he's right, but let's put this in context. There are probably a lot of talented engineers who are doing startups now who shouldn't be. The problem is how do we know who is going to be a good entrepreneur and who is not? We don't. The only way we know is by doing it, which is what a lot of talent is doing. I think many startup founders are a bit grounded in reality and know that they face real possibilities of failure. Just as long as they keep this in mind and don't become obsessed with caviar dreams, then I see no harm done. Being an entrepreneur is one way for an engineer to develop non-technical skillsets and learn something about himself in the process.

Let's not forget that this entrepreneurial activity is forcing the big guys (Google, FB, etc) to be more innovative than they would otherwise be - either to attract talent or to compete with startups.