I agree that the shine has worn off startups, but don't agree with the reasoning. Startups were great in the 90s because VCs and investors were desperately hurling money at businesses that they didn't even remotely understand. Why work long hours in a stuffy traditional office at MSFT or IBM when you could make more at a carnival-like startup where you spend more time playing foosball than working?
Nowadays that equation is somewhat reversed. Most startups expect crushingly long hours, and typically pay considerably less than the tech giants and have vastly lower total compensation. There's still the slim chance of a huge payday, but the expected earnings are far lower. And while there's still plenty of "keep doubling down on legacy tech" at the giants, there's also way more genuine innovation than there used to be, whereas startups are the output of tech industry magnetic poetry sets as often as not: "instagram for urban farmers", etc.
In the end I also agree that we should push for facilitating more new companies, and the article underlines a lot of the hurdles we need to lower or remove. The big questions in my mind: whether that can be done in a way that doesn't just further empower the big established players, and whether anyone has the political willpower to hamstring successful, generally well-liked companies to increase competition.