Some bad takes from the internet
It's all about risks
The maker of Keras argues that nothing has changed in SaaS

Here's what's wrong in the argument - it's a contorted form of survivor bias. It's absolutely true that for any succesful SaaS concept, the original investment has very little to do with either the lifetime investment, or the value of the business.
However: When experiments drop 100x in price you can make 100x more experiments making it 100x more likely that one of these will turn out to have some merit. Sadly - and we're seeing this already - attention is a zero sum game - so 100x more software launches means 100x less average attention available for your idea. So Chollet could be so wrong that in the end he's right - just not for the reasons he thinks he is.
It's still all about risks
A Meta AI researcher tasked with working on AI risk was experimenting with Open Claw and it ran amok and deleted all her email. I guess we all know Open Claw is risky so that's hardly a faux pas however the alt-right memelord Elon had opinions

You would think a man who routinely blows up rockets would understand the purpose of risk a little better. "Somehow who blows up more rockets than anyone else since 1970 is definitely gonna solve space travel. 100%"