I read this in an opinion piece that promotions and bonuses at Google are issued for launching things - products, face lifts, new features - not necessarily maintaining applications (maybe with the exception of search and advertising).

So taking that lens it's easy to see why so many new products are launched but then fall away and are killed, because the people involved can get their money and then move on to do other things. Also explains the constant user hostile redesigns of existing products.

To fix the reputation in this regard the internal reward structures would need to be changed so that building new things is the only way to get ahead.

"The people who should be most wary of this policy are Googlers who are looking for a team to join."

Only building new shiny things from scratch and never having to support them? Doesn't sound like a bad deal.

At the start of my career I was all about the tech, and I would probably have agreed with you. But now, with around 23 years of building web things behind me, I've realised that what keeps me interested is the fact that tens of millions of people have found what I've contributed to useful and interesting. Without that I would have given up and got bored.

Most people (certainly everyone I've ever worked with) need to have positive external input and feedback on the work they do to drive them forward, and if you're not getting that and the feedback is repeatedly "Meh, it'll just be killed in a few years" it would eventually start to impact your confidence and ability to do good work.