The dynamics are exactly right, but I would say engineers can't "let" "politically bad" projects fail, because fixing that is, in a very real sense, above their paygrade. That responsibility lies with the executives.

The engineers' role should mostly be as technical advisors, i.e. calling out bad projects for technical reasons (UX, architecture, etc.) But even the seniormost engineers do not have the corporate standing, let alone political cachet, to call out or fix political issues (empire building, infighting between orgs, etc.) They can and should point out these conflicts to leadership (very diplomatically, of course) but should bear no responsibility for the outcomes.

However, as an engineer you should ABSOLUTELY be aware of these dynamics because they will impact your career. Like when the project is canceled with no impact delivered.

The example given of the latent turf war between the product and platform teams might have been avoided via a very clear mandate from senior leadership about who owns what exactly. This would probably have involved some horse-trading about what the org giving up its turf gets in return. (BTW if you've ever wondered "Why so many re-orgs" this is why.) That this didn't happen is a failure on the execs' part.

As an aside, I know this happens in every large company, but somehow it appears to be a lot more common at Google? Or at least Googlers are more open about it. E.g. I observed something similar on that recent post about lessions from Google: https://news.ycombinator.com/item?id=46488819