[wegs2]

What's been a bit amusing is seeing a string of startups go through the same lesson about Google Cloud.

I'm increasingly coming to the conclusion that the mistake Google made was failing to recognize and focus on its core competencies. In ads and search, everything is scalable, statistical, and returns are decoupled from investment. You hire a few really smart people, and have them blow everyone else away.

It worked.

They should have looked for more businesses like that. Maps, gmail, etc. all did pretty well too. Even Youtube, despite playing whack-a-mole with demonetizing or removing channels with millions of subscribers, does well enough. There were plenty more businesses like that to diversify to.

Then Google brought that same mindset to things which required customer service (which doesn't scale), reliability (which isn't statistical in the same way as ads and search), and other domains, hired a ton of mediocre people to do it, and it seems like a bit of a dumpster fire right now.

[draklor40]

They simply cannot handle business that require long term support, lower margins, and communication overhead. As you said, given how effective Stats and Math have been for Search + Ads, they approach every problem with the same Hammer, which even Elon learnt the hard way, is not a great idea (Tesla's production issues with too much automation).

Reading Steve Yegge's posts (on Medium, IIRC), Android and Youtube are exceptions that the management leaves the teams to work indepedently instead of hosting the Google culture on top of them (he cites Android's API compatibility to this regard, which is in a way the opposite of how Google supports its products).

[loa_in_]

It makes sense from their perspective: they have datacentres upon datacentres of computing power that power their own endeavours, but are underutilized. So they monetize the remaining capacity in any way they can