[sanderjd]

> What about AI/ML, Cloud, Workspace ? (+ of course ads and search).

"AI/ML" isn't a product, it's a tool, and one that is used effectively in a number of successful Google products (including the big moneymakers).

Cloud and Workspace are the non-ads bright spots, which is why they keep receiving plenty of investment, but they are both still very far from dominant in their markets, behind both Amazon and Microsoft. They are good businesses but not at all a full solution if the ads business declines substantially.

> Why didn't they launch something like ChatGPT?

Because billions of people would have used it immediately, and it's way too expensive for that. I suspect nobody yet has a way to serve these chats without the marginal revenue being negative. Google can't launch search features with rate limits, they can only launch things that are capable of serving billions of requests per day.

I'm not sure what argument you think I'm making that the Cloud thing is a counterargument to... My point is very much not "Google has always done everything right and made no mistakes". They absolutely missed the Cloud opportunity, and that's probably the only thing they've tried that could have been big enough if they'd entered the market early enough. (Maybe also if they had started making their own iphone-competitive hardware much earlier.)

But they thought going the social route, competing with Facebook, was going to be the answer, and the failure of that strategy was a major setback.

It could still be the case that this new era of AI is their saving grace. They still have a very real chance to be the first ones to figure out how to turn it into an economically sustainable product in some form. For all the hype, I sincerely doubt anyone has accomplished that yet. Definitely not OpenAI, and I'm guessing this Bing integration is a loss leader.

[Patrol8394]

>they can only launch things that are capable of serving billions of requests per day

They did not have to launch something like ChatGPT at full scale. I understand very well the cost associated with it. But, they should have been the one setting the stage and the narrative. Building expectation releasing cool demo, experimental features backed into Docs. Things that could show where "potentially" Google could have taken the technology in time ..

But, it seems like, that while they were working on the "perfect working money printing AI product", they completely left the stage to OpenAI and friends.

I think, that was definitely a strategic error.

Perhaps Google will come ahead, who knows. I hope this was a wakeup call for them.

[sanderjd]

> They did not have to launch something like ChatGPT at full scale.

I think this is really a misunderstanding of the constraints Google has to operate under. They have released plenty of things that "show where "potentially" Google could have taken the technology in time". Not least among them, the original transformer paper that this is all based on... But also, there are features in Docs and elsewhere that are backed by this kind of AI. Have you seen the "complete the sentence" thing in Gmail? How do you think that works?

What they didn't release is a generically useful product that everyone could use. Because if they had, everybody would have used it! Which, as you stipulate, would have been untenable.

But that doesn't mean it wasn't a strategic mistake! But it also wasn't a strategic success... The insightful and maddening thing about "the innovator's dilemma" is that incumbents can make the right decisions and still be disrupted. An incumbent just doesn't have the same options as a new entrant. This is an excellent demonstration of the effect!