Hacker News Reader

Comment by vl | Hacker News Reader

Urs is an engineer who imagines that he can run business and manage people, neither of which is actually true.

Urs is kind of quite assured of himself about a lot of technical decisions. Probably because he blueprinted TI, and contributed directly to some of the key decisions (warehouse-scale computer, as one example, although I am not sure if he or Luiz is more principal). He largely no longer made any good decision since the google plus era?

I read somewhere it states that Urs was behind the push of google plus.

Urs missed Cloud. Half-heartedly built GAE, which is technically superior, but mediocrely implemented.

Urs decided to defund internal development, in order to boost Cloud, after realizing google was years behind. That was a laughable technical decision.

Personally, Urs shows some form of personal distaste towards a few old but still working systems, and tried to make them disappear by cutting off their resources. borgcfg was one of them. Another evidence of Urs losing touch of the technical reality. borgcfg code is right behind C++, Java, the 3rd most human written code in Google. Make it disappear by cutting off its resources is essentially making almost every engineer suffer at Google... (I am behind the effort to revitalize borgcfg, and experienced the ridiculousness behind this view).

For all Urs' bad points, the point about internal vs external is incredibly true. XAML was a slow, painful mess in Windows 8 because all of the internal UI was largely developed in a parallel private internal framework. Once the internal shell UI started being built with the exact same frameworks and tools as public customers had to use, there was suddenly a bunch more pressure and resources dedicated to making them suck less, and the ecosystem as a whole was better for it.

Urs wasn't behind Google+. He may have supported it strongly, I don't remember, but G+ was a Larry&Sergey thing. They became obsessed with the idea that Facebook would beat them at everything because social was magic unicorn dust.

Urs missed Cloud

He didn't merely miss it but actively rejected it. I remember pushing for Borg to be turned into a product when I was there. The stock line (which I think came from Urs) was that Google's opportunity cost from renting out its hardware was so huge that it'd never make sense to do so, because ads was a superior business model to anything else and machine demand was infinite.

> Google's opportunity cost from renting out its hardware was so huge that it'd never make sense to do so, because ads was a superior business model to anything else and machine demand was infinite.

This seems reasonable though. Cloud is much lower margin than ads, so what’s the proposition to move into it?

Just because I’m good at something doesn’t mean I should do it. I make a mean scrambled egg, but don’t want to go into that business.

Good question but there must be one, because Google very much wants to win in cloud.

It doesn’t seem like Google wants to win. They’ve been distant #3/4 for 10 years.

If google wanted to win, I’d expect them to be more competitive.

I don’t really use GCP much but it seems like a hobby product line.

For example, they only have $4B revenue last quarter with almost $1B of losses (9% of Google’s revenue) [0], Compared to AWS’ $13.5B (12% of Amazon revenue) and Azure’s $15B (not sure what percent as this number includes lots of things that I don’t really consider cloud).

[0] https://www.parkmycloud.com/blog/aws-vs-azure-vs-google-clou...

It's hardly a hobby given the huge amount of features they've added to it. It's not winning, but that's not due to lack of trying. Probably more due to lack of enterprise nous and a bad reputation in certain areas.

Well, Urs actually gave answer at all-hands meeting, when everything became “cloud, cloud, cloud”. Clouds grow so fast, that Google would eventually stop being hardware purchaser #1 in the world, loose best discounts, which would make Search/Ads unprofitable. So there was no choose but to join cloud arms-race.

But of course it was too late. It was essentially IE moment for Google.

> Google's opportunity cost from renting out its hardware was so huge

ATM it is not wrong.

That's maybe the moment when an once visionary leader lost their grasp over the future. Instead, they start to focus on maintaining their own competitive advantage.

Exactly what innovator's dilemma prescribed...

His personality really comes out in this statement:

and will drop by the office occasionally, saying he plans to be in the Bay Area "on a regular basis" as travel restrictions ease.

Anyone who thinks they can jump this many time zones 'regularly' and be just as productive, either doesn't travel often, is deluding themselves, or is trying to paint a false picture of themselves to the public.

“Regular” is the key point, that doesnt include frequency. Im in australia but did 6 trips to NA/EU in the year before the pandemic. Ideally Id aim for ~3 to cover the major business process/planning/promo cycles. Thats not too frequent, but still regular.

Many high-level businesspeople travel from ~US to ~China and back weekly. There's no way I could do it as jetlag is brutal on me, but you and I don't mainline young boy blood.

> mainline young boy blood Thanks for that joke :D

I wonder if it still is a common practice among rich people? After being so ridiculed for this.

i think if you have to ask.... it's likely happening

You'd be surprised how quickly you can get over jetlag when traveling in first class. It makes a world of a difference.

(used to travel a lot for work, and was lucky enough to get upgraded a few times to business and once to 1st).

100.000% as productive isn’t a reasonable standard. Flying first class in lie-flat seats helps immensely to reduce the impact of the travel itself. Jet lag is still an impact, but most frequent travelers can find a mechanism that works ok for them. You might be at 50% for a day, 80% for another two, then at 90+% per trip. That’s usually workable.

Trying to do that sitting in a middle economy seat isn’t directly comparable.

New Zealand is a long flight but IIRC it’s about a 21 hour time change to California (give or take DST) which only feels like about 3 hours. (Not defending the subject of the article, just an interesting time quirk).

21 hours is equivalent to 3 hours though, things just get wonky when you cross the IDT.

Reply on news.ycombinator.com