4 ms·

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 deci

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).

Arainach 5y ago
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.
native_samples 5y ago
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.
- prepend 5y ago
  > 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.
  - native_samples 5y ago
    Good question but there must be one, because Google very much wants to win in cloud.
    - prepend 5y ago
      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-cloud-market-share/ https://www.parkmycloud.com/blog/aws-vs-azure-vs-google-clou...
      - native_samples 5y ago
        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.
  - vl 5y ago
    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.

Arainach 5y ago

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.

native_samples 5y ago

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.

prepend 5y ago
> 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.
- native_samples 5y ago
  Good question but there must be one, because Google very much wants to win in cloud.
  - prepend 5y ago
    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-cloud-market-share/ https://www.parkmycloud.com/blog/aws-vs-azure-vs-google-clou...
    - native_samples 5y ago
      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.
- vl 5y ago
  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.