4 ms·

Right. I think it all comes down to the fact that with Google, maybe 2/3rds of any product is software running on servers they control. You can never see that s

Right. I think it all comes down to the fact that with Google, maybe 2/3rds of any product is software running on servers they control. You can never see that software, let alone modify it. Even the decision to keep running that software on Google’s hardware is totally out of your hands - Google could turn it off literally any day, for any reason. This is the downside of relying on someone else’s cloud software: ultimately if it’s not your code and not your computer, its fate is out of your hands. I think its healthy that our industry is finally coming to understand that.

throw10920 3y ago
> Google could turn it off literally any day, for any reason. Google has SLAs and service contracts that they are contractually bound to. They technically can turn off their services any day, but they won't, for the same reason that they won't liquidate all their assets into cash and drop it out of a helicopter - because it'd be extremely financially stupid. This comment makes no sense because it's discussing something that is extremely unlikely to happen - far, far less likely than a service failure caused by self-hosting.
- josephg 3y ago
  Those SLAs didn't seem to help people who bought these Jamboards. Or any of the laundry list of beloved products google has shut down over the years. They have a reputation for killing products for a reason - PMs on Stadia were promising that it definitely wouldn't be shut down just months before it was killed. I'll grant, Google has SLAs on some of their products (like Firebase). But trust arrives on foot and leaves on horseback. Wave, Inbox, Reader, etc. I still remember - with horror - that time google suddenly announced they were massively increasing GCP (Google Apps?) pricing without warning anyone beforehand. Is there a prediction market for the life expectancy of google products? Some things will probably be around more or less indefinitely: Docs, Gmail, etc. But the base rate for "this google product, service or API will still be available in 1 year" is probably somewhere between 1% and 10%. I'd consider any "all in" strategy on a lot of google's tech to be a serious business risk for that reason. Especially if losing that bet means you need to rearchitect your entire application.

throw10920 3y ago

> Google could turn it off literally any day, for any reason. Google has SLAs and service contracts that they are contractually bound to. They technically can turn off their services any day, but they won't, for the same reason that they won't liquidate all their assets into cash and drop it out of a helicopter - because it'd be extremely financially stupid. This comment makes no sense because it's discussing something that is extremely unlikely to happen - far, far less likely than a service failure caused by self-hosting.

josephg 3y ago
Those SLAs didn't seem to help people who bought these Jamboards. Or any of the laundry list of beloved products google has shut down over the years. They have a reputation for killing products for a reason - PMs on Stadia were promising that it definitely wouldn't be shut down just months before it was killed. I'll grant, Google has SLAs on some of their products (like Firebase). But trust arrives on foot and leaves on horseback. Wave, Inbox, Reader, etc. I still remember - with horror - that time google suddenly announced they were massively increasing GCP (Google Apps?) pricing without warning anyone beforehand. Is there a prediction market for the life expectancy of google products? Some things will probably be around more or less indefinitely: Docs, Gmail, etc. But the base rate for "this google product, service or API will still be available in 1 year" is probably somewhere between 1% and 10%. I'd consider any "all in" strategy on a lot of google's tech to be a serious business risk for that reason. Especially if losing that bet means you need to rearchitect your entire application.