1. Some of the things they launch they probably shouldn't be launching because they have next to no chance of ever being a meaningful business for Google. (eg. Google Helpouts)
2. Some of the things they launch are attempts to get into a particular area and Google's interest in that area lasts longer than their interest in whatever they initially launched. Chat is probably the best example of this; Google Talk, Hangouts, heck, I don't even know what it's called now! In such cases Google should be more disciplined about supporting whatever it is they've launched. Re-brand & iterate as much as they want, but never leave customers hanging.
I think if Google followed these two rules they would be sunsetting a lot less stuff. They might still need to retire the odd product/service, but at least they wouldn't be doing it so much that customers doubt every single launch.
Their current approach hasn't been working well for their customers, but it's actually going to begin affecting their customers less and affect Google more. Who in their right mind would put any medium/long term stock into Tables, for example? Customers no longer affected. Now Google can't launch a service that the market will take seriously.
Most of it is just the HN echo chamber. Folks commenting that GCP my be shut down because it doesn't make as much money as Azure, totally oblivious to the fact that Google has billions invested in physical buildings and hardware being hammered on and built out around the world as we speak, just shows that folks here are out of the loop on this type of thing.
From what I can tell, there isn't a single person in this thread claiming GCP may be shut down. And if there is, it certainly isn't a meaningful number of people.
What you're doing is claiming people hold an easily attackable position that they don't hold, and attacking that position. It's called strawmanning.
Not in this thread, but it has happened a couple of times in the past any time Google products / GCP have been brought up, and that I've replied to saying that it is nonsense.
I don't think this is true. Plenty of "normal", non-HN people bring this up too (I mean, not my mom, but like regular somewhat tech-savvy people). You can point to GCP all you want, but people are just going to keep pointing to https://killedbygoogle.com/.
Aside from an easily dismissed hardware-backed GCP case, Google _does_ have a long history of useful and interesting ideas that launch, gain some users, look really useful if some effort is put into it, and get dropped instead. Even if some solid use cases are building.
I'm going to look at this idea, see what it might be able to do, and look at whether there's another alternative that would be likelier to survive.
That is the exact thread I have a comment in from ~9 months ago. Anyone thinking that Google is going to shut down 24 physical datacenter regions spread all around the world if they're not #1 in the cloud market in 3 years is insane. The cloud market is multi billion dollar business and is only going to continue growing (even moreso now in the last ~9 months thanks to COVID). It isn't a free RSS aggregator that isn't maintained anymore and thus going to be sundowned.
They should release the financials of the app and let companies bid to acquire the app and its userbase. Smaller companies get a shot at growing the business, users stay happy(er than they would otherwise) and google gets a shot at acquiring the original idea back should it succeed under the hands of people who give a damn.
Unfortunately, such a program would likely be sunset after a year.
The way Google builds software makes that approach tough.
Even if you had the source code to the app (which Google isn't about to give out at any price), what you'll find is that it's an app architected to link against libraries nobody has ever seen and run atop a distributed computing fabric that's loosely related to Kubernetes but, really, nobody's ever seen, storing data to a backing store nobody's ever seen, and identifying users via an authentication model that nobody's ever seen.
Dangling references all over the place, leading to special sauce Google is even less likely to publish, because it's deeply tied into a physical hardware architecture that Google can't publish, because even if they did nobody's going to build it.
Some of Google's user-facing stuff runs on the architectures they make public, like GCP. But a lot of it runs on Google's proprietary fabric of service management and distributed storage, which is an alien planet relative to the world outside their walls. Publishing an app out of that part of the ecosystem would be like Google handing a company a koala with no eucalyptus trees. It'd be dead in days.
More importantly, this is all released on Google's internal infrastructure. Any such work would require moving to the external infrastructure, and the cost of it likely wouldn't make sense.
The real issue is this: Why am I paying 10 dollars/user/month for this when G suite Business is 12? I could see using this for the organization, but not at that cost.
