Google really needs to start publishing LTS commitments on their new services - or put "beta" tags with disclosures on them - so at least you know how long they're guaranteeing support on a service you are getting committed to. The reputation they're accruing by killing all of these services is really affecting the consideration for new services. That is a very hard to measure and consider data point, but I can say first hand for my own services and those of companies I advise - this point comes up every time someone mentions electing for a google service.
Imagine considering building gemeni into your product knowing full well they will likely sunset it in a year or two or force you into the newest version (or replacement service entirely) vs OpenAI who are supporting and reducing cost on legacy models and service as they release new ones.
The other ugly side of this is the sunsetting of free tiers and plans on many products they do choose to keep. Gsuite free edition is a very good example - early adopters of gsuite for businesses were promised it would remain free, only to find that last year google walked back on that and sunset nearly all the free gsuite accounts, forcing early adopters (and supporters) to pay or get off. Google One is another example - they removed the ability for gsuite accounts to subscribe to Google One individually, and the aftermath of that was "take your data out or upgrade your entire organization to a premium storage tier per user". Zero intention to accommodate or support those users that had been happily paying for that service for years. Icing on the cake: they kept billing for it even after removing access to the service.
Yes. I'm actually worried about Golang too. Moved to it from Common Lisp but I am worried about the long term liability of it.
They also started doing telemetry and other things in the language because it is being treated as a product and then what happens if Google loses interest in it?
Would have gone all in on Rust if it just was aesthetically pleasing.
Golang is 14 years old, used by Google in production projects, and still in active development so it has a good track record. It's also open-source so worst case scenario it would get forked and community-maintained without Google backing. I wouldn't be concerned about Golang being deprecated.
Google does for paying customers. Google Cloud and Workspace (or whatever their "cloud office suite" is branded now) both have deprecation windows for launched features/products.
Imagine considering building gemeni into your product knowing full well they will likely sunset it in a year or two or force you into the newest version (or replacement service entirely) vs OpenAI who are supporting and reducing cost on legacy models and service as they release new ones.
The other ugly side of this is the sunsetting of free tiers and plans on many products they do choose to keep. Gsuite free edition is a very good example - early adopters of gsuite for businesses were promised it would remain free, only to find that last year google walked back on that and sunset nearly all the free gsuite accounts, forcing early adopters (and supporters) to pay or get off. Google One is another example - they removed the ability for gsuite accounts to subscribe to Google One individually, and the aftermath of that was "take your data out or upgrade your entire organization to a premium storage tier per user". Zero intention to accommodate or support those users that had been happily paying for that service for years. Icing on the cake: they kept billing for it even after removing access to the service.