There is a certain amount of arrogance to always being afraid of vendor lock-in. Most companies don’t survive, even the best ones might be just around for 20-25 years. The big worry should be on building a business that won’t immediately die.
And even with Oracle (probably the primo example of lock-in) it’s not like there aren’t firms who’s sole speciality is pumping data out of the Oracle DB and transforming it magically into T-SQL. It’s never the end of the world with vendor lock-in.
NOTE: now vendor lock-out does scare me like no other ironically
> It’s never the end of the world with vendor lock-in.
I think that's a matter of scale. Using Oracle for your counter example isn't very persuasive, as that's a huge vendor, so there's a market to extract you. Not so for many other vendors.
> Most companies don’t survive
...and ensuring they can control their costs and pivot away from solutions that prove too complex is part of being able to survive.
I mean, I agree, vendor lock-in doesn't have to be the end, but it also makes sense to extricate yourself when you can. When I worked for Virginia state govt they made a deal with Northrup Grumman...one that the legislative audit group came by later and said it was such a terrible deal that we should drop it...except we couldn't afford the exit costs, so we had to stick with a deal that was bad in terms of both money and quality.
That's a position you don't want to be in, and that result "It's bad but we can't afford to get out of it" is what the fear of vendor lock in is all about.
The only YCombinator company to go public - Dropbox - started life completely dependent on AWS, proved product market fit, got funding and slowly moved from AWS.
Other companies like Netflix decided that they didn’t want to manage infrastructure and proudly announced they closed down their final data center and moved to AWS.
Twitter desperately needs to move everything as quickly from self hosting as possible.
Not the person you're responding to, but I worry about (and have experienced) both with my tech stack, even as I've purposefully switched vendors multiple times with minimal headaches.
Locking yourself into a single vendor is easier to voluntarily work your way out of than your vendor shutting down or shutting you out unexpectedly. But the good news is if you plan for one you get the other for free.
And even with Oracle (probably the primo example of lock-in) it’s not like there aren’t firms who’s sole speciality is pumping data out of the Oracle DB and transforming it magically into T-SQL. It’s never the end of the world with vendor lock-in.
NOTE: now vendor lock-out does scare me like no other ironically