Back in undergrad around ~2016, a professor who shared the same sentiment once said something funny along the lines of 'No/low-code is temporary. But one thing is certain - HTML will never die!' during a class discussion around the concept of 'clicks, not code' and if it presented a real risk to future job security.

Services like this are cool and all - if one of them gets someone to play around with it and find they like it, that's a solid win. But relying on a 3rd parties usually leads to some sort of heartbreak down the line if and when it disappears and your projects along with it.

Surely AWS could commit a developer to spend a week or two writing a basic wrapper, even if it's a MVP-ish set of node or python scripts that lets folks download their apps. Or for a more AWS-specific solution, export them to EC2 instances.

Where do you arrive at a week or two as an estimate for taking a cloud platform and making it self-hostable, even as an MVP? I know nothing about Honeycode, but I would bet that it's piecing together a bunch of AWS services behind the scenes rather than being something you can just stick on an EC2 instance.

Wouldn't even need to be the same application, just something that lets the existing data form a functional application. Could even plug into Amplify, Step Functions, etc

Isn't what you're proposing an entirely new product? How could an engineer put that together in two weeks?

No, just a wrapper really. As you said, most of the heavy lifting is done by the AWS services. It needn't be high quality at all; just something that someone could download to not lose all of their work.

Of course I meant "two weeks" as something more or less metaphorical - could be 16 weeks, and a 2-pizza team, given the amount of time until Honeycode shuts down. I just think Amazon has the resources to do that.

If I was a competitor like Retool I'd be building a migration tool right now.

I suspect literally nobody used this tool.

Totally untrue. i know this for a fact.

I'm kind of shocked that Glitch is still around after all these years.

It's pretty much "dead" though, right?

Personally, I think commercial open source (COSS) technologies are pretty much the only solution to this problem. This way you aren't beholden to a company stopping their service and leaving you on your own. COSS ensures the project is maintained and if it isn't you can always self-host and migrate on your timeline instead of a vendor mandated timeline.