What I hate the most is that they forked the Apollo codegen repo, and call it aws-appsync-codegen.

Yeah, I’ve been following the talk about that on the Apollo slack

Yeah, I emphasized that over there as well. AppSync powered by Apollo, seriously?!

It does look like the clients add some additional stuff on top of Apollo, though, with how they manage sync.

Yeah, they took apollo client + redux in about 50 LOC.

It'd be one thing if AWS wanted to piggyback open source and contribute back. But they just want to appropriate it.

They even updated the copyright in the forked repo to Amazon. That's not 'reusing open source', that's annihilation.

Those things said, I can't imagine the VC conversations within GraphCool. Eek

I apologize: not 50, but 62 sloc: https://github.com/awslabs/aws-mobile-appsync-sdk-js/blob/master/packages/aws-appsync/src/cache/offline-cache.js

I was burned by the RethinkDB debacle, so I'm overly sensitive to VCs 🙂

Well, Graphcool went full open source, so I'm not worried even the slightest in this case.

... would you put your money into GraphCool given that AWS is now in the same space?

Definitely.

I still would. Look at tools like Serverless. AWS has awesome services, but they still haven’t gotten their tooling accessible to the masses.

Their UI and UX have sucked since forever, they're too big to be flexible.

We're responding as developers. That's not the proper mode

They do not nearly offer the level of support Graphcool does.

Perhaps Google would be interested in a GraphCool acquisition to compete with AWS. But that's about the only exit left.

I doubt that Graphcool would be interested 😄

Their duty is to their shareholders, not their ego

I think the biggest mistake anyone can make at this point is move in a direction where companies like Google and Amazon are buying up the competition.

Completely agree

Think about things like managed clusters. Just like when VPS became big years ago, running one isn't the hard part. Keeping it running, operational costs, is the biggest issue.

But VCs don't want healthy businesses. They want 100x returns. So when GraphCool needs their next round, will there be money? Hopefully GraphCool can become / is a healthy business and they don't need to ask for more money. Then they sidestep the issue

Any developer can create an AppSync GraphQL endpoint, but maintenance, scaling, management, monitoring, etc. is not something Amazon is going to do for you.

RethinkDB -- despite being, arguably, the best NoSQL database -- couldn't make itself into a business and the VCs cut funding

The biggest overlap I think, is with the already free tier at Graphcool. Anything above that does not overlap with Amazon's offering.

Why does a full stack solution matter? Mongo seems to do fine as an "implementation detail"

The same applies there. Just clicking 'create new database' in the Amazon console doesn't make it a proper business solution.

Rethink had a similar offering. I understand that just because AWS moves into the space, it doesn't mean other products collapse. AWS DBs aren't shutting down Postgres. My point is, though, that their entry into the space makes potential suitors of GraphCool more hesitant to also move into the space which lowers GraphCool's expected value

Well, RethinkDb also still exists in its open source form

Horizon was more the issue. Horizon was their JS "realtime" client /server solution. And Rethink closed before Horizon was feature complete. I don't think we'd call GraphCool feature complete in its current form, though it offers escape hatches that Horizon didn't have

I think the focus should very much be on those escape hatches. I've seen many many frameworks and hosted platforms come and go over the past 20 years. And for me, the single most important decisive argument is: can I work around any limitations I might encounter. With the focus on GraphQL Gateway tooling, and the experience I have with that building my own framework around that concept, that strong point really shines. There are no limitations, just missing features, and most of the time they are easy to implement on top of Graphcool.