29 comments

[ 2.7 ms ] story [ 62.8 ms ] thread
I don't want more Amplify platforms, I want them to fix the existing issues. Their github is filled with them.

There is so much pain with Amplify. They don't listen to anybody.

I guess one of the reason is fixing issues won’t get anyone promoted beyond L5
Well until then I'll whine under every Amplify post for bad PR since this might affect promotions.
It's called promotion-driven-development. It's a quite popular methodology on big tech these days ;)
This is the Amazon version of the Google Graveyard. They never retire the product, but they never really fix the bugs either.
This sounds much worse, though, as they are clearly still working on it, but only things that make good sounding announcements.
100%,

I hate their new releases, it makes me feel duped. My team has so much pain working with amplify. The very basics of their API functionality and lambda layers are extremely broken and it seems like it will be impossible to scale beyond a team of 5 people without insane conflicts.

> it seems like it will be impossible to scale beyond a team of 5 people without insane conflicts.

Tell me about it. We have to use different versions of the CLI to push different type of changes to the schema. It is insane.

100%. Their existing SDKs are so broken! I don't understand how they can ignore fundamental issues that absolutely make it impossible to use the SDK.
> There is so much pain with Amplify. They don't listen to anybody.

Indeed. Amplify sounded great on paper and I was really excited to use it in my latest project but it’s simply not living up to its promise.

I spent one month trying to get Sign in with Apple to work with Amplify + Cognito.

Frustrated and annoyed, I gave Firebase a chance. I had it working in 40 minutes.

Now the entire project is using Firebase.

Had the exact same experience. Excited about Amplify, but ran into bugs that hadn't been fixed for months blocking AmplifyJS social login from working as documented.

Switched to Firebase, and it all just worked.

Same here - and I like AWS and am on AWS for the most part. I never figured out Apple / iOS sign in. Their ECS container / S3 etc stuff works totally fine.
Yep. The Amplify JS still doesn't handle clock skew correctly for API calls after years of people complaining. I am still not sure they understand the problem.
Has AWS caught up to firebase yet , we went with Firebase as a backend for our Flutter web project and it's been a dream.

Web hosting , auth, Nosql data storage !

It is so far behind! It can't even handle social login reliably.
Would love to hear others experiences. I’ve used both firebase and amplify.

My concerns with firebase are: Google kills products at a whim, security rules are horrible.

Amplify seems more powerful so more likely to work long term, directly uses AWS “full strength”.

Concerns with Amplify: their npm libraries seem to be a colossal mess (they have old and new libraries that have to be partially used in tandem, and the docs use a sprinkling of each), plenty of odd bugs seems to crop up, the DataStore cache confuses me, idiomatic usage is not obvious - the fact that it’s composed of so many layers means it is not obvious which level of abstraction to interact with, and CLIs have flaws that seem blindingly obvious in design, which makes me worry about quality overall.

Should I stick with Amplify of return back to warm and fuzzy firebase (before Google dump it at random) and risk its limitations?

I would say either use Firebase or roll your own ( a open source self-hosted framework would fit this).

Firebase seems to be one of Google's stable projects, so I love it. I've used it for at least 4 years. I'm only using it as a hobbyist, so take all my experiences with a grain of salt.

Google hired the firebase team, integrated their product into the platform and completely rebuilt the DB with Firestore, and made Firebase into their mBaaS powerhouse that supports hundreds of thousands of apps. Its not some free product that's drawing a lot of resources, it's their entire mobile app developer strategy, and under many contracts to maintain the service.
I really like Firebase but for me it is way too risky for anything but personal projects. Vendor lock on the entire stack.

Amplify tried to compete with Firebase by glueing together a bunch of disparate AWS services. It has less overall vendor lock, but you are still functionally tied to an AWS service. Again, for me, too risky.

If you're comfortable with the vendor lock, Amplify is just not even close to Firebase. Amplify feels like some hacky side project in comparison to Firebase.

It depends, if your company has the money to hire a dedicated backend programmer then Firebase wouldn't be needed.

But if you're trying to solo develop or working on an extremely small team Firebase can do a ton of the heavy lifting. I can put a server up ,but I find backend programming to be very boring

I don't really understand why people say this. I'm using Firebase in my Flutter app but I've completely abstracted out the implementation into service classes. The business logic just calls my service auth, fetch data, etc... doesn't know anything about firebase. I could swap Firebase out for another provider no problem, no lock in.
No problem? Come on. The entire Firebase stack is closed source, and only available through Google. It's not like you can take your Firebase back-end and move it to Digital Ocean. If you want to go elsewhere you have to export everything, and put it somewhere else, with some other technology stack.

Your idea of vendor lock is radically different than mine.

Of course the "getting started" example is "here is how you wire up this button to send an analytics event".

Is anybody looking at all this trash data?

I try to avoid Amplify as much as possible.
Don't use Flutter lol.
Anyone that hitches their wagon to a Google embedded framework gets what they deserve.
Ubuntu rewrote it's installer GUI in Flutter.
I’ve actually found Flutter to be significantly more productive for developing an enterprise Android app when compared to React Native. One in production and one being prepped to launch.

Why do you have a negative opinion of Flutter?

Beating a dead horse never works. Let's try beating TWO dead horses!