While building has becomes way cheaper (and probably is going to become even cheaper in the future), is building something exceptional really that much cheaper now?
AI has certainly made it so much simpler to just pump "something" out (slop), but did it actually make building something that went through hundreds and thousands of iterations significantly cheaper?
I also like to think AI is really raising the bar for everybody. In the past, you could easily get away launching a product with a crappy landing page and a couple of bugs here and there, is that still the case? Don't people just expect a perfect landing page at this point (when's the last time anybody specifically talked/ thought about responsiveness?) paired with a flawless onboarding etc.?
Compiling to WASM is a huge reduction in size (as we don't need to ship a runtime etc. ourselfes anymore), but that requires some browser runtime. Dart specifically doesn't support standalone WASM (outside of the JS/browser context) yet.
I was able to get a WASM binary as small as ~234 KiB!
I haven't actually tried WASM yet. The demo on the landing page is using Dart transpiled to JS (embedded in xterm.js). But Dart does also support compiling to WASM directly, I'll give it a try tomorrow and let you know about the stats!
Binaries in Dart are definitely larger than 200kb, I just AOT compiled one of the examples and it was ~7.4MB (arm64 MacOS). Dart has a few different targets it can compile to, but it generally requires a couple of MBs, as it packages the whole runtime.
Yeah, my goal definitely is to have to as close to Flutter where it makes sense. You should mostly be able to copy it over 1:1, with a couple of caveats (GUI and a TUI app have a few differences here and there).
Yeah, I fully agree. UI has become a pretty obvious example for declerative programming, but it probably makes sense for so many other domains as well!
I'd say it depends on where you are coming from. With Vide, I'm approaching this problem from the code side. In my opinion, any application that is supposed to go into production and scale should be built on a solid code foundation.
My value proposition is to make developers more productive by skipping the boring stuff, while FlutterFlow is more of an "all-in-one" app platform.
While Cursor stops after writing great code, Vide goes the extra mile and has full runtime integration. Vide will go the extra mile & make sure the UI looks on point, works on all screen configurations and behaves correctly. It does this by being deeply integrated into Flutters tooling, it's able to take screenshot/ place widgets on a Figma-like canvas and even interact with everything in an isolated and reproducible environment.
I currently have a web version of the IDE live but I'm going to launch a full native desktop IDE very soon.
Original Author here, thanks for the feedback. I totally agree with most of the points made.
Right now, I wouldn't call it production-ready. It is too complex for non-coders and too trivial for actual developers.
I've been working on this for quite a while and I wanted to get something released so I wouldn't keep coding in my basement without feedback.
Right now, I'd say this tool is handy for Flutter beginners getting to know the layout system.
Addressing the problem regarding this being just a "Klickibunti" application.
I feel strongly about having this tool produce actual code a human developer would have written, therefore this could be considered the "groundwork" for future actually useful features.
Such features could include a full sketch integration that strongly connects to the actual widgets which work with different screen sizes and gesture behaviors.
One big feature which I've been planning for a while is a full 2-way code integration, meaning you'd be able to open up a random Flutter project in the widget-maker, tweak something and proceed with writing code.
Again, the points mentioned are very valid and I'm trying my best to build something which doesn't fall into the category of being a "nice idea that just doesn't work".
Author here, the widget maker is actually written in Flutter and it can definitely compile to Linux/ Mac/ Windows/ Android and anything that has a screen.
Flutter Desktop, unfortunately, is still in a very early phase and I had some troubles getting the same experience one all platforms. I'm working on fixing those issues and I will provide a Mac version pretty soon and eventually a Linux and Android build too.
AI has certainly made it so much simpler to just pump "something" out (slop), but did it actually make building something that went through hundreds and thousands of iterations significantly cheaper?
I also like to think AI is really raising the bar for everybody. In the past, you could easily get away launching a product with a crappy landing page and a couple of bugs here and there, is that still the case? Don't people just expect a perfect landing page at this point (when's the last time anybody specifically talked/ thought about responsiveness?) paired with a flawless onboarding etc.?