Comment by bredren
6 days ago
Congratulations to everyone who made this company go. That said, this company's growth seems threatened by AI adoption rather than boosted by it.
6 days ago
Congratulations to everyone who made this company go. That said, this company's growth seems threatened by AI adoption rather than boosted by it.
One thing I thought would have been incredibly useful for me is to go from HTML -> Figma
There is a ton of focus on going from Figma to something you could presumably dump into react or html. But I found nothing in the reverse.
Realistically, for a lot of applications, there are more things in production than in Figma. It's just not practical to spec everything out when you're moving fast. But when you do want someone to look at it or tinker, it's a huge lift to migrate your current production to Figma. I wish they would use some AI for that. Just take a webpage, and build the Figma design docs. Doesn't even have to be perfect, just good enough to get help from designers
The feature we need most is Figma > SwiftUI, etc. The ability to take the Figma UIs and export to web and mobile would be such a time saver.
jsx-lite had it in experimental mode for a while: https://github.com/lamppkk/jsx-lite
Anyone with good SwiftUIfu care to take over? Then hopefully jsx-lite gets submitted to the ES specs like E4X. Then we can have UI write once and runs in all places.
There is a Storybook-to-Figma [0] plugin, the same company built a whole slew of X-to-figma plugins.
[0] https://www.figma.com/community/plugin/1075741140914731351/s...
Ah yes Dreamweaver. Look how well that turned out.
Turs out people don't want "quick dump as HTML" but rather "maintainable, understandable, performant HTML". I don't see how that has changed with AI.
Dreamweaver was a widely used tool back in the day and definitely started a trend in web design tools.
1 reply →
If by “people” you mean “developers”, then yes you’re right. But I don’t think anyone else ever really cared.
The problem with Dreamweaver is that you still needed a developer to upload and run the site. And back then, you couldn’t run single page applications (web stacks hadn’t evolved that far yet), so still needed developers to write the backend.
Thus there wasn’t a huge amount to gain in using Dreamweaver for the professional world.
AI has changed that in that it doesn’t have the same limitations as Dreamweaver. However, like yourself, I don’t think we’ll see AI replace developers. Or at least the current crop of LLMs still have a long way to go before they can be used without developer oversight.
Edit: also worth noting that Flash was everywhere back then too. So many web designers opted for Flash instead.
3 replies →
Take your current project, then run `figma-cli-ai --fix` and have a modern design.
I guess I'm missing the joke as this doesn't seem to be a thing?
Check out html.to.design
I want to be able to refactor designs, and not entirely from the command line. I do not see why Figma should not be able to do this. Coding comes later.
They’re adopting AI so that designers within figma can create. I don’t know whether this is good or bad (I don’t design) but if the tool everyone uses to mock things up gains coding abilities, we’re cooked.
You’ll be able to go from figma to production in weeks.
And we'll have 10x of the Tea app incidents.
AI tools are still just that tools - they're not abstraction layers from "intent" to "production product".
There's also code snippets so that developers can provide the HTML or React or whatever code to implement a component or set thereof. And that stuff feeds into an MCP server, so that in theory an AI / code assistant can implement a design in whichever framework you build in, but within some limits if you provide the right code.
It won't just be figma. The market is already filling up with lots of players in this space.
There may be no moats. Just distribution winners.
Never underestimate the power of market share. Adobe has had a bunch of competitors but still dominates design.
21 replies →
> It won't just be figma. The market is already filling up with lots of players in this space.
MongoDB wants a word.. another "this will never amount to anything" HN special.
It'll depend on other vendors to make the models who will eat up all the margins
"but if the tool everyone uses to mock things up gains coding abilities"
adobe tried this 10 years ago (Adobe dream weaver) and failed
I literally can drag and drop design photos from drawing board to claude or open ai chat and they recreate it themselves instead of needing figma
not sure where you know that "we are cooked" its for them not us
Figma is what all the designers are using. If they turn that into a Claude for design and add the ability to complete the wireframes with AI generated code. There’s no need for engineers until after acquisition. Just maintenance.
The difference here is - Adobe tried using a proprietary stack. Figma can spit React. Everyone knows React - even if you use something else. It’s perfect timing with AI coding agents and every, single, UX team using it.
4 replies →