I know Dylan Field recently posted they made the Llama's 70B model do amazing things but the AI training field is an absolute snake pit.
But then again they have money to burn ig.
Fine-tuning an open-source LLM with reinforcement learning in order for it to be better at a specific task does not go against the bitter lesson at all.
That's fair because I am stretching it.
I think there's AI companies that understand and scale AI. Then there's the rest that are doing a modern version of writing domain specific knowledge using skills, and maybe a large collection of their own data.
Maybe these focused models outperform in their own niche similar to how hand-written rules outperformed in their own niches. But long-term, they get outpaced quickly by generalized approaches.
For context, this blog was released the same day Canva made their announcement for Brush, which looks to be further along in this area: https://www.canva.com/newsroom/news/canva-brush/
I'm glad Figma is pushing in this area. I do think that Figma has one of the best UI datasets out there in the world, and for the most part they've been sitting on it letting it get unused. Excited to see where they take this, but I do think they should be looking at image models in addition to LLM-based edit models (I'm biased as I left figma to start a image-mode-based UI design tool) as I find they're more capable in this area.
meanwhile, another commenter was right, the figma mcp is dogshit
Bring that layer up, move this box, etc
https://youtu.be/hHwjceFcF2Q?is=GWE74YfHzoEDaiWL
Ed: arguably there's some interesting bits in that scene beyond foreshadowing dejected designers slumped in a recliner, sipping cheap whiskey in front of a screen, adjusting pixel perfect layout across multiple devices using only a rasping monotone voice...
At the same time I mention that the "I am not a Robot" widget on some websites is the same as a Voight-Kampff-test.
Btw, the scene has an absolutely cringeworthy German translation where Deckard says at the end "Gib mir davon eine feste Kopie" with a literal translation of "hard copy" that nobody would ever use.
I haven't edited anything manually in Figma for a while now other than small tweaks.
This will be pushed to feature and UI/UX parity. The community has spun up an open source Raft/Paxos collaboration server, and are working to connect it with open source cloud and local compute.
Same thing for Canva.
The community is jumping onto this motion quickly. Just look at how well SolidCraft is progressing.
Companies are not going to be able to lock you into tools anymore.
Bitter lesson, etc, and training an entire VLM to prevent text overlap seems a bit much, especially since GPT 6.* /Claude 5.*'s vision are good enough that it rarely ever happens for me, and if they do make a mistake, it's generally a 5 minute fix?
Figma's current AI Agent is essentially useless, especially when compared to these other tools. Not sure what effect the work in this post will have, but even if it's great I feel that designers have already left Figma and are working elsewhere.
Yeah, no surprise there.
Its surprising to me that the market is still stuck on SaaSpocalypse not noticing that these companies have pivoted hard in an impressive and revenue generating way
Its to the point where their consumer facing products are a sideshow and distraction, but thats what people are familiar
They’re selling tokens just line all AI forms, with huge enterprise commitment payments up front
The cost of Figmas tokens was a frequent source of consternation at my prior engagement, and this was with a company that gave anybody $750/mo in Cursor credits if they asked nicely and filled out a form. Figma is also facing competition from Claude Design as well as other design tools that have a much cheaper price per seat, and allow you to BYOM. I wouldn't be surprised to see enterprises support a design tool switch if a lead showed how much money the company would save by switching away from Figma to something else.
Claude Design isn’t going to change a corporation’s workflow, an individual contributor isn’t going to cause a workflow change in their own department and especially not another department. Management operates their department in a silo. The outside consultancy driving the executive’s desired AI transformation will. Figma/Adobe/Canva are operating there, and not just coming in with their own software. They just steer things in the direction of also burning tokens.
They’re doing workflow automations not really so vendor locked.
Figma has consistently done better than most at nailing the UX. Maybe I'm enough of a generalist that it always feels great to me, and super specialists lament the lack of their super special needs being met (genuinely, a valid gripe for the specialists).
I've tinkered with Figma AI, and it's just good enough to keep me tinkering. It's not creating magic yet, and it's certainly not outpacing claude design to the point where I always pick figma.
I am enthusiastically optimistic that Figma will push the UX on agentic design to where I come to figma every day again.
Feels to me like a big issue that a smart competitor could take advantage of. I want to use my agent, I don't want to pay Figma to use their agent.
I'm also a bit salty about Figma ratcheting down hard on API calls, which creates customer service issues for any product that integrates with the Figma API. So maybe I am biased, but based on comments in here it also seems like the Figma MCP isn't very good either. Is Figma's long-term strategy to go full walled garden?
This is how Penpot does it, just use whatever you have, no bundling
Hit me with that feedback! Also thank you for the kind words.
-diffui creator
- If using a mouse without a side scroll wheel, how do I navigate the grid? I may have 20+ screens defined, and want to be able to grab an empty space in the grid and drag my way around. A mouse with side scroll solves this, because the main mouse wheel goes up and down. But without a side scroll mouse, I really cannot navigate.
- There's a ZOOM IN icon, but no ZOOM OUT icon in the left panel.
- I'd like an cleaner way to provide an existing design, or page, and get a new version. Yes, this is currently possible but the way you do this is (IMO) cludgy. I have to take a PNG, drag and drop it into the grid, then create a new empty prompt, then drag the output of the PNG into the input of the empty prompt. Then I reference the image with natural language "the image I've attached" ... but it'd be nice if at the very least, the prompt window gave me a way to reference the image, or some visual indicator that it sees it. If there's some better way to do this (create a new screen off of existing ones) please let me know.
- Performance. I have one screen that has like 40+ generated routes. The entire app. Its awesome but, it really starts to lag my M5 pro.
- Add some way to set the brand for an entire screen? Maybe I'm using this incorrectly but, every time I create a new prompt, I have to select the brand. I'm never cross pollinating brands so it seems like I should be able to set the brand ONCE on a grid, and any prompts started there use that brand?
- Maybe add some suggestions to help users generate different styles of designs? IDK maybe its just me but I've used this now across many different apps and the designs look great but, I'd love suggestions on how to improve them or mix it up. Sometimes I just find myself saying things like "make it modern and really awesome" but I know that's not the right term. There's industry standard terms, ideas, and directions for apps to be made in (for instance, retro pixelated) and it'd be cool to see suggestions somehow implemented. Not entirely sure how but, I could see it being very helpful?
- I see you recently added Teams. This was a big blocker at my workplace that I had to fight to get through. We need OKTA integration so that this can be added as an app which we login to through OKTA. At the moment I have a single account which is very non-standard for us, but your app is so good I pushed it through to get access on an individual basis.
Three ways. 1. Hold space and left click pan. 2. Press H to select the hand tool to navigate around the canvas. 3. Press "N" or "Shift+N" to focus the next/previous node (carry over hotkey from figma).
> There's a zoom in icon, but no zoom out icon in the left panel
Right click to zoom out, or shift+click.
> I'd like an cleaner way to provide an existing design, or page, and get a new version.
Totally hear ya on this one, I'm still working on a more fluid "redesign" flow. One note - you can @mention referenced elements. IE say "Redesign @webapp into dark mode". It'll bring up a selector when you press @ that lets you select imported nodes.
> Performance. I have one screen that has like 40+ generated routes.
Would you mind sharing the file with me? Top right share button, add j@diffui.ai. I'll take a look.
> Add some way to set the brand for an entire screen?
Good call. I'll make it so the brand selector maintains the previously selected brand, if there is one. Should be live in a couple hours.
> Maybe add some suggestions to help users generate different styles of designs?
I defo need to do this more, since you aren't the first user to mention that they don't necessarily have the language to convey a style they want. One thing that might help here is the shuffle appearance feature (open the brand dropdown, select randomize style). It'll inject some entropy which will let you window shop styles. I need to make a flow though that actually teaches the users the right terminology. Lemme think on this.
> - I see you recently added Teams. This was a big blocker at my workplace that I had to fight to get through. We need OKTA integration so that this can be added as an app which we login to through OKTA. At the moment I have a single account which is very non-standard for us, but your app is so good I pushed it through to get access on an individual basis.
Noted. Will work on this, but this will take a bit longer to get through. One note is it's very likely I'll have this as an enterprise-only option (eg the highest price tier). I am about to redo those pricing tiers however, so enterprise will be cheaper per image.
In my experience, Figma Agent and Figma's MCP leave something to be desired. I'm sure Figma Agent is useful for people who haven't used Claude Code or similar before, but it really requires most of your product's context to exist in Figma for it to do well.
As soon as your product's context exists out of Figma, then it starts to drift and hallucinate its own styles. Using the MCP with an AI that has access to your code provides better results, but even it will have trouble remembering what page I'm asking it to edit.
I know they're trying to address this by adding features such as guidelines and specs, along with linking your Figma files to GitHub repos, but my complaint with these features is a lot of the companies I have worked for do not use the very specific tech stack that Figma supports with these features. If it's anything like code connect, it may take years to get more broad support. Every company I have worked for is a C# Microsoft/Azure house.
This does not help with my broader frustration with Figma which is there is a lack of focus/vision. They're trying to compete on all fronts by developing all these different products/features and it doesn't really feel like they have a cohesive story for what the future or product design is. Granted, I think most of us in this industry don't either, but when am I supposed to use Figma Make vs. Figma Agent vs. Code layers?
They keep blurring the line between them when in all honesty I'm more interested in just using the AI agent I pay for to code what I need and bring it into a canvas for archiving/exploration when I need it.
Better results meaning it gets closer to what I had imagined in my mind or the design is subjectively better.
I loved the now legacy Claude design experience apart from the fact that the project lived on Anthropic’s servers.
My ideal product would allow the creative loop I had with Claude design, but more tightly integrated into my actual component library so there’s nothing to keep in sync.
I found some joy with Penpot but it didn’t allow me to use my real components.
I’m down to trying to hand roll this in each project now as a living style guide, but I’m missing something from the Claude Design system prompt that made it feel so quick and easy to explore and arrive at something I like and share prototypes
Then the whole thing is effectively useless because of server-side font licensing. Your expensive brand font cannot be used by the MCP so you then need an additional layer of client-side Figma plugins to replace the placeholder font with your real font.
What?
I know figma has a strong fanbase and companies are all rolling their own "post-trained" models for their built-in AI, and figma is a great product, but the fact frontier models can basically do in html what figma does in it's intermediary format sort of point to where the chips will land in terms of what you will need to use figma for in a few iterations.
I haven't used canva in 6+ months either for the same reason and I don't think either's userbases are growing honestly.
2.) You can use any agent with Figma
3.) You can use any agent with not-Figma
I used to really respect Figma's engineering team and have been blown away by the degree to which they've dropped the ball with AI.
All they need is to expose their existing API very well over MCP and give a "screenshot" feature agents can utilize and the agents are good to go.
I find it a huge distraction to ever draw anything in Figma now
Highly recommend it. I think we should make a .well_known url for them or something, honestly.
Oh I like that idea, I'm going to try it.
Wouldn't it be great if there was a standard template, or reference implementation that can be copied to modify for each site. Even better if they all had a shared vocabulary.
Easier to maintain, scale, and can use most of the components of the design system directly if you're doing web.
Why would you need an agent for figma if you can just tell an agent to create HTML+CSS directly?
(real question: what workflows does this support? is the sharable links to something in the cloud?)
Open formats are the clear winner in the age of LLMs.
Figma is an excellent product and works drastically better for me than the various tools I've used in the past for UI design.
I also don't see a future for the product though. Anthropic can now generate a design canvas of sorts as an artifact. Is it as good as Figma? Absolutely not. Will a designer use it? I'm not sure, but it has worked well enough for me in times where I'd previously have needed to find a designer to partner with for client work.
People can generate new variations of such tools by conversing with the AI, like a personalized Figma. In the near future they may generate a new Figma for every new project, or even every time they interact with the computer.
Any kind of AI editing can (and will) be rolled into any number of other models.
The irony is that Adobe tried to buy Figma and I'm sure that would've been a good exit for the Figma founders and maybe the employees too and that was called off over antitrust issues.
I just heard this from our designer today:
> After seven years, I've now cancelled my Figma subscriptions.