Posted by martinald 3 days ago
While it has a strong potential to let people iterate on using a design without the nuts and bolts of going back and forth with a designer, CD operates at the "leaf-node" level, where the output is generated.
However, a lot of design has a deeper life-cycle than that. There's the collaboration, pitching, review, iteration, asset management, etc.
In fact, the first step for using CD is "onboarding", where it sucks up a design system from your existing assets/resources. It presumes you already have a design.
As it stands now, CD is one way... existing design -> task specific resources. This could be very useful, but only touches on a part of what a complete design tool does. But for iteration it's not so great. E.g. task specific concerns don't have a way to feed back to the originating design. Changes to the originating design don't have a direct path to feed back to the task specific output (e.g., when a logo or branding focus changes, or maybe just spacing guidelines are updated, the ad hoc processes around CD will have to be repeated if the changes are to actually land.)
I'd think AI design integrated with Figma is in a much better position to address these more complicated scenarios.
I doubt Claude Design even cares about these deeper scenarios, BTW -- it's intended as a leaf-node tool. Just pointing out it's not about to replace Figma or other more comprehensive design tools.
If corners can be cut, they will. All those steps would be flatened to something like CD and a couple of side tools.
Companies did "collaboration, pitching, review, iteration" because they had a designer in the loop anyway for the actual final work. Now that they don't have to, how many will just skip those steps, and if it means the end product gets less intented and "defined", they'd be fine with that?
The requirements are so unstable—the product team has few strong beliefs—that they change the next day. And then again every few days after. Hopefully, the changes are small enough that design isn't full resetting each time, but it's not rare to have big changes. The entire project gets swapped not infrequently. What eventually slows the changes is the engineering deadline and the fact that the developers need to start. But the slow drip of product requirements means whatever time budgeting went to design shrinks. And whatever time went to engineering is eaten into such that now the design needs to be something that can be built in half the original amount of dev time. Each day the designer takes at this point eats into that window and so it's dictated by what can get built.
I don't think that has to strictly be viewed like an entirely bad outcome, but for what it is and how it's accomplished, you could just cut the design part out. Besides, you're going to iterate later, right? Right?
Absolutely. A chance for middle management and C-levels to bikeshed inconsequential bullshit and feel like they're doing something.
>That end result might be bad (perhaps it would help to be more respectful of the designer's efforts), but it's the same either way.
Can't be that worse than the slow to load, 50MB for a page, flat design full of wasted space shit redesigned every year or so to follow the new stupid trends that we're getting for the past 15 years
I do think Figma will have a problem that people with think Claude Design competes with Figma directly.
I expect people in leadership positions aren't comparing "Claude Design vs. Figma", but area comparing "Me and my product manager using Claude Design vs. A designer using Figma."
Sometimes people want some graphics without it being like they're planning a wedding
The reaction that designers I know have given Claude Design couldn’t be different than how Claude Code was received by software devs. It’s simply useless for designers, their workflow is very different from software devs. You can’t “oh let Claude Design come up with a quick logo for this” in the same way that Claude Code was able to quickly solve small annoyances for devs.
People that think that Claude Design is going to replace Figma don’t really understand how both products relate.
Claude Design empowers non-designers to make decent designs. It’s not aimed at designers.
Figma will probably better integrate AI in their own offering at some point which will help designers become more productive. And that will be the end of it.
Haha, that's exactly how cc was received initially. It's just autocomplete. It's useless. It can't even x. I tried to y and it gave me z. Over and over all over the internet this was the reaction. Then the bargaining began. Oh, it will maybe speed up some simple things. Like autocomplete on steroids. Maaaybe do some junior tasks once in a while. And so on...
To me, it seems obvious that AI will attack this from both directions - upskilling developers to make more design changes AND upskilling designers to make more design iterations and more changes to the codebase -- the design artifact is "new react components" (which can be re-implemented or not) instead of a figma design.
It is pretty good for internal apps and dashboards or small hobby pages and websites where being generic look and feel doesn't matter much.
Most web design is already crap to begin with, so AI web-design will fit right in.
Plus compared to the totally open-ended video generation, web desisn is mostly samey (follows a few trends and conventions), way more restricted, and doesn't include difficult-to-recreate (due to uncanny valley effect) humans in it.
It still produces subpar code, with horrendous data access patterns, endless duplication of fucntionality etc. You still need a human in the loop to fix all the mistakes (unless you're Garry Tan or Steve Yegge who assume that quality is when you push hundreds of thousands of LoC per day).
Same here.
Oh, and Claude Code is significantly worse at generating design code than almost any other type of code.
Not really. 2024 called.
Opus 4.7, high effort. Literally 30 minutes ago. There's a `const UNMATCHED_MARKER = "<hardcoded value>"` that we want to remove from a file. Behold (the first version was a full on AST walk for absolutely no reason whatsoever, by the way):
function restoreUnmatchedSelectors(cssPath) {
const cssContent = readFileSync(cssPath, "utf-8");
const marker = UNMATCHED_MARKER + " ";
const count = (cssContent.match(new RegExp(escapeRegExp(marker), "g")) || []).length;
if (count > 0) {
writeFileSync(cssPath, cssContent.replaceAll(marker, ""));
return count;
}
}
function escapeRegExp(s) {
return s.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
}
Maybe if you had a little more faith, you'd get better results :)
Ah yes. A sed one-liner in a Javascript code
> Maybe if you had a little more faith
Ah yes, all I have to do as an engineer, is to ... checks notes ... have faith
Haha, maybe by you. By many on HN, but HN is a bubble of its own. By plenty of others it was received very differently. Many of us had been doing agentic coding for more than a year already when Claude Code was released, because we found it valuable.
We will see if such groups of professional designers also form for Claude Design or other such tools.
I guess I had expected something like Claude Code with visual tools added on top, but that’s not what this is.
At this point, Claude now writes > 99% of my code. I wasn't an enthusiastic early adopter; it took me a while to be willing to let go of the reins. But in tandem with LLMs getting better, I also began to realize that what happens inside the code is very rarely important enough for me to care about. Like, I care that it's secure, and performant where it needs to be, etc. -- but mostly I just care about its outputs. If it does what I want it to do, then how it does this doesn't really matter to me or my clients or my users. On the development side, my attention has focused to writing specifications and monitoring the correctness of the test suite, and > 99% of the time that's good enough. It's been a lesson in non-attachment to let go of lovingly crafting every single line of code, but the tradeoff in terms of productivity has absolutely been worth it.
What makes this viable is the fact that there's essentially a "hidden layer" (the code) upon which Claude can operate. My clients don't actually care about the code, and within certain bounds (correctness, security, performance, extensibility, etc.) it turns out that neither do I. This gives Claude a lot of latitude to solve things in its own way, and I think that's a bit part of its effectiveness.
On the other hand, with design there is no hidden layer. Every single aspect of the design is visible to the user and the customer. So the design reflects upon my work in ways that code does not. This means that the conditions which allow me to relax my grip on coding just don't exist for design. It's very difficult for me to imagine delegating design in the same way that I've become comfortable delegating coding.
That said: I suspect that the use-case for Claude Design will be for applications which today receive very little design attention. There are loads of applications where design is less than an afterthought, and the product suffers terribly for it. Delegating to Claude, in those contexts, would likely be a very big win. But for applications which already have designers obsessing over every pixel, I see a very limited role for this. Figma's market is mostly the latter -- the former, by definition, is not part of the market for design tools -- so I don't see them being threatened by this for a long time.
But as a person employing designers, I have already accepted that I will let go.
We did a marketing website redesign for our b2b saas product with a 3rd party design firm, we gave a lot of input, but the thing isn't perfect, at some point we had to call it done. It was still a significant improvement over what we previously had, but I am under no illusions that it is a masterpiece.
Now, coding tools do have some clear shortcomings for design atm, but how long they will be like that is not clear.
For example, in the world of e-commerce, one goal is improving conversion rate, as long as we get that and the design looks nice, that's OK.
These sort of issues are what you trust the judgement of a good designer to navigate through. I have no doubt that Claude Design can be better than no designer, and probably better than a bad designer, too. But better than a good designer? I'm more skeptical of that than I am of software engineering.
So...we can shitcan the designers and offload the work to the 10 developers still keeping the lights on?
Enterprise is not gonna lay off all their designers any time soon.
Sure you can, and many otherwise would-be human-designer-clients, will do exactly that. Nike might not, but 99% of the companies hiring designers are not Nike scale.
Many companies are not really based around their web design being a big factor in their business. Some very influential companies in my area look like they last updated their website in 1998.
Well, when you put it that way, that sounds bad for designers, and, by extension, Figma.
ps. I do like commas.
There is a bunch of repetitive work in design as well, once you start working on larger projects. Yes, everything should be setup with components/reusability and what not, but just like programmers take shortcuts sometimes, so do designers, and you have to repeat the same change across many instances/files whenever you have to pay back the "technical debt".
Probably Claude Design could be quite helpful in those cases, and the same goes for other domains too, same happens in video editing and 3D work, probably any creative effort has moments of dull, repetitive "do this change across X" where any automation would be of serious help to reduce that. It seems like a quite good thing to try to address with LLM tooling, still driven by actual humans.
People that think that Claude Design is going to replace Figma don’t really understand how both products relate.
I think the most effective teams will be working within Claude, not within Figma.
For individual creators, this will definitely replace Figma. I bought Sketch for use as an individual creator because I wanted to create mocks before coding them. There's no way I'd make the same purchase today.
Anyways, I'm sure Claude Design will incorporate some of Figma's features such as a company wide design language.
What makes you think that I didn’t read the article, but rather just disagree with it?
“which of course you did not read” is such a negative/toxic statement that adds no value.
obviously developers use the product to collaborate with designers. but it’s not the developers that are buying this product. they’re just stakeholders.
If designers still want Figma then the other people are along for the ride (unless the idea is the designers are being replaced with a PM+Claude.)
There are Figma plugins that let you extract a static HTML website into a Figma file. Copying that over Figma Make and prompting for a while can make pretty good prototypes that need very little adjustment back in Figma.
However, I believe that being able to do something doesn’t mean you should do it. Prompting back and forth can easily introduce a lot of cognitive load on top of all sorts of other daily task.
I feel the modern human in the loop is similar to the factory processing line workflow where just almost anyone can learn how to use a tool and produce output.
Maybe it will replace (a large share of) Figma users.
They will gladly use something like this (many have already started experimenting with other similar products) to get them even 60% of the way there and then they can polish the rest in code...
Which is basically how they used Figma before. Visualize to close enough and then iterate to final in code.
If Claude Design can ingest your design system and previous examples and go further than templates and scaffolding, if it can help you brainstorm ideas and variations so you can - as the human in the loop - get to your final design faster..
Why wouldn't you do that?
I don't think this is correct. In my experience no one buys Figma because of Dev Mode only. Dev Mode just makes it easier/faster to go from an existing design to working code. So it is/was a means to increase Figma's moat, not to get new customers or users. (Devs already needed access to Figma before the introduction of Dev Mode.)
Perhaps this will change soon if AI models reach the "army of geniuses in a datacenter" level, but current models are a far cry from just being able to clone Jira or Asana.
All they have to do is hold back a super capable model like Mythos while using it to clone your entire product. There’s nothing Figma, Salesforce, Workday, etc could do.
Correct, design is really about understanding something and curating a solution for it. In tech, that just happens to mostly be distilled as mockups.
That being said, my team is increasingly not needing Figma itself for many features:
* The first pass is often a lovable prototype or some low-fidelity mockup. These are enough to get us all aligned on what's being built.
* Engineering takes a first pass. Gets a UI base in place.
* Designer (who can vibe code) or UI engineer comes in to put the high fidelity touches on things.
Basically, the product itself becomes the mockup.
These are a design communication tools, in the same way FrontPage and Dreamweaver back in the day - they enable a designer to communicate intent to create solutions to UX design problems in a way that can be understood by downstream production.
Using Figma doesn't immediately make something aesthetic or useful, its just a tool. Claude Design enables the same thing, it enables someone to communicate design intent, with a much lower bar.
These AI tools like Loveable just do more of the groundwork than Figma does
Design is the process to you follow to solve a given problem/need/desire balancing the user needs with the business/tech constraints. The output could be a a digital UI, a physical object or an intangible process. But often people think about design just as the aesthetics of a product.
A common pattern is the Empathize (Research, Diverge), Define (Converge), Ideate (Diverge), Prototype (Converge), Test (and then iterate).
The main benefits come from the divergent phases: empathize and ideate. But it’s far too common already that some executive has an “illumination” of how something should be and just wants to build as is, without any research or validation.
They can use Claude Design (and similar) to just build a prototype of their first idea, skipping all the design process and end with something that looks good but doesn’t solve the problem adequately or fits the actual user needs/context.
Of course, LLMs are useful tools that can be used in the right way: to build better prototypes in less time, to synthesize research insights, to explore ideas…
Design is its own thing. Fundamentally it's problem solving, but design is thinking about the user of the 'designed' thing and how will they use it, what will they do with it, what will it enable them to do, how will they learn how to use it, how can it be ergonomic to the environment it will be used in and the user.
A lot of people can conceptualise a factory making widgets or a programmer writing code, but drastically few people really see the design process: sketches on paper, wireframes in figma, design as a solution to a problem, design as story telling, or aesthetics. But those outputs are not the point.
The output of a few hours in figma is not really the images of a website made, it's more about communication or articulation of the the problem being solved and the solution that will solve it.
Which is why it doesn't really matter about the tool, design is an expression, the medium whether it's a sketch or a figma mock up or a vibe-designed UI in claude design is less critical than the thought that went into it.
And wait it gets even worse!
Why?
- Figma is sending Anthropic a bunch of training data from its own LLM assisted data. As much as Anthropic claims that it won’t use it, we all know what Amazon did with third party sellers.
- Anthropic hasn’t started to play hardball yet. Why wouldn’t they just hold back a model like Mythos (or better) while they use it to gut a few SaaS companies? It’s an easy way to increase their revenue!
What on earth does this even mean and how does it relate to design?