May 2026
Last week if you'd asked me how I felt about Figma I'd have been kind of an asshole about it. “Oh, they've spent the last year doing everything wrong, you can't use AI to generate stuff on the canvas, like just look at their stock price, everyone agrees with me on this, I am of course right about everything.”
And then Claude Design came out.
AND I THINK I WAS COMPLETELY WRONG.*
I want to say I am still incredibly bullish on Anthropic. Have you heard their headquarters are basically a skiff? They know they're on to something. I've been trying to put it like this:
Claude connects everything. Give Claude access to all of your things (haha, fun, privacy, idk what to tell you†) and you can do stuff like, “Pull this PRD from Atlassian, look at this epic in Jira, grab context from the latest DMs with my manager in Slack, and put together a flowchart for how this new feature should work.” And it'll do it. It's super neat. If I never have to write another bug ticket myself ever again it will be too soon. I like how the little robot dances in my terminal. I'm looking at new token multipliers across the industry with a tear in my eye because I suspect my license is about to become a lot less useful. I ♥ Claude.
But I'm a designer, and designers hang out in Figma. So let me take you on the saga of how I ~how I wasted all my fucking time for the last three months~~~
B2B SaaS right now feels like like some weird psychological experiment where every week or so someone in a meeting demos some tool you've never heard of, says it's the best thing since sliced bread and we should all begin using it, and so you try to start using it and then get burnt out because you have no time to play with stuff because you have your regular work. This then repeats ad infinitum until you lose your goddamn mind. The tool exhaustion is real. If you're a designer, it feels like every engineer around you has been given a machine gun and you're showing up to every meeting with like, a knife or something (idk I'm bad at analogies). You get the picture.
Design tools are just... slow. As an industry we've moved away from pencil wireframes because they're hard to track digitally, which I think has been a detriment to testing ideas in low-fidelity. That said, they were also really awkward. “Hey stranger, wanna test out my mobile app? Pretend you're clicking this app icon I drew on a piece of paper.”‡ You spend a lot of time as a UX designer moving pixels around really slowly. It's just kinda a bitch.
Which is why access to AI for in-browser prototyping is so awesome. You can actually move with some speed, and demo and feel how something looks and operates in the environment it will actually deploy in when it's finished. I myself have completely switched over to vibe coding because I'm a terrible engineer and the code I write is no better, typically, than your average new engineer fresh out of college. It's made coding a lot of fun again.
However! Brute-forcing learning just enough of something means that I am one of those, “Should Designers Learn to Code?”-types who said sure i guess and is living with the consequences.
Which is my fun little segue into Helen lore you didn't ask for.
The year is 2015. I am at my first corporate job. I am a “Front End Developer” (heavy emphasis on the quotes for that) on a Design team. I build prototypes for ideas that Invision doesn't work for. I am learning what web sockets are so I can make mocks of 2-step authentication because that wasn't really a thing yet. I am learning just how much coffee I can consume on no sleep before I vibrate out of my skin. Etc., etc.
There is heavy discussion in the industry of, “Should designers code?” and so every so often, as a person who knows how to find the terminal on her computer, I sit down with a designer in my office, install Xcode on their Mac, and get them set up with a repo where they “just want to make a few CSS changes” which they will inevitably not do, but want to, because their Design QA ticket has been sitting untouched in Jira for three weeks post-launch and it's driving them batty. (I actually really empathize with this, so if anyone in Product has stumbled upon this article, please be kind to your designers. Give your engineers some time to address these sorts of things.)
I say they will inevitably not do these CSS changes because engineering tends to have ever so slightly more to it, so it's not uncommon I imagine as an engineer to receive a PR from a designer where you go, “Neat”, and then close it and re-do it because it's just easier. No shame in the game. I've already ranted about this. There's utility in having the people who are good at the thing you pay them to be good at doing that thing, which in the case of engineering, is... an engineer.
BUT, since this is all just a ton of context for how I wasted a bunch of time over the last three months, guess what I spent a bunch of time doing in the year of our Lord 2026?
Did you guess attempting to set up dev environments on a bunch of designers' computers remotely???? BECAUSE THAT'S WHAT I DID!!!! And I have learned a very painful lesson that somehow I didn't learn a decade ago, which is that the designers who want to get good at this sort of thing are just going to do it, and the ones who don't won't. Which is... fine.
All right, so back to the top of this post, which is that truly, all of a week ago, I was so excited about Claude Design. And to be frank I think it's still really, really neat that they basically just made a 30,000 word prompt and turned it into an entire tool. Talk about dogfooding your own product. Geez.
That said, I'm recommending to people for now that like... you should just keep using Figma Make. Which is wild. A week ago I was so gung-ho for Claude Design. Everything in that one little Mac app? Hell yeah!
I think TJ at Southleft has done a better job of summarizing this than I can do justice here. Basically: if you're trying to build prototypes that are grounded in your company's actual design system... Claude Design just isn't it yet. I imagine that could change over time: it'll be interesting to see how Anthropic decides to position this tool. Chipping away at Adobe and Figma's market share is fun for me to watch as a consumer, I'll say at the very least.
But for now: if you're trying to use design systems, need enterprise-level protected password prototype sharing, need tooling that abstracts away the development environment set up for designers, need tooling that meets designers where they actually are... I think the right tool for now is still Figma Make.
(The messy bitch who lives for drama in me though cannot wait to go to Config this year.)
What a fun and stressful time to be in tech.
* Always fun and humbling, obviously
† I feel like I've sold my soul at this point
‡ idk i kinda miss 2015