Look, it’s cool getting systems to talk to each other. But….before that…
why do they have to talk to each other anyway?
Some of the UX fellows see their job being compromised by the advancements of AI. Indeed, it’s true, if all you’re doing was some black and white boxes and arrows wireframe on Figma, XD, miro (yes, people are doing stuff there) Axure (okay, I’m old), Illustrator (did I say illustrator? Unfortunately, yes. I saw a website design done in Illustrator in 2023!).
I was never in agreement with the segregation. Design is design. It involves lots of things, depending on the area you design. But this is a conversation for another time. Another post.
Back to the MCP hype. I got excited. Figma IPO, figma MCP. Let’s try it (the second one, for now).
I plug it in to VS Code, and voila. Relatively simple. Then I look at the use cases. Can’t understand it. Can’t imagine this in the real world. Although it looks cool, it’s a scary freedom in the software lifecycle. Even if you are experimenting. How do we roll back? Your code has something that your design doesn’t. In my head, we are breaking so many rules.
Okay, I get it. Imagine how quick you could run a user testing, with real product, just by prompting from existing code? But the fact of selecting stuff on Figma, vs just writing a prompt in Copilot directly, did not convince me of why I needed a Figma MCP.
By the time we have a two-way flow, design to code to design, I am happy to meet you again, MCP.