Learning Where AI Amplifies the Design Process—And Where It Slows Us Down

We haven't found as much value in image generation yet. And so we aren't using it as much for that as we are as more of like a writing assistant and early design phase ideation sort of tool.

Chad kind of talked about that at the conference. Can you say more about like why you found like ImageGen isn't landing for Snohetta? And there was this great quote in that section of your presentation, like, I don't use AI for design, I use it for ideas.

Yes.

Which still makes every time I read it or hear it, it makes me laugh. But like, can you talk about more what that means?

Yeah, absolutely. What we found is that when we experiment with it for image generation, the images that it comes out with, it depends on the prompt, of course, but oftentimes the images that's coming out with is of designed or it appears to be a designed building or a designed landscape, it's like already delivering the whole package. Whereas when in our typical conceptual process we are looking at abstract ideas, we're drawing metaphors with natural forms or you know, geometric forms, whatever. There's a lot of abstraction is truly the word abstraction. And metaphor is what we're usually looking at in early concept design. And so seeing a finished building, it goes way beyond that level and it starts to be like a period at the end of a sentence rather than more of an ellipses, which is I think what we're going for in the early design phase. So we're having more luck or more interesting results just by using text based conversation. If we use AI for ideation at all, it is more often for text based sort of back and forth of feeding it a bunch of ideas that came out of a group brainstorm and trying to like get it to escalate those ideas, amplify them. And that's how, that's what the person was referring to when they said I don't use it for design, I use it to generate ideas is they're usually they are literally throwing a bunch of their ideas into an LLM and seeing what happens when they try to dial.

All of that up.

And then they're extracting ideas from that and taking that back and using that to do more image brainstorming. And that is generating more interesting, richer results than the image generation tools themselves.

I love this idea that you're talking about and it seems to me it really requires knowing your values. Like this idea that you proposed a few minutes ago about refocusing the way you do budgeting based on the values that the firm has and using technology to augment the values of the firm. But like megaphone out to the world. This is who Sno etta is. And this is why we do what we do is super important to keep that feedback loop going internally, but as well, you know how that makes it out in the world. I love that idea of kind of this whole value based tech stack that you're talking about.

It's a conversation that Chad, our design technology lead and I have often, because it is something that's been very important for him throughout his whole entire career, is the impact that technology has on the design of spaces. It's something that he was interested in, you know, even before he started his professional career. And it's still the case. And we see it every day. Like the buildings and spaces around us change because of what the technology we use to design them is capable of doing. And that's wonderful and so interesting. But it also means that we can start to develop spaces that truly do leave behind the humans and the other beings and ecosystems because the technology isn't connected with all of that in the way that we are so maintaining that connection and making sure that we are intervening to create the balance between those two things. We want to embrace technology and everything it can do, but we don't want it to have. We don't want it to be in the driver's seat, in my opinion.

Chad kept using a Marshall McLuhan quote when we were preparing for your talk about like, every new technology is both an amputation and a augmentation, I think is how it went.

Every augmentation is also an amputation.

Every augmentation is also an amputation. And so miro augments you, but then it cuts off pinups and whiteboards and, you know, all that kind of stuff too.

It's true and it's possible. I mean, it will always be true. The augmentations will lead to amputations. But what we've talked about is that we want to be involved in making the decision about what's amputated.

You don't want it to just happen to you.

No surprise. Amputation.

Amputations. No surprise. Amputations, please.

That's not the title of this episode. Evan. I know what you're thinking. Making an executive call right there.

I see.