Bonus

|

Episode 2

How you can get the most out of AI tools

Helen Zhou

Co-founder of Galileo

Feb 13, 2024

Feb 13, 2024

|

32 mins

32 mins

music by Dennis

About this Episode

Galileo recently broke design twitter with the launch of their new generative AI tool for builders. So in this episode we talk with their co-founder (and lead designer) Helen Zhou about:

  • Her journey as a design founder

  • Why user research for AI tooling is unique

  • How AI will change the day-to-day role of a designer

  • Her response to concerns about LLMs and creativity

  • The potential breakthrough use case for UI generation

Share on Twitter
Share on Twitter
Share on LinkedIn
Share on LinkedIn

Deep Dives

Get our weekly breakdowns

Insights + resources from top designers πŸ‘‡

Lauren LoPrete

Director of Design Systems @ Cash App

David Hoang

VP of Marketing and Design @ Replit

Adrien Griveau

Founding Designer @ Linear

James McDonald

Designer @ Clerk

Femke

Design Lead @ Gusto

Join 10K+ designers

HC

HC

HC

HC

Deep Dives

Get our weekly breakdowns

Free lessons from πŸ‘‡

Lauren LoPrete

Lead designer @ Netflix

David Hoang

VP of Marketing and Design @ Replit

Adrien Griveau

Founding Designer @ Linear

Femke

Design Lead @ Gusto

Join 10K+ designers

HC

HC

HC

Deep Dives

Get our weekly breakdowns

Insights + resources from top designers πŸ‘‡

Lauren LoPrete

Director of Design Systems @ Cash App

David Hoang

VP of Marketing and Design @ Replit

Adrien Griveau

Founding Designer @ Linear

James McDonald

Designer @ Clerk

Femke

Design Lead @ Gusto

Join 10K+ designers

HC

HC

HC

HC

Transcript chapters

[00:00:00] Galileo's launch day

Helen: We've been in private beta for about six months now, and the goal of the launch is really, I think two things. One is that we wanted to make this product really public to the world and get the message out. , and the other thing is, it's a very.

, technological challenging products. So we wanna make sure that we nail the scalability and stability. And especially with AI products, , it could be very tricky if you get a large traffic onto your product. I think you've seen the, the Twitter posts and we're like, well, you know, I think we, , had an idea of like, if we get something like a hundred thousand impressions, that would be pretty solid.

And if we get 200. Thousand That would be amazing. , but after we get it out, we quickly realized that this thing is, , exploding. It was getting like hundreds of thousands of impressions and, it really kind of brought this like exhilaration, , internally. 'cause our whole team had be like, oh my God, now we have to add servers.

Now we have to upgrade. We upgraded probably five, six different infra. Services that we use in like 48 hours, uh, because of the traffic that , was coming in. I mean, We were anticipating it. It wasn't like, oh, let's just go out and see what happens. So our team worked for weeks to prepare the lunch, but when you're living it, , it still feels different,

Ridd: I just looked at the tweet. You're almost at 750,000. Impressions, like you took over design Twitter and then of course you have like a few hours later, you know, your, your co-founder tweeting like, hey, like, servers are like basically going down, we're running around, everything's on fire.

And it's like, that's kind of the startup dream a little bit.

Yeah, I mean, it was very exciting. Definitely, like, I would say greatly exceeded our expectations we've been building this, , in stealth for a while and we've been testing it in private beta for a long time. , and, you know, you're very nervous. It's like, oh, what's gonna happen when you like, put it out in the world?

And we're super happy that I think most of the reactions we get from people , are quite positive. So . it felt quite rewarding for not only for me, but for my co-founder and for my whole team as well.

Helen: I think we've gotten, over 50,000 designs generated in twenty-four hours. I. And we've gotten in less than three days, probably somewhere around a hundred thousand designs. So when you think about a number, it's, I, I think it's quite phenomenal.

[00:00:00] Galileo's launch day

Helen: We've been in private beta for about six months now, and the goal of the launch is really, I think two things. One is that we wanted to make this product really public to the world and get the message out. , and the other thing is, it's a very.

, technological challenging products. So we wanna make sure that we nail the scalability and stability. And especially with AI products, , it could be very tricky if you get a large traffic onto your product. I think you've seen the, the Twitter posts and we're like, well, you know, I think we, , had an idea of like, if we get something like a hundred thousand impressions, that would be pretty solid.

And if we get 200. Thousand That would be amazing. , but after we get it out, we quickly realized that this thing is, , exploding. It was getting like hundreds of thousands of impressions and, it really kind of brought this like exhilaration, , internally. 'cause our whole team had be like, oh my God, now we have to add servers.

Now we have to upgrade. We upgraded probably five, six different infra. Services that we use in like 48 hours, uh, because of the traffic that , was coming in. I mean, We were anticipating it. It wasn't like, oh, let's just go out and see what happens. So our team worked for weeks to prepare the lunch, but when you're living it, , it still feels different,

Ridd: I just looked at the tweet. You're almost at 750,000. Impressions, like you took over design Twitter and then of course you have like a few hours later, you know, your, your co-founder tweeting like, hey, like, servers are like basically going down, we're running around, everything's on fire.

And it's like, that's kind of the startup dream a little bit.

Yeah, I mean, it was very exciting. Definitely, like, I would say greatly exceeded our expectations we've been building this, , in stealth for a while and we've been testing it in private beta for a long time. , and, you know, you're very nervous. It's like, oh, what's gonna happen when you like, put it out in the world?

And we're super happy that I think most of the reactions we get from people , are quite positive. So . it felt quite rewarding for not only for me, but for my co-founder and for my whole team as well.

Helen: I think we've gotten, over 50,000 designs generated in twenty-four hours. I. And we've gotten in less than three days, probably somewhere around a hundred thousand designs. So when you think about a number, it's, I, I think it's quite phenomenal.

[00:02:28] Use cases Helen is excited about

m kind of curious, like you've seen all of the different ways that people are exploring and using product and the different types of people even using the product. Are there certain use cases that you are particularly excited about right now?

Helen: There were two use cases that stood out during our private beta period that was quite obvious that we had lots of designers using us to kind of doing fast. Ideation. , we spoke to this designer. he's part of a design agency in Norway, and they were designing this like cryptocurrency app.

, and he was like, oh, like I. I use Galileo and it gives me all the key screens of that crypto app. He said in under three hours. And he said, without Galileo, I would've, it would've taken me days to put all these screens together. He was like, oh my God. Like this tool is amazing. So we've heard a lot of those type of use cases from designers is like, oh, it gives me this initial.

Set of screens super fast, and then I'm just gonna go, you know, tweak it and, and finalize it and saves me same amount of time. That's the first use case that , we've seen. The other one that stood out to us is , the startup founders. , there was, , one person that they, , are doing this like healthcare startup in San Francisco.

And he doesn't know how to design, but he's like, oh, well now I can generate ideas with Galoo and pass it on to. My dev team super quickly so they know what I'm talking about. And he's also using us to do customer interviews. He's like, oh, well now I have this, like five new ideas. I'm just gonna generate it before I build it to test it with the customers and to see how they like it.

So I think that's another kind of use case I think is really beneficial to the audience because as a startup founder, you have to move really fast. And there is this third use case. I think what we realized after we launch, which is a lot of, , software developers started to use us. It's really interesting, they kind of like, oh, I, I'm, I'm developing this news app. And now with Galileo, I generated some designs , and then I quickly, you know, put the prototypes together.

Ridd: I think all of those are very interesting, especially the fact that you're in a sense democratizing design for non-designers. There are a ton of implications about that that I do wanna just like save time and get to, , and have a little bit more of a meaty discussion.

[00:02:28] Use cases Helen is excited about

m kind of curious, like you've seen all of the different ways that people are exploring and using product and the different types of people even using the product. Are there certain use cases that you are particularly excited about right now?

Helen: There were two use cases that stood out during our private beta period that was quite obvious that we had lots of designers using us to kind of doing fast. Ideation. , we spoke to this designer. he's part of a design agency in Norway, and they were designing this like cryptocurrency app.

, and he was like, oh, like I. I use Galileo and it gives me all the key screens of that crypto app. He said in under three hours. And he said, without Galileo, I would've, it would've taken me days to put all these screens together. He was like, oh my God. Like this tool is amazing. So we've heard a lot of those type of use cases from designers is like, oh, it gives me this initial.

Set of screens super fast, and then I'm just gonna go, you know, tweak it and, and finalize it and saves me same amount of time. That's the first use case that , we've seen. The other one that stood out to us is , the startup founders. , there was, , one person that they, , are doing this like healthcare startup in San Francisco.

And he doesn't know how to design, but he's like, oh, well now I can generate ideas with Galoo and pass it on to. My dev team super quickly so they know what I'm talking about. And he's also using us to do customer interviews. He's like, oh, well now I have this, like five new ideas. I'm just gonna generate it before I build it to test it with the customers and to see how they like it.

So I think that's another kind of use case I think is really beneficial to the audience because as a startup founder, you have to move really fast. And there is this third use case. I think what we realized after we launch, which is a lot of, , software developers started to use us. It's really interesting, they kind of like, oh, I, I'm, I'm developing this news app. And now with Galileo, I generated some designs , and then I quickly, you know, put the prototypes together.

Ridd: I think all of those are very interesting, especially the fact that you're in a sense democratizing design for non-designers. There are a ton of implications about that that I do wanna just like save time and get to, , and have a little bit more of a meaty discussion.

[00:04:53] Helen's pre-launch user research process

Ridd: But first I want to. Zoom into a couple little parts of your process, learning more about the early stages of this journey, and the first, you hinted at a little bit, which is some of the learnings from that alpha period. I'd love to kind of just hear you talk about. How you've approached the user research process for this new startup?

Like what were some of the key things that you were trying to learn? How did you structure the process? And then ultimately what was the output? Like , how did impact your product strategy?

Helen: Now it's like you type something and AI generate you something, and everyone's like, yeah, it's, it's a thing. But back in late 2022, you know, we talked to investors, we talked to users, and I think the general sentiment we got was like, that sounds kind of weird.

Why would you wanna do that? We did user interviews, which there was a lukewarm kind of reaction. But, you know, we went with , the direction anyways and when we actually, , developed the technology and put the product in front of people, I think they immediately got it.

So I think the biggest realization I had is that I think I. When you work on AI product, it is very different than maybe some of the other verticals, is that it's a little bit like consumer product, which is you really have to ship it to let people, to use it to get the real reactions right? I think because the experience could be super customized and personalized by just by talking to people about what do you think of this idea?

I don't think you're gonna get a fair. , evaluation of what they really think of what you're building. If you are designing the Instagram app, if you're just showing a mock of Instagram to people versus them using and seeing the photos of their friend, I mean, it's a very different experience, which I think for AI products, it's kind of in the consumer bucket, which is , the most important thing I learned is that you really have to ship fast. The best way to learn is that you have to put the product in front of your customers and see how they're using it.

I think that's the best way , to learn, , kinda compared to what we did initially. Just talking about concept with people.

Ridd: I think it's interesting to call out for people that you are a first-time founder, like your background is in design. You were a long-time designer at Facebook, site manager at Cruise, so you're learning a lot of these things and going through a lot of these challenges for the first time. My question is, what skills have you had to learn on the fly to complement what you already bring to the table with your design background?

Helen: I think it is interesting that I, sometimes I feel like I almost have to unlearn some stuff I learned as a designer. I think designers, you know, we value quality and , we do thorough research. We make sure that we design a great end-to-end experience. But when you work at a startup,

what I found myself is that as a designer, , I actually had to do less design work. I think that there is simply just, you know, not enough time, sometimes not enough resources to do , the most ideal experiences that you would hope for.

So you really have to learn. How do you do the trade-offs with the resources and the time you have? And then , the next thing is that as a founder, you kinda have to juggle between a lot of different priorities. You know, , there's product, there's marketing that there's hiring, there's legal, there's a lot of different things, I think being a maker, being a builder, like we tend to spend time on what we're good at because that's kind of your comfort zone. But at the same time, you have to take a step back and think, well, what is the most important thing? For my company right now. Maybe it's not design, maybe it's something else and I need to spend more time on those things.

You kind of have to battle with your instinct a little bit of like, oh my God, I'm gonna do a perfect design, versus, okay, there is a lot of other things that I really need to take care of.

[00:04:53] Helen's pre-launch user research process

Ridd: But first I want to. Zoom into a couple little parts of your process, learning more about the early stages of this journey, and the first, you hinted at a little bit, which is some of the learnings from that alpha period. I'd love to kind of just hear you talk about. How you've approached the user research process for this new startup?

Like what were some of the key things that you were trying to learn? How did you structure the process? And then ultimately what was the output? Like , how did impact your product strategy?

Helen: Now it's like you type something and AI generate you something, and everyone's like, yeah, it's, it's a thing. But back in late 2022, you know, we talked to investors, we talked to users, and I think the general sentiment we got was like, that sounds kind of weird.

Why would you wanna do that? We did user interviews, which there was a lukewarm kind of reaction. But, you know, we went with , the direction anyways and when we actually, , developed the technology and put the product in front of people, I think they immediately got it.

So I think the biggest realization I had is that I think I. When you work on AI product, it is very different than maybe some of the other verticals, is that it's a little bit like consumer product, which is you really have to ship it to let people, to use it to get the real reactions right? I think because the experience could be super customized and personalized by just by talking to people about what do you think of this idea?

I don't think you're gonna get a fair. , evaluation of what they really think of what you're building. If you are designing the Instagram app, if you're just showing a mock of Instagram to people versus them using and seeing the photos of their friend, I mean, it's a very different experience, which I think for AI products, it's kind of in the consumer bucket, which is , the most important thing I learned is that you really have to ship fast. The best way to learn is that you have to put the product in front of your customers and see how they're using it.

I think that's the best way , to learn, , kinda compared to what we did initially. Just talking about concept with people.

Ridd: I think it's interesting to call out for people that you are a first-time founder, like your background is in design. You were a long-time designer at Facebook, site manager at Cruise, so you're learning a lot of these things and going through a lot of these challenges for the first time. My question is, what skills have you had to learn on the fly to complement what you already bring to the table with your design background?

Helen: I think it is interesting that I, sometimes I feel like I almost have to unlearn some stuff I learned as a designer. I think designers, you know, we value quality and , we do thorough research. We make sure that we design a great end-to-end experience. But when you work at a startup,

what I found myself is that as a designer, , I actually had to do less design work. I think that there is simply just, you know, not enough time, sometimes not enough resources to do , the most ideal experiences that you would hope for.

So you really have to learn. How do you do the trade-offs with the resources and the time you have? And then , the next thing is that as a founder, you kinda have to juggle between a lot of different priorities. You know, , there's product, there's marketing that there's hiring, there's legal, there's a lot of different things, I think being a maker, being a builder, like we tend to spend time on what we're good at because that's kind of your comfort zone. But at the same time, you have to take a step back and think, well, what is the most important thing? For my company right now. Maybe it's not design, maybe it's something else and I need to spend more time on those things.

You kind of have to battle with your instinct a little bit of like, oh my God, I'm gonna do a perfect design, versus, okay, there is a lot of other things that I really need to take care of.

[00:08:44] What Helen believes about the future of design

Ridd: Let's zoom out a little bit 'cause I want to get your perspective on kind of where this all might be headed. Obviously we're all guessing to an extent, but guess is as educated as anyone else's that I'm going to talk to. So maybe we could even start by talking a little bit more about.

What do you believe about the future of design that ultimately led you to start Galileo in the first place?

Helen: know, I,

I was a designer at, , Facebook for. More than six years. And, , then I led a design team at Cruise, the self-driving car company. , so I've been kind of in design for almost a decade. And, you know, the longer that I'm in it, the more that it's, it's a feeling that has been built up stronger and stronger is that, I think as a designer in your kind of earlier stage of your career, you, you, you spend a lot of time in the tools. You think about the craft. You think about, okay, I'm gonna draw out this buttons. I'm, I'm going to, design this form. But as you grow as a designer into kind of a more senior level and, and you started to lead a team and you started to help other people grow and you realize kind of.

The, the really important skills is that you, you really need to understand the users that you're designing for, the business that you are designing for. Like what is the, , the objectives of your business and why is your business doing certain things and.

What is everyone else doing in your organization and how does your work fit into the context? I think knowing that, being able to communicate it, being able to kind of distill all these information and make the right decision, I think fundamentally that is a really critical skills for a designer.

But you know, I've seen a lot of designers. They're, they're amazing strategists. You know, they're great thinkers, but. They simply just don't have enough time , to spend on those things , to learn about the business, to develop those skills because they had to spend a lot of time in the tools to do, you know, kind of like a repetitive task they've been doing, you know, again and again and again.

And to me, I think when I see the. Advancement of the large language models in 2022, I, I knew that it, it has reached a point where it could potentially help designers to automate some of the repetitive tasks that they've been doing so they have more time to focus on things that I think where they can truly shy.

Ridd: I can't just give you like softball questions because I think there are some people out there that are a little bit skeptical and back on some of these changes, and so. I'm gonna put on my like design Twitter troll hat really quickly , and push on this just a little bit because even like the use cases that you listed, you talked about like founders being able to design, maybe that prolongs the time that they're able to go before hiring a designer.

[00:08:44] What Helen believes about the future of design

Ridd: Let's zoom out a little bit 'cause I want to get your perspective on kind of where this all might be headed. Obviously we're all guessing to an extent, but guess is as educated as anyone else's that I'm going to talk to. So maybe we could even start by talking a little bit more about.

What do you believe about the future of design that ultimately led you to start Galileo in the first place?

Helen: know, I,

I was a designer at, , Facebook for. More than six years. And, , then I led a design team at Cruise, the self-driving car company. , so I've been kind of in design for almost a decade. And, you know, the longer that I'm in it, the more that it's, it's a feeling that has been built up stronger and stronger is that, I think as a designer in your kind of earlier stage of your career, you, you, you spend a lot of time in the tools. You think about the craft. You think about, okay, I'm gonna draw out this buttons. I'm, I'm going to, design this form. But as you grow as a designer into kind of a more senior level and, and you started to lead a team and you started to help other people grow and you realize kind of.

The, the really important skills is that you, you really need to understand the users that you're designing for, the business that you are designing for. Like what is the, , the objectives of your business and why is your business doing certain things and.

What is everyone else doing in your organization and how does your work fit into the context? I think knowing that, being able to communicate it, being able to kind of distill all these information and make the right decision, I think fundamentally that is a really critical skills for a designer.

But you know, I've seen a lot of designers. They're, they're amazing strategists. You know, they're great thinkers, but. They simply just don't have enough time , to spend on those things , to learn about the business, to develop those skills because they had to spend a lot of time in the tools to do, you know, kind of like a repetitive task they've been doing, you know, again and again and again.

And to me, I think when I see the. Advancement of the large language models in 2022, I, I knew that it, it has reached a point where it could potentially help designers to automate some of the repetitive tasks that they've been doing so they have more time to focus on things that I think where they can truly shy.

Ridd: I can't just give you like softball questions because I think there are some people out there that are a little bit skeptical and back on some of these changes, and so. I'm gonna put on my like design Twitter troll hat really quickly , and push on this just a little bit because even like the use cases that you listed, you talked about like founders being able to design, maybe that prolongs the time that they're able to go before hiring a designer.

[00:11:52] How AI impacts the market for design

Ridd: Maybe they don't even need a designer if they have even a little bit of an eye for design. You talked about engineers using this tool, Hmm. so. How does this kind of advancement where you're just just pushing the bar for what it takes to design something down to such a small amount, how does that impact the role of the dedicated product designer moving forward?

Helen: So I think on a higher level. believe the designer role is becoming more and more versatile and there is a lot of skills on communication, understanding the businesses and.

, understanding the users. And I don't think. AI is here , to replace those skills.

Now, in the use case that I talked about where founder be like, well, you know, I now can use Galio to generate ideas and pass it , to my devs. I think what that replaced is kind of the more time actually from the founder themselves that if he didn't have.

Being a small team like they are, he, they didn't have a tool like this. He probably would have spent a day kind of trying to cobble together some wireframes or write a really long, , product spec doc and try to pass it to the engineers he works with. Right. , and now he can just do this and which saves him a few hours, I think, as their business involved and things are getting.

More complicated and there's so much more customer context and business context they'll need someone , to be able to understand them and still into the right features to prioritize and build. That's the time that, as a founder, that's usually when you hire a designer to kind of help you lead those whole area.

And so I think to answer your questions, at least that what I've seen is like how people have been using our tool. I don't think that necessarily Replacing the person that they could have hired, I think they were replacing.

A lot of the, laboring and the time spending that they had to them to do themselves. Now with this tool, it just makes it easier.

Ridd: I was looking through your product launch comments just to kind of get a sense , of people were responding to it. And it's quite clear like the product is good. Like it was so many comments of people being like, you know, I went in with low expectations and actually this is legit.

[00:11:52] How AI impacts the market for design

Ridd: Maybe they don't even need a designer if they have even a little bit of an eye for design. You talked about engineers using this tool, Hmm. so. How does this kind of advancement where you're just just pushing the bar for what it takes to design something down to such a small amount, how does that impact the role of the dedicated product designer moving forward?

Helen: So I think on a higher level. believe the designer role is becoming more and more versatile and there is a lot of skills on communication, understanding the businesses and.

, understanding the users. And I don't think. AI is here , to replace those skills.

Now, in the use case that I talked about where founder be like, well, you know, I now can use Galio to generate ideas and pass it , to my devs. I think what that replaced is kind of the more time actually from the founder themselves that if he didn't have.

Being a small team like they are, he, they didn't have a tool like this. He probably would have spent a day kind of trying to cobble together some wireframes or write a really long, , product spec doc and try to pass it to the engineers he works with. Right. , and now he can just do this and which saves him a few hours, I think, as their business involved and things are getting.

More complicated and there's so much more customer context and business context they'll need someone , to be able to understand them and still into the right features to prioritize and build. That's the time that, as a founder, that's usually when you hire a designer to kind of help you lead those whole area.

And so I think to answer your questions, at least that what I've seen is like how people have been using our tool. I don't think that necessarily Replacing the person that they could have hired, I think they were replacing.

A lot of the, laboring and the time spending that they had to them to do themselves. Now with this tool, it just makes it easier.

Ridd: I was looking through your product launch comments just to kind of get a sense , of people were responding to it. And it's quite clear like the product is good. Like it was so many comments of people being like, you know, I went in with low expectations and actually this is legit.

[00:14:15] How AI increases iteration cycles

Ridd: There was definitely that theme though of people saying, Hey, I just took something that previously would've taken me days and I accomplished it in an hour or two hours, and so. I want to drill in even more. Like what do you think designers can be doing in their day-to-Day roles to move the needle at the companies that they work with When they are able to do so much more with less, given these technological advancements where maybe they are spending less time in the tools themselves.

Helen: My hypothesis is that. I don't think people would necessarily, spend lesser time , in designing the thing. I think they're just gonna have a faster iteration.

They're gonna be designing more things in a shorter period of time. I, I remember when, , I used to run like design Sprint at , Facebook, right? The design sprint is, you get a lot of folks together into a room and you spend days there and every day you like pump out. 10 ideas, , and at the end of the week you have like, oh, 30 ideas. Let's all take a look at them.

Lots of the teams and the companies that I've seen, they spend a huge amount of resources on these type of visioning exercise because it takes time to develop ideas, but the things don't have to necessarily to be finalized because it's essentially, it's a, it's a brainstorming exercise.

So I think that's kind of where a tool like this can shine. I'm like, well, if I have a quick idea, I'll just use this , to generate it because I'm not looking for a finalized version. I'm, just, communicating my ideas to my stakeholders and like I think people wanted to see the ideas being visualized.

I think with a tool like Aleo, you can have 20 ideas in three hours if you want. it's not like you're designing less, I think it's more like the things are going faster and now you can just, at the end of the day, hey, we have a lot of things being produced .

You can essentially go to your team, your product peers, your engineer peers. It's like, Hey, what what do we think about this? Like, which direction should we go deeper? I don't think , the overall process is necessarily changing that much, but I think that, that the velocity of the process is gonna be different.

Ridd: Yeah, I like that answer a lot. And I've been asking similar questions to other people and really just trying to understand like what are the key traits that you can identify across the best product teams? And so often it comes down to just being able to compress these. Iteration cycles to run more loops more quickly.

And I can see how a tool like this would help teams make that jump.

Helen: Yeah, absolutely.

Ridd: I've been kind of thinking about. Like where a tool like this fits into my personal design process, because it's not realistic, at least anytime soon for AI to be at the point where, you know, I just hit a button and the whole design is done. All of the thinking, all of the iterations, and it's pixel perfect, and all this kind of stuff.

[00:14:15] How AI increases iteration cycles

Ridd: There was definitely that theme though of people saying, Hey, I just took something that previously would've taken me days and I accomplished it in an hour or two hours, and so. I want to drill in even more. Like what do you think designers can be doing in their day-to-Day roles to move the needle at the companies that they work with When they are able to do so much more with less, given these technological advancements where maybe they are spending less time in the tools themselves.

Helen: My hypothesis is that. I don't think people would necessarily, spend lesser time , in designing the thing. I think they're just gonna have a faster iteration.

They're gonna be designing more things in a shorter period of time. I, I remember when, , I used to run like design Sprint at , Facebook, right? The design sprint is, you get a lot of folks together into a room and you spend days there and every day you like pump out. 10 ideas, , and at the end of the week you have like, oh, 30 ideas. Let's all take a look at them.

Lots of the teams and the companies that I've seen, they spend a huge amount of resources on these type of visioning exercise because it takes time to develop ideas, but the things don't have to necessarily to be finalized because it's essentially, it's a, it's a brainstorming exercise.

So I think that's kind of where a tool like this can shine. I'm like, well, if I have a quick idea, I'll just use this , to generate it because I'm not looking for a finalized version. I'm, just, communicating my ideas to my stakeholders and like I think people wanted to see the ideas being visualized.

I think with a tool like Aleo, you can have 20 ideas in three hours if you want. it's not like you're designing less, I think it's more like the things are going faster and now you can just, at the end of the day, hey, we have a lot of things being produced .

You can essentially go to your team, your product peers, your engineer peers. It's like, Hey, what what do we think about this? Like, which direction should we go deeper? I don't think , the overall process is necessarily changing that much, but I think that, that the velocity of the process is gonna be different.

Ridd: Yeah, I like that answer a lot. And I've been asking similar questions to other people and really just trying to understand like what are the key traits that you can identify across the best product teams? And so often it comes down to just being able to compress these. Iteration cycles to run more loops more quickly.

And I can see how a tool like this would help teams make that jump.

Helen: Yeah, absolutely.

Ridd: I've been kind of thinking about. Like where a tool like this fits into my personal design process, because it's not realistic, at least anytime soon for AI to be at the point where, you know, I just hit a button and the whole design is done. All of the thinking, all of the iterations, and it's pixel perfect, and all this kind of stuff.

[00:17:11] Where Galileo fits in the spectrum of "jobs to be done" for designers

Ridd: And so if you think about maybe the design process as a series of jobs to be done. What's the right entry point for a tool like this? And you've been talking a lot about ideation. Is that kind of where your mind is at? Because it's like, maybe you can hook into like a team's design system and you're actually building out production ready specs and there's a really wide spectrum of. cases in between. So can you talk a little bit about where you see Galileo fitting in that series of jobs to be done?

Helen: That's one of the things I'm very excited about doing this company and developing this product is that there's so many different directions you can go and there is the. The breadth and the depth. Right? And, and you could potentially go very wide in your generation and or you could potentially go very deep like what you said, connecting with your design systems and, doing super tight integration with the company's branding kit.

I, I think there's many, many different directions that we can go and based on. What I've seen, ideation now is kind of a very strong, , use case that we've been observing and essentially , is kind of like the, at the top of the funnel of the process that you described, right?

I don't think AI will replace any. of the human side of things, like thinking about ideas, like what is it that we're trying to build, right? Why are we building this? I don't think AI is here , to replace those thinking and , those creativity. But I think once you're kind of, I. Have a general direction, you have some ideas.

I think that's where a tool like Galileo comes in , is when I say top of the funnel, I mean top of the funnel of the start of the creation is like, wow, I can start use this to formalize my ideas and maybe to pump out a lot of ideas to get feedback. So. At least for now, that's kind of the layer that we wanted to focus on is that how do we help people to edit it fast?

How do we, to help people , to visualize what they're wanting, to build real really fast.

[00:17:11] Where Galileo fits in the spectrum of "jobs to be done" for designers

Ridd: And so if you think about maybe the design process as a series of jobs to be done. What's the right entry point for a tool like this? And you've been talking a lot about ideation. Is that kind of where your mind is at? Because it's like, maybe you can hook into like a team's design system and you're actually building out production ready specs and there's a really wide spectrum of. cases in between. So can you talk a little bit about where you see Galileo fitting in that series of jobs to be done?

Helen: That's one of the things I'm very excited about doing this company and developing this product is that there's so many different directions you can go and there is the. The breadth and the depth. Right? And, and you could potentially go very wide in your generation and or you could potentially go very deep like what you said, connecting with your design systems and, doing super tight integration with the company's branding kit.

I, I think there's many, many different directions that we can go and based on. What I've seen, ideation now is kind of a very strong, , use case that we've been observing and essentially , is kind of like the, at the top of the funnel of the process that you described, right?

I don't think AI will replace any. of the human side of things, like thinking about ideas, like what is it that we're trying to build, right? Why are we building this? I don't think AI is here , to replace those thinking and , those creativity. But I think once you're kind of, I. Have a general direction, you have some ideas.

I think that's where a tool like Galileo comes in , is when I say top of the funnel, I mean top of the funnel of the start of the creation is like, wow, I can start use this to formalize my ideas and maybe to pump out a lot of ideas to get feedback. So. At least for now, that's kind of the layer that we wanted to focus on is that how do we help people to edit it fast?

How do we, to help people , to visualize what they're wanting, to build real really fast.

[00:19:20] Does text to UI cap creativity?

Ridd: put on my design Twitter troll hat. Just one more time, I think there is a narrative around. Prompting that basically says it caps creativity because it's pulling from this constrained data set and maybe even reduces the incentive to create something truly novel when it's so incredibly easy to just type something, hit, enter, and pull from a set of patterns and templates that work.

Do you have a take on that or a response to that design Twitter troll? I.

Helen: Essentially the whole, like AI technology, the way it works is that it just read everything on the internet and it kind of gives you a version of what his has already seen Not necessarily very good at creating something that is super novel. Something that I hasn't seen before. and I think I've seen kind of the point of view, , on social media, which is, well, what about the uniqueness? What about something that is really novel? Like, is everything gonna look the same?

I think being able to push the boundaries , of innovation, I think that's where humans can shine, right? The people who have the real kind of.

Ambition and attention and creativity to create something super novel. I don't think , that's where AI is gonna shine. I think where AI is gonna shine is that, I'm thinking of an example that how I use chat GPT, which is,

I'm not here to write a Shakespeare novel. , I'm here to write an email. I just need it really quickly.

Like that's kind of the, the idea of it. I don't think A.I is here to necessarily.

Push the ceiling of , the design innovation, I would never be able to write a Shakespeare with ChatGPT but now I can write emails and marketing copy super fast. And I think to me that's enough. Like that's what I'm looking for.

[00:19:20] Does text to UI cap creativity?

Ridd: put on my design Twitter troll hat. Just one more time, I think there is a narrative around. Prompting that basically says it caps creativity because it's pulling from this constrained data set and maybe even reduces the incentive to create something truly novel when it's so incredibly easy to just type something, hit, enter, and pull from a set of patterns and templates that work.

Do you have a take on that or a response to that design Twitter troll? I.

Helen: Essentially the whole, like AI technology, the way it works is that it just read everything on the internet and it kind of gives you a version of what his has already seen Not necessarily very good at creating something that is super novel. Something that I hasn't seen before. and I think I've seen kind of the point of view, , on social media, which is, well, what about the uniqueness? What about something that is really novel? Like, is everything gonna look the same?

I think being able to push the boundaries , of innovation, I think that's where humans can shine, right? The people who have the real kind of.

Ambition and attention and creativity to create something super novel. I don't think , that's where AI is gonna shine. I think where AI is gonna shine is that, I'm thinking of an example that how I use chat GPT, which is,

I'm not here to write a Shakespeare novel. , I'm here to write an email. I just need it really quickly.

Like that's kind of the, the idea of it. I don't think A.I is here to necessarily.

Push the ceiling of , the design innovation, I would never be able to write a Shakespeare with ChatGPT but now I can write emails and marketing copy super fast. And I think to me that's enough. Like that's what I'm looking for.

[00:21:13] How designers can use "image to UI" technology

Ridd: That's a really good analogy. It makes sense and it's a nice way to reframe it because yeah, there's always gonna be people that strive to write Shakespeare, and that's fine most of us are not trying to write Shakespeare and this utility can really, really help. We've been talking a lot about the text to UI.

I'd like to hear you talk a little bit about the image to UI, which is a slightly newer feature and maybe even. Your vision for how you see this being used by designers moving forward.

Helen: Yeah, I'm glad that you mentioned that. That's a newer technology that we've been kind of developing in the recent month. And it came from the idea, that, you know, it's not always easy to prompt things.

Because a lot of us, especially designers and builders, we're kind of like visual thinkers. You're like, oh, that's kind of what I'm thinking. Now I have to translate it into words and describe it. And, for some people it takes a mental hurdle , to do that. And the idea of image d UI, it's like, well, can I just draw it on a napkin or, in a Google Slides, can I just draw some boxes?

And , you just take screenshot and give that to our AI and be like, , I'm thinking of designing a homepage of an e-commerce app. , this is kind of the overall structure I'm thinking, can you. Gave me something higher fidelity.

Ridd: I'm excited to see where that feature goes. 'cause something else that I, it feels almost like inevitable that I'm really excited about. The potential for AI in the long run is like being able to feed it not only just wireframes, but visual languages that I'm really by or specific types of, uh, UI or layouts.

And I could take a screenshot of something that I really, really like and say, okay, now. Here's my flow. Here's the vibe. Make the magic happen.

Helen: I think that is a really, . Interesting idea, which is, you know, kind of, there is many different aspects of the visual input that you can give, right? You could give, , content, you could give, architecture of a UI, or you can give a visual style of a UI. I, of, our MVP kind of took approach of like, okay, everything is one input.

You put it and then you go, I think. It's a really good experiment that we can grow in the future. It's like, well, what if you let people to input different aspect of things and see how that would actually translate to the end results? So I think that would be something that would be really exciting to experiment.

[00:21:13] How designers can use "image to UI" technology

Ridd: That's a really good analogy. It makes sense and it's a nice way to reframe it because yeah, there's always gonna be people that strive to write Shakespeare, and that's fine most of us are not trying to write Shakespeare and this utility can really, really help. We've been talking a lot about the text to UI.

I'd like to hear you talk a little bit about the image to UI, which is a slightly newer feature and maybe even. Your vision for how you see this being used by designers moving forward.

Helen: Yeah, I'm glad that you mentioned that. That's a newer technology that we've been kind of developing in the recent month. And it came from the idea, that, you know, it's not always easy to prompt things.

Because a lot of us, especially designers and builders, we're kind of like visual thinkers. You're like, oh, that's kind of what I'm thinking. Now I have to translate it into words and describe it. And, for some people it takes a mental hurdle , to do that. And the idea of image d UI, it's like, well, can I just draw it on a napkin or, in a Google Slides, can I just draw some boxes?

And , you just take screenshot and give that to our AI and be like, , I'm thinking of designing a homepage of an e-commerce app. , this is kind of the overall structure I'm thinking, can you. Gave me something higher fidelity.

Ridd: I'm excited to see where that feature goes. 'cause something else that I, it feels almost like inevitable that I'm really excited about. The potential for AI in the long run is like being able to feed it not only just wireframes, but visual languages that I'm really by or specific types of, uh, UI or layouts.

And I could take a screenshot of something that I really, really like and say, okay, now. Here's my flow. Here's the vibe. Make the magic happen.

Helen: I think that is a really, . Interesting idea, which is, you know, kind of, there is many different aspects of the visual input that you can give, right? You could give, , content, you could give, architecture of a UI, or you can give a visual style of a UI. I, of, our MVP kind of took approach of like, okay, everything is one input.

You put it and then you go, I think. It's a really good experiment that we can grow in the future. It's like, well, what if you let people to input different aspect of things and see how that would actually translate to the end results? So I think that would be something that would be really exciting to experiment.

[00:23:44] AI's adoption curve for designers

Ridd: analogy that I've been thinking about for much of the last like year even, which is basically like. Let's say I'm driving downtown and there's this plot of land where construction is happening and one day I drive past this little plot of land and all of a sudden a six-story scaffolding for this big building is just there and it, it just came out seemingly overnight.

And wow, I cannot believe how quickly. That was thrown up. But then you drive back, you know, the next month and the next month, and it's still the scaffolding. Nothing has really changed. And you're kind of like, man, like is there anything even really happening? And it actually takes quite a long time to get to the point where you have this ready to go.

Building, and I think for a while now we've kind of been in the scaffolding phase of ai, where it came outta left field. It almost like hit us in the face a little bit, where we're like, oh my gosh, look at this. What is possible? Let's think about all the different ways that this is gonna fundamentally change the role of design and the overarching industry for software.

And now we're like three, maybe six months into it and. I don't think there are that many designers who are actively using AI powered tools in their day-to-Day workflow. So this week and those a hundred thousand designs, it's a pretty big deal in terms of the overarching narrative and where we sit within this metaphorical building cycle.

So I'd be curious to hear from you, like where do you think we're at? What needs to happen for the majority of designers to be using AI tools basically on a daily basis?

Helen: Yeah, I like your analogy. I think it's a very smart one. I think the analogy that you give is not just describing this AI wave. I think you're essentially describing any technological wave, right? If you think about. the internet around 2000 and mobile in 2008. I think they all went through that period.

And people overly excited there so much investment and, and, and eyeballs and attention get onto the thing. And then there is a period of like, please, this gonna be a thing or it's gonna be a fad.

Are we Web-three-ing all over again?

Are we And, , , I think AI is very different than, I think the nature of it is, is a bit different than Web3.

I, I believe AI is here to stay like the internet and mobile. But it is gonna take the, the amount of time , that any technology I think it needs to take to. Kind of cross the Chasm, you know, you have the kind of, the, the innovators and then you have the early adopters, and then it takes a while really for the new technology starts to go mainstream.

I don't think there is like a one particular thing that needs to happen that for us to get there. It's not something that you we'll see overnight.

It's like, oh my God, tomorrow and now everything has changed. I think it's gonna be more like a gradual process, , of you're gonna see some, . Businesses, especially my hypothesis that it could start somewhere like a smaller businesses start to adopting these things because they're under the pressure of their competitor, right?

If your, your competitors are using AI to drive efficiency , to move faster. , you will have kind of the peer pressure to do the same thing as well. , and the same on the individual level. If you work in a company and you're seeing your peers are all using GitHub copilot , to kind of speed up their coding speed, and you are going to use that as well, but this adoption is gonna take some time.

But I think it's eventually it's gonna go mainstream.

[00:23:44] AI's adoption curve for designers

Ridd: analogy that I've been thinking about for much of the last like year even, which is basically like. Let's say I'm driving downtown and there's this plot of land where construction is happening and one day I drive past this little plot of land and all of a sudden a six-story scaffolding for this big building is just there and it, it just came out seemingly overnight.

And wow, I cannot believe how quickly. That was thrown up. But then you drive back, you know, the next month and the next month, and it's still the scaffolding. Nothing has really changed. And you're kind of like, man, like is there anything even really happening? And it actually takes quite a long time to get to the point where you have this ready to go.

Building, and I think for a while now we've kind of been in the scaffolding phase of ai, where it came outta left field. It almost like hit us in the face a little bit, where we're like, oh my gosh, look at this. What is possible? Let's think about all the different ways that this is gonna fundamentally change the role of design and the overarching industry for software.

And now we're like three, maybe six months into it and. I don't think there are that many designers who are actively using AI powered tools in their day-to-Day workflow. So this week and those a hundred thousand designs, it's a pretty big deal in terms of the overarching narrative and where we sit within this metaphorical building cycle.

So I'd be curious to hear from you, like where do you think we're at? What needs to happen for the majority of designers to be using AI tools basically on a daily basis?

Helen: Yeah, I like your analogy. I think it's a very smart one. I think the analogy that you give is not just describing this AI wave. I think you're essentially describing any technological wave, right? If you think about. the internet around 2000 and mobile in 2008. I think they all went through that period.

And people overly excited there so much investment and, and, and eyeballs and attention get onto the thing. And then there is a period of like, please, this gonna be a thing or it's gonna be a fad.

Are we Web-three-ing all over again?

Are we And, , , I think AI is very different than, I think the nature of it is, is a bit different than Web3.

I, I believe AI is here to stay like the internet and mobile. But it is gonna take the, the amount of time , that any technology I think it needs to take to. Kind of cross the Chasm, you know, you have the kind of, the, the innovators and then you have the early adopters, and then it takes a while really for the new technology starts to go mainstream.

I don't think there is like a one particular thing that needs to happen that for us to get there. It's not something that you we'll see overnight.

It's like, oh my God, tomorrow and now everything has changed. I think it's gonna be more like a gradual process, , of you're gonna see some, . Businesses, especially my hypothesis that it could start somewhere like a smaller businesses start to adopting these things because they're under the pressure of their competitor, right?

If your, your competitors are using AI to drive efficiency , to move faster. , you will have kind of the peer pressure to do the same thing as well. , and the same on the individual level. If you work in a company and you're seeing your peers are all using GitHub copilot , to kind of speed up their coding speed, and you are going to use that as well, but this adoption is gonna take some time.

But I think it's eventually it's gonna go mainstream.

Deep Dives

Get our weekly breakdowns

Insights + resources from top designers πŸ‘‡

Lauren LoPrete

Director of Design Systems @ Cash App

David Hoang

VP of Marketing and Design @ Replit

Adrien Griveau

Founding Designer @ Linear

James McDonald

Designer @ Clerk

Femke

Design Lead @ Gusto

Join 10K+ designers

HC

HC

HC

HC

"

I've been binging Dive Club lately and the quality is nuts

Literally the only show about design I watch”

Eugene Fedorenko

"

I've been binging Dive Club lately and the quality is nuts

Literally the only show about design I watch”

Eugene Fedorenko

hello@dive.club

β’Έ Dive 2024