Dylan Field, how does it feel? You just got off stage.
Uh, it's a relief obviously. Um, you know, you stress a lot about these keynotes and, uh, I just am really proud of the team. They did an amazing job and yeah, now the real work's ahead of getting all the feedback, responding to it, making our products awesome. I'm really excited to share this with the world.
The crowd was roaring. They were roaring, especially at motion. I think it was like, was it similar or less similar to the roars for hyperlinks at the first?
I don't remember the magnitude of hyperlinks. Hyperlinks were well received. Uh, but yeah, motion, I mean, as we play with it internally, we've been so psyched. Uh, and I think it's going to be creating so much good content.
So, what is your hottest take today about AI and design?
My hottest take, I have a lot of hot takes, but I'll only give you one. Um, I would say that it is not enough to wear a thinking cap. You have to actually think. And there's a world where a lot of people are getting into this space of they're letting the AI think for them and they're tricking themselves into believing that they're thinking themselves and especially when the AI is sycophantic and it's telling you you're so great and such a genius but then kind of giving you what it wants to do and you're like oh yeah that's kind of good. I'll follow that. That's no longer your idea. That's no longer your intention. And you get to an outcome and you look at it later and you're like, what is this? And so what you need to do is use AI as a tool to learn, a tool to understand the world better, a tool for leverage to get things that you know are not the critical stuff done faster. But don't give up thinking. Don't give up your brain. Don't join the hive mind.
Don't join the hive mind. So, are we going to need instead of recapture tests, we're going to need thinking tests?
I don't know how we do that, but we'll figure it out.
Okay, lastly, thank you again for being here.
Of course. I mean, I've now tasted, you know, the drug that is Config, and I think I'll be back next year.
Well, that we Config build you. Well, thank you, Dylan. Well, I'm very excited. Today, we have the king of robotics. We have Brian Ringley. Brian, welcome to Sourcery.
So, for people who don't know you, what is your role at Boston Dynamics?
Yeah, I am an interaction designer at Boston Dynamics. Um, I've actually also been a product manager. I've been in business development and I started as a customer. So, I've kind of done the whole gamut over the past seven or eight years. Yeah. I was one of their first customers.
Uh, for construction applications. I was working on the construction automation team at WeWork. Did you know they had one?
That's right. Yeah. He wasn't like right right there, but yeah.
What's your hottest take from working with Adam Newman?
Okay. He really doesn't he really didn't wear shoes. That was true. And he's he was so nice. He really is charming. Yeah.
Did you have to keep that in consideration when you're making construction robots for Adam Newman that they have to be light on their toes?
Like a specific like barefoot avoidance feature.
God forbid it steps on his toes.
That was like our pre- like ground clutter avoidance technology. So that actually could have been a problem for me. Yeah.
Okay. So if people who are not in your world want to know about robotics, like what should they be looking at today? What is the hottest trend?
I think that they should be looking at what's happening right now with your ability to have a conversation with software and think that it's, you know, we're not too many steps from you being able to do that with a machine and think about what it would mean to have a machine in your physical environment.
Yeah. And you're taking stage today. What are you going to be talking about?
I'm going to be talking about Atlas, our uh generally capable humanoid robot. So, the title of the talk is the product that can do anything because that's the design challenge when you create this general purpose hardware platform. How do you actually task that robot so that it can do anything? Because that's the point of creating that humanoid form is that you have that flexibility and you can just retask it to do whatever you need.
Humanoid robots are super hot right now. Uh-huh. If you didn't know that, you picked a good thing to work on. What is the biggest challenge with building humanoid robots and designing them?
I mean, I think it's your like traditional hardware product challenges. It's It's really not enough to have the world's best hardware, which I believe we do or to connect that to all the various agents and models and policies that actually give it the intelligence that it's required. It has to be reliable. Um, it has to be serviceable. It has to be cheap and affordable. So, I think all of those things are what will actually make for like a successful humanoid product in the long run over the kind of the more obvious and sexy design elements.
When do you think we're realistically going to start to see humanoids in the real world?
Oh, I I always hate this question. Um, but obviously as soon as possible.
It's like the AGI question.
Yeah. Yeah. Um, I mean, I think I think it'll happen faster than you think. I remember going through this with Spot, our quadruped robot. Um, where it was so exotic, like before I ever saw one in person, I just thought it was the most exotic thing. I didn't think it had any practical purposes. I thought it was like R&D. And it was really only like two or three years later that I was in, you know, the Bay Area, of course. And it was the first time I just saw one where I didn't know who was controlling it, right? It like wasn't one of it wasn't me or one of my colleagues. I just saw one walk by. And that was the moment like, oh, this is going to happen. It might take 10 years. It might take 20 years, but you know, this is going to happen. Not not because it's like some inevitability, but because it's possible to do and we'll figure it out.
Well, I guess of all use cases for humanoid robots, they've been promised they can do generally everything, right? What are you most excited to use a generalized humanoid robot for?
I mean, I I like the space that Boston Dynamics is playing in right now, which is automotive applications. We're working with Hyundai. We're like in their plants. Uh we're doing like automotive part sequencing and trying to work our way up to like the holy grail of general assembly. Um and I like building a robot that just like it just looks like a power tool. Like we're not trying to trick you into thinking it's a human. We're not trying to give it any kind of, you know, agency in that way. We just want you to look at it and be like, "Oh, that's that's a useful thing that belongs in a factory." And you kind of just understand that by the visual language of the design of the robot and the way that it behaves.
Should humanoid robots have faces or not?
This is part of my talk actually. Yeah. Well, because at first I was like, does it need a head? You know, does it need anything? Um, but you know, does it have a heart? Yeah. Does it need a soul? Um, but we, you know, at first we're like, okay, well, you want to elevate the perception system, like the cameras up to like roughly like average human eye height because you're doing human-purpose tasks in human-purposed environments. But then we're like, why the face and the directionality? Why the two degrees of freedom in the head? And the truth is, it's because we're putting these robots in the world with you. How do you communicate with other people? You look at my face. If the robot looks at you when you're approaching it, you have a sense that it recognizes your presence and it's going to act safely around you. If the robot looks at an automotive part, you get that predictability of its next movement. So, it's really about that human-robot interaction that it's necessary both to have the face and the kind of ability to look around naturally and gaze at things.
That's really exciting. I could literally ask you 500 questions about humanoid robots. And I'd love to take a tour. Whenever you want to give me one, I'll be happily Brett will come. We will happily come by. Yeah. But okay, so you were at Config last year. What was your biggest takeaway from last year?
I was I think it was two years ago that I came. Um and uh well, I was so ignorant of everything because I was like at Boston Dynamics starting to get into the interface design side of things. We were moving more into software. We need software to interface with our hardware. And I was like, I think that people use Figma for this. I've heard good things about Figma. And I was like, I've also heard of something called a design system. Like, so I just came to kind of look at what these other big companies were doing. Um, and to kind of understand the culture and the purpose and how it related to our business objectives. And now it's crazy. We have like the second version of our design system. You know, it's compatible with like AI agents so that we can use all of these new emerging breeds of software to design our interfaces. So really just in the last two years um like this culture of design and technology has exploded and become highly relevant for things like robotics. And I think part of what I'm excited about is like did you know that we also use Figma that we have the same challenges as you that we also need design thinking to solve these problems.
It's very exciting. So you're going on stage today. Yeah. You said the last time that you watched the keynote, they had a lot of Gen Z terms. So, are there any in particular that you're looking to include in your presentation today?
Oh, yeah. We're going to we're going to robot max today.
We're going to robot max today. Okay. As we close up, what is your hottest take in robotics right now?
Well, I think I think there are a couple things going on. I what I'm most excited about I think my hottest take is that there's a lot of conversation about how the AI brain will commoditize hardware. So whoever solves the AI layer of robotics, the hardware won't matter anymore. They're all just physical agents and we can deploy the same AI on everything. But I actually think that the opposite is true. Is that no matter what that intelligence is, you still need this physical body that's capable of interacting with the world, that can do it over and over again, that won't break down, um, and that can move in kind of new and and efficient ways. So, I think hardware is more important than ever by virtue of this software explosion.
It's a great way to end it. Hardware is more important than ever. Thank you, Brian.
Right now we have Ian Silver, head of product design at OpenAI. Ian, welcome to Sourcery.
We are at Config today. What are you most excited about?
Uh well, for one, just to see a bunch of familiar faces. It's such a cool opportunity that the design community comes together like this every year. Um, uh, I I did get a chance to like quickly look through, uh, some of the stuff that launched today and I just think that like it's amazing how quickly the tools are advancing like on a daily basis almost. Uh, but just seeing how the design tools are evolving. Um, it's just cool like what people are have access to and what they can make now.
So what does product design mean at OpenAI? What are you building?
Uh good question. So our team um kind of sits between like a bunch of different functions, right? So you have engineering that's actually like doing a lot of the hard building. Although now with some of the tools is designers are also building a lot too. But um and you have the researchers that are like making the models. um you have like product managers that help kind of run everything. Um and but but designers sit uh sort of at the center of that and help uh figure out like how do we kind of bring these novel new technologies and and put them together into products that actually make sense for people that solve problems that are easy to use that are delightful hopefully. um and kind of package it all up and and and think about like when you have these new model capabilities, how do we actually want people to interact with them?
What's the favorite product that you've worked on?
Uh that I've worked on. I mean, I've I've uh had a chance to work on a few a handful of really great products. Um, obviously recency bias with ChatGPT is kind of amazing. Like to work on something that uh has just touched so many people's lives that I get to use every day. It's cool and and it like um I don't know it sort of surprises me every day what you're able to do with it. Codex is a recent favorite of course too. It's like really changed the way I and our entire team has worked and and that's amazing. Not to shill our own products, but like they're both like so such fun to to work on.
Is it true that you invented ChatGPT?
Yes. Yeah. Put that put that in writing. No, I did not.
Okay. So, as ChatGPT has been developing, there's been a lot more integrations. What like how do you think about functionality when you're including them within the chat interface? And I know Codex is different.
What do you mean like functionality in what way?
Like how users are interacting with it.
Yeah. I mean it again like it's such a cool time to be designing because I think that um just the way you work with computers is changing. Um I I like it's funny. I remember this moment when I like just typed a bunch of uh like typos and all this thing into it. This is when I first started working on ChatGPT and and it just like understood me and it worked and I was like that's just not how computers worked up until like a couple years ago like the fact that they're non-deterministic and you can just talk to them. You know, we're working on lots of exciting like voice interaction type stuff and just changing the way you you build you interact with things. A lot of assumptions you can throw out the window and you can like start to reinvent things. And you know, we're starting to get into the world of generative UI, which a lot of people talk about, and I think that is like an exciting future, and we're just at at the beginning phases of that. And so, you know, there's just basically any point of how you interact with something, it's kind of just been flipped upside down and changed, and the way tools are built now are starting to shift. And um I think it's just super exciting time.
Did you expect people to fall in love with ChatGPT?
Well, I mean, uh, I mean, fall in love with it as a tool.
No, like fall in love with it. People fall in love with it. People are in love with it like they come to it all the time because it gives them emotional support and validation.
Abs. Yes. Absolutely. I love Chat as a tool. Um, and I think that but but yeah, I mean it's able to fill uh I think a lot of different roles for people and and and being a like a safe space where someone can talk to I think is is like kind of an interesting product for sure.
So taking a step back, one of the main headlines right now is that design is dead. I think it's very obvious we're in a conference center with 10,000 people. It's not dead. But why do you think that that is why like why do you think that's a headline and and why is that wrong?
Yeah, it's so interesting. Like I think that you know you see because you can just theoretically type into a box and and get something like maybe you know you could argue the role of design is changing or something like that. Um, I actually think that the role of design hasn't changed at all in some in many ways. Like um I think if you think of a designer as someone who's like drawing rectangles and and that sort of thing, then maybe maybe that's dead. I don't know. But like I think that like designers do a lot more than that, right? um or at least like great designers are are really thinking about and like understanding the psychology of how someone interacts with something and the the way people are going to approach and and and use your product and who you're building for. Um you know, it's like theoretically you could you could now in in one prompt make DoorDash. Uh but that doesn't mean you have a successful business, right? like there's still a lot of work that goes into understanding what problems you can go after and solve and how you do that. And then of course I think like the human touch I think is going to be more and more important than ever like making sure that uh or I think that products that have that like true care and human kind of touch will really stand out. Um as like more and more people can make more things. I think the the bar will be higher, which is great. And and I think that means that designers who uh typically have a very high bar, I think, will have a really great chance at kind of standing out and helping take a mediocre thing to great.
Historically, what have been the products that have inspired you most? And what do you think is maybe the best design product of all time?
Oh, best design product of all time. I mean, I'm I'm always inspired by hardware because I work on software. Yeah. Uh and you know, I've recently had a little bit of a window into that world. And um it's just so much crazier to work on something that has to be sort of made and can't be changed. Like with software, it's all malleable. And you know, the way we work now is very iterative. we like try something and we throw it out and we and we just like keep iterating. We get a lot of things wrong all the time, but then we continue to make it better with feedback. That's just so much harder to do with hardware. Um, you know, of course I I think there's just so many great pieces of hardware out there in the world. Of course, like, you know, early Apple products uh inspired me a ton. Um, and uh, you know, I think like they they did such a good job bringing together the entire package in a way that nobody else did. Um, and so yeah, that that's always been an early inspiration.
Do you think the new jalapeno chip is the right shape?
It's not really shaped I are there any chips that are shaped like that? It should have been maybe like triangular or something.
It should have been in the shape of a jalapeno. So, our jalapeno would have been maybe a little bit hard to cut, but it could be a jalapeno flavored chip.
Yes, I guess you're right.
Yeah, maybe. I don't know.
So, as we close out, I have to ask you, what is your hottest take right now in AI and design?
Okay. I think that if I were just starting in my career, I I Okay, I'll say that. I think it's an incredible time to be just starting out in your career. Um, and I want to just encourage everybody to just obsess over the changing landscape and tools. Don't be afraid of it, but embrace it. Um, technology has always changed very fast, but it's never quite changed, I think, like this. And I really think that the accessibility of these tools now is like mind-boggling what people can do. I mean, I like when I was learning, I had to like like pirate Photoshop and I had to like I did that, too. steal stock photography and like cobble together anything to to be able to to design. Um, and I just think it's gotten way more accessible and and and then to learn like I think the fact that you can, you know, learn to code, you can learn to um, you know, what whatever it is that you're excited about, I think that there's now these new tools. Um, and I think that just your output now can be so much uh, so much bigger and your ambitions can be uh, bigger. you know, the the the team that won the um the Apple Design Award this year uh was uh this app I think called Grug. Is that right? Anyways, it was like it's really funny app. It's kind of like caveman caveman aesthetic. Yeah. Um, but it was just two designers, no engineers with Codex. And I and they wrote this really inspiring kind of note about it where they were like, it wasn't like we just prompted Codex to do it, but it empowered us to like build our idea. Um, and so I think people that come with this like unique approach will be able to build so much. And I think that stands for whether you're senior and very seasoned or brand new in your career.
Amazing. Very inspiring place to end it. Thank you so much, Ian.
Yes. Thank you. Congrats and have fun at Config.
We have Nico here with us. Nico, welcome to Sourcery. We're at Config.
Yeah, very excited to be here.
So, you stole the show. You were pretty much the first one on stage after Dylan. What did you talk about?
I talked about code layers and how like code is material that needs to become more collaborative, needs to become more fun. The workflows around code are not fun right now.
no. I think they have their place, right? They're built for production for so you don't ship bugs into production, right? But it's not about like getting to one idea. It's about getting to like 600 ideas. Like Figma's in the business of getting the bad ideas out so you can find the great ones.
That's a good line. Have you used that before?
Um, okay. So, for people who don't know, what is your role at Figma?
Uh, I'm a product manager now. I was a product designer for seven years at Figma. Uh, and I switched into product management, but those lines are blurry for sure.
And you came here right after university. Where did you go to school?
I studied in Did you get that?
That's in Australia, right?
Uh, it's in Germany. It's in German. Do you want to try saying it?
There you go. There you go. Brett is dying laughing behind the camera right now.
I studied interaction design there. Yeah.
Well, that's exciting. So, what's the difference between German design and American design?
I actually do think that there was a lot more or like there's a lot of more service design, design as a practice, design as something where you like shape user journeys, not necessarily individual pixels basically. And I do think that like a lot of that like strategic service design part I've seen a little bit more in Europe um than than in in US and so there's design can be bigger.
We were talking to Holly earlier and she was saying I mean she lives in Germany now. She grew up in the US but she did say it's a little bit cold over there and she came she came over here and she was like wow everyone's so optimistic.
That that that's true. I I I do think that like like there's a lot more friendliness generally in the US, but also especially in California. Yeah.
So, what should everybody be looking out for other than code in Figma?
I my favorite one like the motion stuff is so cool. Uh but I think what I love the most like at the end of the config demo, you should definitely tune into the keynote again. At the end, there's one section where Pierce is talking about animating the shader effects and there's a little like there's a knob on the shader effect and it's like moving as the timeline progresses and that's so cool. There's this little thing where like the knob moves around because the timeline moves around and the timeline is just such a cool indicator because everything kind of like comes together in the timeline. Everything gets animated. That's sick.
It's really cool. I know Brett said he loved that, too.
And he's he's the one doing all the magic behind the show. But I haven't actually asked anybody uh this question yet, but how do you rehearse on stage? Because you all have really good stage presence.
We just try it out a couple times. So, I think you got to like write it and then you got to rehearse it and then you got to unrehearse it, right? Like you got to once you wrote it, you have a script and then you go on stage and you're like, "Ah, this doesn't this doesn't fully work together. I need to like I need to like unlearn and bring the jokes back." Like my first script was a lot more scripted. Mhm. And was like, "Where's the Nico? You got to get the Nico back. You got to get the jokes back." And so I tried to get get a couple jokes here and there.
I know. I was laughing my ass off the whole time. So if you haven't watched the keynote yet, you should definitely watch it. even just for entertainment sake. Okay. So as we step back, we think about broader macro code has become a big part of design and even expanding that outside of traditional designers and some other people who just really want to try making things. So what do you think is going to come next? Like what is the next evolution?
I think that as as we said earlier, I
I think that there is more stuff that is going to come back to a canvas because the canvas is just a natural way for people to explore, put things side by side, kind of like put all the ideas out there because the key thing is that everyone is going to get to the average a lot faster.
Yeah. And to get to something that stands out, you just need to be kind of like obnoxiously pushing on your idea, seeing how far you can take it, push it further than you could before. And that's what your responsibility as a designer, as an engineer, as a creator is going to become is to not be satisfied and add more things that make it you, that make it your unique take on something. And you got to do this by trying out over and over and over again. Like we had one thing that was really funny when I joined. Every designer showed their most embarrassing earliest work.
And I joined and I was working with Raasmos, who was like one of the earliest designers or maybe first designer at Spotify, and I was like, I'm afraid, right, and other people too, but then you see their first work and it's terrible.
Yeah. And you're like, wait, my first work is terrible too. And so I think there's so much of this where you just need repetition, you need to do it, you need to do it over and over and over again, and you need to analyze what has happened, how you have progressed an idea because the moment you can explain it properly, you're kind of done solving it. I've worked on problems for a year and then at the end I'm explaining it to someone new and after a sentence they're like, oh yeah, that makes sense. I'm like, I just, this took me a year to get here to this explanation. And so that is what I think we need to do more. We need to explore more, put things side by side on a canvas.
What kind of risks should people be taking?
I think they should just start trying out stuff that they don't think will work. Because if you in your head think something doesn't work, you might be wrong. You need to be able to see it and judge it for yourself. Evaluate this and then decide that it's wrong because your brain like plays tricks on you with how good an idea can be and it can be so much better once you feel it or if it's, you'll learn from that too.
Mhm. Okay. So, it's a hard question. What is the worst designed product of all time?
The worst designed product of all time. Oh, that's a good question. I think there's products in the enterprise space that I really struggle with.
Yeah. Like in the actually it's...
It could also be hardware. Like it could be like a car or a toaster.
No, I'm not good at cars. I really struggle. I don't know if it's the worst one, but I really struggle with the UI of, I think it's Workday, the HR management software. I don't know where to click and I don't use it enough to learn it properly.
It's a good one. Sorry, Workday.
Sorry. Sorry.
If they are our customer, we'll just remove that one. But it's okay. Okay. So, as we close out, what is your hottest take right now in AI and design?
I think that experts will work in clusters of representations, not files.
Thank you so much. Thank you so much.
So, we have Zach Liverman of the MIT lab here today. Zach.
What am I wearing? Some I don't know. Muji, J Crew nonsense.
You're the most dripped out guest that we've had today.
Wasn't expecting that. Yeah.
Well, you're bringing something from the MIT lab. So, what's going on at the MIT lab right now?
Okay. So, I help run a group there called the Future Sketches Group, and we're looking at kind of what are tools that artists and designers will be using in the future. What does it mean to be making art in 5 years, 10 years, 20 years? And the Media Lab has all different types of groups. There's about 25 groups there.
