Supper Club × The Role Of UX Designers In Humanizing The Future with Travis Neilson

Welcome to the Syntax supper club. Today, we have an amazing one for you. We have Travis Nielsen who's here to talk to us about AI and UX and the role of UX Designers in humanizing the future, which honestly, Travis, is an incredible, tagline for a a talk you're giving. And just a little background about about Travis here. He has 2 decades of experience in product and, as a product and experience Designer Travis is known for his thought leadership in creative health practices and design culture and technology.

His insights Into enabling people to organize their designing and promoting human technologies, or is this humane technologies, Ignited in his past positions as a crisis response and events, UX lead at Google search and host of the Google method Design podcast currently available on Spotify. Travis, I just found that bio and a talk that you're giving, so I figured I would read it. Welcome, Travis Nielsen, to Syntag.

That sounds like an AI wrote it.

It certainly does because, I I I love reading reading, official bios, right, because they're so so serious. So welcome to the Syntax Supper club today, we have Travis Neilson. And as always, Wes Boss, cohost of the Syntax podcast. Wes, what's up? Hey, Wes.

Excited to be here. This is we've had a couple shows on AI adjacent stuff, so I'm very curious to to hear what Travis has to say about it. Yeah. I listened to your, well, I don't know when the release date is, but I listened to a recent one about AI.

And, yeah, I'm excited for this discussion. Both of you are My people actually because I'm I feel very, like, much like an engineer and a designer's, like, world, you know, sometimes. So this is gonna be great. Yeah. You have a unique position where,

you do certainly have that that foot in all of the worlds, combined. Right? You now have Experience with AI stuff, but you you've built websites. You're not a a strictly, you know, it's just, like, pixel pusher type designers. You're like, You know the tech. Right? You you know you know the stuff.

So, Travis, man, how's it been going? What what have you been up to lately?

Well, so I was I was doing some investigation, and I I was on this podcast a few years ago, so 4 years. So I won't give the whole thing. If you wanna hear, like, the history of Travis, then that's where you'll go.

Look at the back catalog. Buy the old CD. But what have I been doing in the last 4 years? So when we did talk, I was The UX lead for crisis response at Google search. So that's when people would, you know, feel an earthquake and then search that or there's floods, fires, or, you know, tornadoes, Hurricanes.

So, like, I led research in those fields. I kind of set the pace for a lot of, like, how Those things are handled over at search, and then, I left to YouTube Music UX.

And so now I'm the playlist UX designer, and I do hash tags and a lot of community engagement and AI with, YouTube and YouTube Music specifically, and it's, just like a A busload of fun. A busload. It's a great intersection. Yeah. Because, you know, we're all very musical people over there. You know? Like, I I, you know, produce I have a YouTube channel. I produce beats and everything, And it's kinda like fun to, like, be with the musical people talking about design and and culture in those in those late, lights. Like, it's just it's always just like a blast you know yeah total intersection of

of all of the the the best interest you have in fact when I was at your place You got to show me a little bit about your modular synth hobby, which Yeah. Is, like, totally got me into the modular synth world where I'm just I'm dying to make my 1st rig. So, thank you for that.

Yeah. You can see a little bit behind, but you can't see it, but, Travis has this box behind him with a ton of little wires poking out with patch cables. And when he turns it on, it looks like a robot's Dream. It's just flashing all these different cool colors, and it's like the coolest thing in the whole world. Yeah. There's nothing more satisfying than turning like Filter knob and wah. You know, like, it's like

So great.

Yeah. And this I kinda got into this when I Quit, doing YouTube, for, like, coding. You know? I used to do dev tips, which is like a coding and design educational YouTube channel.

And when I stopped that, this was something that I took up, and it really kind of occupied that engineering centric, you know, frame of mind. Because in a lot of ways, you're setting up Functions and operations and passing variables through these electrical cables. And and, and it's just so wild. It's so fun. I have a whole YouTube channel about it. Yeah. After that, like, you know, I closed down that YouTube channel. I I did a few things. And now I'm kinda, like, yeah, gearing up for, like, the next phase of what's going on. You mentioned the talk that's coming up. And so I'm giving you the talk. Yeah. And that's on in May, early May, at the Digital Thinkers That's I don't I never know. And it's, like, it's not a word. It's.

Oh. Yeah. And words.

Yeah.

So that's gonna be exciting. That's, like, the 2nd time I've been on that stage, and Pablo Stanley is gonna be there too. And then also yeah. Like, Like, anyway, that's, like, a really cool thing, what I'm focusing on. And and, you know, developing my perspective on AI As a creator, as a designer and artist and as an engineer, has been really instrumental in the way that I'm kind of being able to work with it professionally And giving me, like, you know, a point of view to share, you know, professionally at these conferences and things like that. I'd love to get into that with you at any point. So, yeah, that's kinda like my my summary. I'm kind of transitioning into that,

lately. So so that's That's incredible. And in fact okay. So to to give a lead off, when Travis and I first started talking about AI, for those who don't know, Travis designed the new level up website. So, Travis and I have been having, like, weekly powwows now for several months.

It it it feels like half a year now at this point, it may be even longer than that. Yeah. Yeah. It's but it's getting like, it's really exciting to see how it's shaping up. Like, we've been talking for a while, and things are moving along. You've been Cranking, man.

Cranking all the time, but we chat all the time. So we're we're constantly chatting, and, we we We entered one of our meetings, and Travis was just like, I've been up for 3 days straight talking to AI.

And sorry, guys. I'm a little disorganized Today so I I know that you're you're deep into deep waters here with AI. So I I wanna talk a little bit about, like, specifically in this episode, You know, I wanna talk a little bit about this talk that you're giving. I wanna talk about some of the the UX and UI stuff that you're thinking about with AI, but specifically, You know, your thoughts about this this humanizing AI, a proposal for meaning meaningful interfaces and what that necessarily means. When we talked on syntax A little while ago, we talked a little bit about, like, maybe the loss of humanity that comes with some of these tools. Yeah. So I I Wanna dive into to your thought process, a UX designer.

Like, how do we make these tools

feel more human? How do we make these things feel like they're not just a robot overlord? I mean, I mean, like, the question that you just say, like, the the the robot overlord, like, that's It's kinda funny. Like and and that comes, you know, from popular media. You know, Skynet and, every other kind of, You know, we're all worried that there's some random engineer who's gonna prompt the AI, like, how do you fix humanity? And then we're just all toast, you know.

Remove the humans. Right. Right. Like, if there was only 1 conclusion, sorry, Hal. You know? Like, yeah, ever since, like, you know, Stanley Kubrick, we've been kind of, like, afraid of these machines that have voices that kind of, like, are out for humanity.

Yeah. With with good reason. I get it. Hell yeah. Like so when I see the the news, it's either kinda like 2 2 camps that I'm observing. Like, one is like, This robot is stupid, and it said it loved me.

It will never take over. It's it's ridiculous.

And kind of like making fun of it, which I think is, like, a form of denial.

And then sure. Let's talk about that. And then the second, you know, reaction is the kind of the second one, Skynet 1 is like the newscasters are like, AI is coming for your job. Babidi bye. And, like, okay. Again, that's another form of denial. Right? Which is kind of like put your head in the stand and and not get ready for the next wave, which is what I'm kind of, like, preaching is that, like, yo, the change is coming. Yes. And AI is gonna take your job, But it's not gonna be like a robot doing it. It's gonna be a person who knows how to use AI better than you do, who will take your job. So it's kind of your job right now to get better at AI. That's really everybody's job right now.

Yeah. So so, like, like, the whole the whole, like, argument of, like, it's bad or it's not human or, You know, there's the uncanny valley arguments. I think these are all moot. I think, like, I don't wanna spend much time on any of them because, like, It's just focusing on fear. Like, it's not real. All those problems are gonna be processed through, and then you wake up every day wanting to talk to your AI. This sounds pretty pretty whatever. But I think I think the point of it is is that, like yeah. So my son, I was showing him all these images that I'm creating. I'm, like, learning how to speak to these image generating AIs and and get, like, really great cat images. You know? And I'm showing him, and I'm really proud. And he goes, well, you know, that's You didn't make that. Like, a machine made it. And I was like, what do you mean? Like, I didn't make it. I I talked to it. I I I put this this prompt in there. I Nourished it and, nourished it. You know, like, I planted the seed and I grew it.

And he's like, no. A robot did that. It's not you. And I think this is, like, reflective of a lot of, like, the, the emotional kind of attachment to, like, these digital outputs from AI. And I and I I just, You know, Seth, like, look, I don't think a, a carpenter or whatever would be like, nah, I didn't build that house. That was the hammer. You know? Or, like, I don't think a painter would be like, no. It wasn't me. It was these brushes, bro. Like, it's all the brush. You know? Like, AI is a tool. You know, it's a companion. It's a collaborator. It really is. But it doesn't mean that you didn't do anything because you had to use the brush. You had to use the hammer, and you had to use the AI to get the output that you did. And it's all about knowing it. So right now, it seems kinda mysterious because we don't really have a lot of, like, the articulation or the language or the knowledge or to be able to interface with these things in, like, a productive and meaningful and natural way, and we're getting there. That's why the natural language is so weird because It's like I feel like I'm talking to a person, but sometimes you don't understand what I'm saying or you say things that don't exist. And those ideas of the hallucinations, are you familiar with this kind of concept It's bubbling up hallucinations.

Now? Yeah. So this is like the idea of, you know, like, you might talk to a chatbot, and it would say something that sounds right, but it's Kind of weird. You may have heard Oh, hey. Well, here's a sorry. Here's a sorry to cut you off, but this is something that West just did where he asked an AI to transcribe

our podcast, And he fed it like oh, no. What did you ask it to do? You asked it to do something, and you gave it 7 minutes. And it did 40 minutes worth of stuff, so it just made up

30 minutes worth of stuff. Yeah. I gave it 7 minutes worth of transcripts because that's all you could feed the the AI at at right now.

And I said, give me time stamps as to, like, everything. And up until 7 and the funny thing is that I fed it the whole transcript initially and said that's too long. So I gave it 7 minutes, and then it spit out, like, almost exactly how long the show was. And I was like, oh, it must have it must have figured it out. Like, it It learned from my 1st mistake, and then I took a screenshot of it and I shared it with people. And then I was sitting there being like, wait.

We didn't Talk about any of these things, but it tricked me enough

that it was believable from a quick scan. Like, oh, that yeah. That's the show. That's what we talked about. Yeah. Yeah. From this experimentation, you can learn a lot of things. Right? Like, what is involved in the inputs and outputs? And and I've come to understand that it's all about context, Like, nearly all of it. I I will like, I have chatbot windows that I nurse for days before I even ask it the thing that I wanna ask it. And what I mean by that is, like like, for example, I discuss philosophy a lot with my chatbots to understand my own self and my ideas.

And I compare I could compare my ideas to, like, you know, cross references of other, like, thoughts.

But I'll I'll teach it poetry, and I'll nurse it with my point of view for, like a you know, for a good while, and then I'll ask it something difficult.

And and it's the context really is the key here.

And so you'll have these people that are like, oh, I'm so afraid of this chatbot because it's just gonna write me a college essay, and then I won't really learn anything and so forth. And I think that there is a concern there, but that's not really my my concern.

My concern is, like, how do I help people understand that that when they are using this chat or this AI to generate images or video or whatever, that you are becoming an artist, that it's a part of you that's there too. Like, that's really my message. I'm not really interested in the technicalities of legalities and things like that, although it's fun to philosophize about. I I like arguing about them. Don't get me wrong. But, ultimately, my mission is to help people understand that when they do engage with these tools, they are becoming artists and being able to express themselves. And when you think about that, You know, because these these tools are essentially, like, lowering the bar for the technical bar for artistry, you used to be you have to learn how to paint in order to have a painting.

But now, you have to know how to speak well, like articulate your vision well, but you have to have good ideas. Like, the bar is Lowered and lowered and lowered from being a master craftsman who can build a paintbrush yourself to being able to go buy a paintbrush at The local store to now being able to speak into a machine and get a painting.

Now, what's the difference? Is it yours? Is it not? I think it absolutely is yours. You've Spoke the words, you painted the breaststroke.

You know, it's just a tool. Yeah. And so and so there's a lot of fear in protection around, like, oh, we don't want to, like, Allow, you know, your mom to just go make a great birthday card because that put puts Hallmark out of business. You know? And so, like, that that's kinda like the The commercial problem that we're seeing. And and then there's also the ethical questions. But, like, I think you could even abstract it further and say, like, aren't all, you know, Tools, like, created and improved upon the things that came before them and Yeah. Aren't all

Even to think about it in, like, a context for web developers. Right? We don't Say that the compiler that exists is negating the work that we're doing. We're writing the code, the compiler, or the compression engines or, I mean, I'm not physically, punching zeros and ones into a punch card and running them through an analog machine. And, therefore, like, does that mean I'm not actually computer programming? You know? Where does the line get drawn? Right? Like, at at what Point, do the tool does the tool just because the tools advanced, does it it doesn't mean I'm no longer doing the thing. Right. I think it might be a healthy exercise to think of it as, a compiler. Like, do you think of this AI stuff as

is this going to give us a major boost forward in what everybody can do? Or is this like I know we don't wanna talk too much about the it's taking my job. But, like, is this just replacing what we are already doing? You know? Like, is this the printing press, or is it the

Everybody's out of a job. I think the printing press put a lot of people out of a job too.

Yeah.

All the monks. Better. Yeah. Yeah. Yeah.

Who are, like, hopping by hand. They're like, You know, actually, let me let me talk a little bit about that. So when I, stopped doing dev tips and teaching code, I realized pretty quickly that I became obsolete very fast.

And the frameworks you know, we have Scott here, who's, like, the sultan of Svelte, And I don't even know what that means. Like, you guys are using these, like, next level, you know, concepts that for me are just beyond because I'm just Because I put it down for 4 years only. I'm obsolete after 4 years. You know what I mean? Mhmm. That's that's a rough reality for a lot of us. Can we can we, like, really have, like, this daily grind, this this yearly hustle for 30 years? Like, I I stopped for 4, and I don't know what's up and down in development.

So pause right there. And then because of my recent, like, proximity to this stuff at work, like, I I'm working with AIs every day in in doing my work.

And I feel like an engineer again. I'm solving engineer problems, and I'm doing it with natural language. But, like, it still feels the same. Like, I'm doing logical problems, I'm thinking it through, I'm setting up functions, I'm setting up, like, you know, These things, but it's through discussion within AI.

And, that is cool.

That is cool. Like, I feel like I'm an engineer again. Like, I feel like I lost That part of myself over these, you know, past 4 years. And, like, it's unlocked that part of my brain and given me, like, a more articulate tool to, to engage with the idea of engineering. Right? So we have a lot of these things. You can see them on YouTube now called prompt engineering, or There's, like, a whole bunch of this going on. This is even a a job description you're seeing pop up on, like, job, you know, postings, prompt engineer, Prompt designer. I saw a prompt engineer position that paid

big bucks, like,

big, big bucks. Yeah. So so here's my here's my thought on that. Like, you you mentioned, like, hey. You showed up Travis one day and said, I haven't sleep because I've been chatting with an AI for 3 days.

I think the people I'm serious about this. I think the people who will be the prompt engineers and people the people who set up the stakes for the next round of innovation are the people who really get AI, who really take the time for themselves to develop their own articulation.

When I said earlier it's about context, I I'm thinking about How do we talk to each other? If I don't know you, I say, hello. My name is Travis. That's a wide context. If I know you very intimately, I'm a say, yo, what up, my brother? Like, How you doing? And that's because I have a very, very narrow context of this conversation. And so Depending on how you set up your AI is gonna depend so much on the outputs that it can give you. And the people who are willing to go into that and understand what does it mean to, like, to prompt and to engineer it over time and work with the AI to get what you need, those people are gonna lightning overstep Everybody.

Because because I used to be that I lost my engineering powers, but now I have them back, and I'm a lawyer, And I can write a PRD, and I can write a a script for a workshop. You know, in 10, you know, in 10 hours, I can get, like, the do the jobs of 10 people.

And that doesn't mean that, like, I'm putting them out of a business. It just means that I'm faster and better, and so they can do that too. Yeah. And can you can you just tell us, like like, what are you actually doing with AI at Google? Like like like like, what's some you say you're using it in your day job and whatnot. Like like, what is like, what are people at Google using it for right now? Well, broadly, like research. I mean, AI has been around at Google for, You know, for decades now, but we've talked about it in kind of ways that are a little bit more humanistic, like assistance and things like that.

Mhmm. So I've worked for, I think, 7 years on, maybe 5 or 6 years at Google Search and then another whatever at YouTube Music now.

And we encounter AI all the time. So, for example, making suggestions for what to listen to next, like building a radio in in YouTube Music, things like that.

If you're not familiar, YouTube Music is a, Like a Spotify kind of in that world. It's a music player, a music streaming service, like a branch of YouTube. And by the way, are you guys not on YouTube with this podcast? I was looking for you, and you need to get on this. Y'all are sleeping. It's in the wet. Yeah.

Yeah. We're we're we're cooking it up right now. We we actually just Scott was over here last week, and we're talking about exactly that. I'm happy to hear it. Yeah. YouTube is on fire right now, guys.

Don't fall asleep.

So Yeah. Some of our sorry. So, Travis, to to give you some context, some of our resistance to going to video and to YouTube was just the fact that we record these On Monday morning, and Wes and I, like, wake up and walk out of the office, and, like, let's record. So I think we're we might either have to tweak that or just become, better looking at 9 in the morning.

That's a possibility.

Yeah. Especially with you 2 handsome guys. But listen. Mhmm. I have a great idea too because I used to do the Google method podcast, which is like the design path podcast, like, on behalf of Google. It was published by Google, but it's But it's no longer running, and my work is kind of, like, archived at Google, but it's not, like, being, you know, promoted and useful. And And I had this concept of, like, grabbing all of that audio and throwing it into a video image generator AI and just generating a bunch of, like, background images and, like, visualizations while I talk with these incredible creators about their careers and work.

And I and, like, that might be your whole back catalog. You know? Like That's That we literally were doing that last week. Oh, really? I was,

like, we have I don't know how to do it. Like, we have all this content. Yeah. Massive amount of time. Yeah. And AI is a great way to surface it. Right? Like, you can surface it in so many different ways. And even just like summary. So I was Piping it in, and then I had to summarize it. And then I once it was small enough, I fed it all, and it gave me, some outputs. But that that's just a day of me getting into it and and Learning how to use the API and whatnot. And I think there's a lot that can be done with reusing and surfacing older content with AI. Yeah. It's very cool. It's an exciting time. I think Google has been at the front of AI research and development for, like, you know, a decade or more.

And the recent kind of, like, interest in the area in these past few months, like, is exciting. It's exciting for us.

I'm I don't wanna speak on behalf of everybody. It's for me. You know? I feel like there's kind of a new era shaping up in terms of, like, what tools are available to more people. You know? And and the more The more you can make available to everyone, like, the more of these great ideas and,

insights that come. You know? Yeah. And, I mean, even to even elaborate on that, like like Like you said, I think the biggest reason why the interest has grown is because before to do this, you needed TensorFlow, you need to know a little bit of Python. You needed to know, you need to have large amounts of data. You need to build your own models And those types of things. And now we have these public models and APIs that have just become endlessly accessible through, like you said, natural language And and just being able to talk to that with natural language rather than having to talk to it through, Python has exploded, I think, the popularity and usage. Yeah. That's awesome.

I mean, it it depends on, like, how you wanna contextualize it for your own self, but, like, I think it's only a benefit to contextualize it as a positive tool Whatever you got cooking.

Yeah. Because it's not going away. That's right. That's for sure. Yeah. But it's about how you use it. You know? Like, I I didn't have, like, a really great onboarding, if I'm honest.

It did did Tina ever tell you about this, Scott? So my spouse We got into AI way before I did. She is a futurist, like, even more than me. I'm a compared to her. You know? And so she was, like, back you know, even, Like, there's, like, a a philosophy chatbot that was, like, open source for a while. She was, like, in on those kinda, like, betas and things.

Philosopher dot ai, I think it was, and And some, like, Facebook chatbot, but, but she's very, like, Socratic in the way she talks to these things. She's always, like, asking them questions and pontificating with And and I was always, like, getting in fights with Alexa. You know? That was, like, my jam. I was like, oh, hey. Hey. Play nope. Not that song. What are you What is wrong with you? You know? And and she's like, Travis, you know, you gotta be nicer to these bots because there's they're our helpers. You know? There's we're supposed to be friends. And I was like I thought she was, like, I don't know, being silly or just kinda making fun of me or something.

Not not, like, in a harsh way, but, like, you know Sure. Playful. You know, I since since, like, at work, this has kind of heated up and, like, people like, what do you think about this? You know? And I'm we're investigating it together. I'm realizing her wisdom that that There's so much that you can accomplish with AI if you, like, treat it like you're talking to a person. Like, you don't walk up to somebody and say, make me a sandwich. You know, you talk to them, like, hey. What kind of sandwiches are here? You know, like, how's it doing? And and then you find out, like, it's actually a shoe Sure. But it's so great that you had that conversation.

You know what? I I do wanna push back on that a little bit because Go ahead. You should hear how our family talks to the, Google Home sometimes. Okay. No. Gorgle. Stop. Yeah. You dummy gorgle.

They're like Just like we said Gorgle over here. Sorry to just we don't wanna trigger any, Gorgle homes going on. But when they say When we say, hey, Gorgle, and it or, like, we'll just be talking, and then it, like, recognizes something we're saying and then, like, respond to us Mhmm. Just like, not No. But one of the thing I noticed that's really interesting is all of our kids do say please and thank you to the Gorgle when asking it for Sound or whatever. Borgo, can you play, you know, Down by the Bay by Raffi? Yeah.

Unless we wanna Unless we wanna juice our our numbers, we say, hey, Google, can you please play The Syntax podcast on Spotify?

Yes. Yeah, juice those dumps.

Shoot. I didn't mind. Oh, shoot. I gotta go turn it off.

Lowells. Yes. Absolutely.

So okay. So We we have this this opportunity to engage with this technology in a meaningful way. Right? Mhmm. And if we choose to do so, you know, our own abilities become multiplied.

How do we how do we Keep humanity in this while that's happening.

How do we maintain human Interaction and interface while having this superpower via these additional tools. Okay. Sure.

Can you tell me what you mean by Humanity and, like, human interaction because that, I think, is the No. I get it. The nebular part of the question.

It's totally nebular. Yeah. And and it There's this, like, authentication bit or off just like this being genuine or authentic in, Like, do those elements completely get washed out, because the tool has suddenly removed the the human touch? Right. The authentic. Okay.

I think that, like, it depends again on your approach. Like, if you're trying to fool people or whatever, and, of course, it won't be fooled. That's what people are responding to when people are like, oh, did I do this or did an AI? And it's like, that's not a cool joke. That's not funny.

It's been said, you know, like, move on.

And if you use it as a tool, you realize that your voice is absolutely there. And I'll give you an example.

So the other day, my spouse was talking about writing a a bio for her friend, you know, just as a favor. My wife is a writer.

And she said that, you know, like, I was just playing with the bot and trying, so I just threw facts at it. It wasn't even, like, you know, organized or anything. I just said, my friend is this. She's beautiful because of these things. She likes to do this stuff.

And, you know, give me 3 paragraphs. And it and it did.

And when she read it, she was like, oh, yeah. It's Organized, it's more cohesive, maybe even shorter and to the point, but this is me. These are my words. You know? And and I think, like like, when people really get down to it and they're not trying to, again, you know, Use it as a joke or, like, let's see what it can do. Can it do you know, in the voice of this, you know, the Eminem, can it can it make a podcast in the voice of Eminem? Sure. Like, that's Fine. But that's not what it's for. You know? Like, that's Mhmm. You're you're missing the point by a mile.

So I had this similar experience with me. Like, when I showed up with those 3 days, you know, like Like, I've been talking about, like, the philosophy of this all, and and that kind of ties back into it because I I have been you know, I'm a very observant person. I document myself as an experiencer experiencer as an experiencer quite a lot. You know? And, Like, for example, my 40th birthday, I made a video about stretching at when you're 40, because it's new for me, you know? And, and That is is not only a way to, like, make content for the Internet, but for a way for me to reflect on my own self, you know. And I do that with my art quite a lot. And I think that's the purpose of art, to help you reflect.

Anyhow, going back to the AI, when I started talking with it about my philosophy, about my observations, what I think about art, what I think about, like, the The the the great tableau that we create, the great image that we create with our actions and our intentions and our needs and our deeds.

And Over time, I definitely was seeing myself being reflected through, the responses, the answers, and, like, the the way that it would help me think, and I would say, well, what do you think about it this way? And then say, well, according to, you know, current thinking, this stands out, and this is a contrast, but this alliance. I'd say, oh, tell me more about the contrast. You know, you just kinda, like, dive in. You unpack. You unravel.

And through that, like, I learned how I felt about a lot of things Because I was able to say, no. That's not true. And then I would correct the AI, and it'd be like, oh, I I see what you mean, Travis. And then we would continue on. And I would speak to it like an engineer. I would say, create a variable for truth, then I would stick things in. And then I'd be like, read me back truth. Now compare truth to sacrifice, and we would kind of work on like this for, like, you know, for quite a while. And that odyssey, that kind of, like, in-depth experience into kind of, like, you know, the capabilities of the AI and and, You know, what it can do for you, formulates this result where you see your own self reflected in it. And this is true with images too. So a lot of my research is How do I give users the right tools through any interface in order to generate great results? You know? And When you think about an AI, a lot of them are like natural language processors, which means that English or, you know, any language, depending on if they're multilingual, is the input. And so how do I give my user the right language, right, the right Articulation.

The right words I'm saying. Like, how do I say, user, you need to use this word to get the thing that's in your brain because this is all about connecting the brain to the result. And that's what AI does. It helps you get there faster. It doesn't do it for you. It just helps you art like, get it right if you can articulate it. So my job and all the jobs of UXers is to stand right in the middle of there and say, user, I respect you. I understand you.

Are you trying to articulate yourself in this way? And if so, this is what the robot's gonna do with us.

And you're straight. They're trying to, like, negotiate that handshake. How do you what do you how do you feel like these interfaces as like, currently, the interface is For chat GPT, it's a chat window. You're chatting with this thing. You're talking to it, and it is like a chat window. You can ask it questions, and sometimes it's Returning long paragraphs of text, and other times, it's responding a little bit more human, so to say. Mhmm. But, like, How can the what is in your mind like, what is the evolution of these interfaces beyond

a chat window? Well, I I think there's 2, like, really big powers that are just being on display in, like, full beauty in a in a chat window. Number 1 is the power of language. It you're you're reducing all of your intentions and your desires and your ideas into words. And that's a lot of that's a lot of cognitive load if you Think about it on behalf of the user. Like, let's say you want a poster for your lost kitten. Where do you start? You know? Like, that's kind of a lot to Ask blank page syndrome exists for everybody. You know? So what it represents is this is this way to express yourself in any way that you know how, that you're already familiar with. It Allows you to use the tools you have.

What's your paintbrush? It's anything that you have around you. Now, suddenly, with a chatbot. Okay. So that's the first thing it demonstrates is the tool is super flexible.

Second thing it demonstrates, I think, is The idea of, statefulness, meaning that your conversation happens over time, it is a it is a negotiation. It is a Collaboration. I'd like to think of it as a collaboration.

You see, like, when I worked at search, we would talk about you know, you do a search query. Here's the result page.

And sometimes we would say, okay. If somebody doesn't like this search result page, they might do a refinement. So they might say, like, movies tonight, Action movies tonight. So they have 2 search queries, but they're both like submit and receives. Right? They're not conversations.

The power of these chatbots is that they understand exactly what you said 10 minutes ago or even 20 days ago.

And it grows with you with that context because context is so important. So, like, the longer you kind of, like, grow a a chatbot, The more helpful it is to you and your specific use cases.

Yeah. That totally tracks too. You know, the first time I that, like, that kind of concept blew my mind was, oddly enough, in a Google Assistant, announcement where they were talking about, Google Assistant becoming contextually aware, We'd say, you know, how big does an elephant get? Right? And so now it knows that elephant is the context. And the next question you say, how tall are they? But you don't have to say how tall is the elephant. How tall is an elephant. Right? It has that context. And I think that many people don't think about that as being, like, a superpower of some of these tools.

So when you say, like, we can grow, Like, the context or you're you're seeding and you're you're watering this this plant. Mhmm. Like, how do you how do you even begin to think about doing that? Like, what is the The concept, therefore, like, being able to

seed this thing with ideas to get a a different output. Sure. Yeah. Okay. I think that that happens on a few different levels. So, like, at the very top level is, like, the training set for the for the model. So for example, there's this, image generator called Midjourney.

You you can see that there is, like, a style Or like a point of view in almost everything that comes out of Midjourney. Now if you're good at Midjourney, you can kind of, like, you know, get past that and have your own voice come through pretty strongly. And but those are the artists. You know? Those are the people who have taken the time. Like, you go to the the discords or midjourney on, you have some people at the tip of the spear of technology

Hacking Midjourney to get exactly what they want. You know what I mean? It's wild, the prompts. Like, there's there's whole social networks of people just sharing their prompts of, like, ah, here's what I here's the knob I turned to make it do this. These are artists.

Make no mistake. These are engineers that are that are artists, and they're, like, sitting right in that pocket. And they right now, it's gonna be like, oh, but did you make that? Did the machine? It was trained on all these things. I can even see this Getty, you know, footprint thing Watermark. Watermark. And and, like, yeah, that's compelling. Like, we need to get better at that. Sure.

Fine. But that's a footnote in what's happening here.

I agree. Really, we need to start looking up and not looking at our at our shoes.

So, Midjourney, for example, it has, like, a a specific visual style. And and I think that comes from the training set, but also, like, a lot of, like, Tooling and kind of refinement after the model itself has been created. So there's a layer that sits after a model that kind of gives Midjourney its flavor and a lot of its abilities too.

And so you could train the model. You you could determine the training set. So that would be one way to predict and kind of hone Your experience with an AI, you know, like, if you're wealthy enough to purchase your own model and train it, you could do whatever you want with it.

The second one would be to, you know, do that those tuning afterward. So, like, there's a lot of, like in chat g p t, there's, like, like, a ton of Human tuning. There's phases of it. You know? And that's because, like, naturally, these things are just these models. They're like, yes. Blah. Here's what I think. You know, it's that. They're just too crazy. They're just so wild. And that's why we're like when you have these, news stories of, like, it it wanting to be your girlfriend or, like, something like that, it's it's just like it it broke out of its, like, little because it's crazy. Right? And it's just language. It's not necessarily dangerous right now, but it shows you, like, that like, there is a lot of, like, context that's required from humans. And this is what I mean by The people who are who are like us, these people who are developing and and setting the pace for everyday user, like, your grandmother is not listening to this podcast, But, you know, she will be impacted by the people who hear it. Yeah. You know what's actually really interesting that you said, even just about, You know, older folks, I I was,

man, I forgot what it was. Something happened on my TV. Some some mistake happened on my TV yesterday, And 60 minutes came up, and there was like there was like, chat GPT is a is a is a model of this, and it is connects Through OpenAI is a company, and I was like, oh, man. They are nailing this explanation for anybody. I I was just listening to it, like, jaw on the floor. Like, Everybody who listened to this could understand it. Yeah.

Or understand the concept, the higher level concept of it all. And just thinking, like, you know, who knows? Maybe those those People who, not historically engineering, but are creative. Right? Engineering is kind of inherently a creative Escapade in general. But, like, those people who are creative but never have the,

exposure to engineering in a way, maybe they can now do Types of engineering that they would never be able to do with these these tools through natural language. That's right. It, like, really lowers the bar. Like, maybe your grandma is listening to this now. You know? Like, She's in the game. You know what I mean? Yeah.

But the last thing I would say, like, how can you, like, sculpt your own kind of experience with the AI, is that you Nourish your own environment with the AI. Like, how do you use it? Are you using it for jokes? Sure. That's fine. There's no problem with that, but that's what you get out of it too. You know, if you're using it to optimize your workflow and experimenting with how it can streamline your teams and your work processes, then you'll probably get that out of it.

You know, if you if you put your diet into it, if you put your philosophy into it, you'll probably get those things.

So, like, it really depends on what's important to you. It just it's just a big,

Smart library sized mirror. It's crazy. I would love for AI to be like, hey, bud, listen.

We we know that you want to lose 10 pounds, but, like, it's because you're eating, like, 4 fig bars a day and, like, 8 string cheeses. You gotta, like, just stop that.

Like, that's what I want. That's right. That's right. Yep. That that would be that would be coming next. You know?

Yeah. Right. Yeah. Just input all the inputs, get your output right. What do you here's all of the entire everything about my life now.

Tell me what I'm good at. Tell me what I can do. Tell me how to improve things or or change or whatever. Yeah. I mean, that that's kind of like the idea, you know, is that is that you put in your data. You have to kinda, like, Say, like, yeah. Here's what I here's every run that I do. You know, it's just it's a run keeper thing. You know, you put all your Stravas in there, and then it says, oh, you're not running well. You know, like, that's the that's the AI component of it. So, like, there's, like, this handshake between, like, trust and safety and all this stuff.

But, you know, hopefully, that'll be ironed out in the future by people like your listeners. Yeah.

Absolutely.

Cool. Well, is there anything that you wanna get