AI, the brain, and the crowd, with CMU prof Niki Kittur

This episode explores the intersection of artificial intelligence, human cognition, and crowd-sourced problem solving with Carnegie Mellon University Professor Niki Kittur.

1. Crowd-augmented cognition can enhance decision-making and creative processes by combining human and machine strengths.
2. AI's role is seen as complementary to human intelligence, aiding in more creative and expansive thinking.
3. Practical applications of this technology have shown promising results in fields like conservation and automotive design.
4. The concept of Schema, a tool to help manage browser tabs and information overload, reflects ongoing challenges and advancements in digital tool development.
5. Kittur emphasizes the importance of evolving these tools to adapt to changes in human project focus and technological needs.

1: Introduction to Crowd-Augmented Cognition

Professor Niki Kittur discusses his work on crowd-augmented cognition, highlighting the synergy between human cognitive abilities and machine efficiency. Niki Kittur: "We're trying to harness what people and machines are each good at to improve how we understand and interact with information."

2: Applications and Transformations

Insights into real-world applications of crowd-augmented cognition, such as aiding automotive designers and conservation efforts. Niki Kittur: "Using AI to help automotive designers find inspiration from nature has opened new avenues for creative thinking."

3: The Future of AI and Crowd Cognition

Kittur shares his optimistic vision for AI's potential to solve major societal challenges through augmented cognition. Niki Kittur: "Combining AI with crowd intelligence has the potential to address some of the world's most pressing problems."

1. Embrace technology that combines AI with human insights for enhanced problem-solving.
2. Consider crowd-sourced solutions for complex or creative challenges.
3. Explore tools like Schema to manage and synthesize overwhelming information.
4. Stay open to adapting digital tools as technologies and needs evolve.
5. Leverage crowd-augmented cognition in both academic and practical settings to maximize outcomes.

This week on the GeekWire Podcast, we explore the frontier of crowd-augmented cognition, the concept of humans working together with the help of technology, including new ways that artificial intelligence is changing the field.

Our guest is Aniket (Niki) Kittur, a professor in the Human-Computer Interaction Institute at Carnegie Mellon University, where his research focuses on new methods of augmenting human intellect using crowds and computation.

We also talk about a related project that Kittur and his colleagues developed called Skeema, a browser tab manager that helped users organize their work, projects, and ultimately their brains in the process.

People

Niki Kittur

Companies

Carnegie Mellon University

Content Warnings:

None

Aniket Kittur
One of the things that it seems to correlate with relatively well is something that's known as the satisficer maximizer scale. You know, a maximizer is sort of what I describe myself as, and it sounds like you are. We really care about optimizing some particular things that are important to us. It's too bad. As you were starting to explain that, I was really hoping you were going to say that it correlates highly with charm and wealth, but maybe that could be a future study.

Todd Bishop
Yes. Let's just use the two of us as a sample size of two for your next study. There we go. I'm fine with that.

Welcome to Geekwire. I'm Geekwire co founder Todd Bishop. We are coming to you from Seattle, where we get to report each day on what's happening around us in business, technology and innovation. What happens here matters everywhere. And every week on this show, we talk about some of the most interesting stories and trends in the news.

We do focus on Seattle, but every once in a while it's fun to look elsewhere. And we consider Pittsburgh a sister city of sorts to Seattle, especially because the Geekwire team has spent extended amounts of time there, including our hq two project back in 2018, which is an epic story that I'll share with you another time. And I have. Joining me on the show today, a professor from Carnegie Mellon University in Pittsburgh, Nikki Couture. He is in the human Computer Interaction Institute at CMU, where he focuses on crowd augmented cognition, including the specialty of crowds and computation.

And there's an interesting backstory to how I initially connected with some of Nikki's work that we'll share as we go along. But first, Nikki, it's great to have you on the show. Thanks, Todd. It's great to be here. Thanks for having me.

So I was first introduced to your work through a project that you and some of your colleagues worked on called schema, and we'll get into that in just a second. But first off, can you give me just a broad overview of what you do and what your focus is as a professor of computer science? Yeah, absolutely. So my lab is really interested in how do we help people make sense of overwhelming information. So, especially, you can imagine online, we're constantly bombarded with all sorts of information that we have to make sense of, and we're looking at how do we take what people are good at and what machines are good at and put those together to kind of help us understand, make decisions, and be creative with that information better than we can alone.

I was looking at some of your research papers. And this dates back to things like mechanical Turk, which are kind of like the og of crowdsourcing in some ways, you know? So can you give me a sense for how crowd augmented work and crowd augmented cognition have transformed over the past 1015 years? Yeah, absolutely. Great question.

Aniket Kittur
So we actually started looking at this with Wikipedia as one of the greatest examples of one of the largest encyclopedias in the history of the world that was created entirely by volunteers working together. And we were looking at questions like, how does this actually work? How do you take lots of people, a lot of whom have very different points of view, and put them together and create something that results in a high quality artifact and something that, going back to what I'm interested in, this seems like a great way to take all of this fragmented information all over and start to pull that together. So we started by looking at that, and there's some kind of benefits there, but also some trade offs, like, you can't get volunteers really just want to work on the things that they care about. They don't necessarily want to do a lot of the kind of cleanup work or work on things that are less popular topics.

And so we started looking at, well, what are some other approaches that are coming out that you might be able to direct some of that work towards? And that's where we got into crowdsourcing. And again, we kind of hit it at that phase where it was just starting to come along with mechanical Turk and some other examples. And there we were looking a totally different problem of, we have a lot of people now, but we have to incentivize them to do well, and we have to think about, you know, how are they going to, how are we going to ensure high quality work? How are we going to coordinate these people who usually kind of work very independently from each other, you know, very different situation.

And then, you know, there's been a lot of other interesting types of crowd cognition that have come along, too. So things like scientific communities, like, I don't know if you've heard of the polymath projects. No. Really interesting community where they were trying to prove theorems together that had previously been unsolved, and they used kind of a very different structure. They actually had some, like a Fields medal winner who was helping to lead the project, as well as high school math teachers that would just all come together, work together on a wiki to try to explore different ways of.

Of solving this problem. So we've been looking at these very different ways of how we can put people together to go beyond what they're capable of alone? And how do you create these architectures that combine people with machines and the right ways to do that? One of your more recent papers looked at artificial intelligence in particular, and the combination of AI and this crowd augmented cognition. What's the potential these days for the those two things working together?

I'm really excited by this, and I really think that we might hear about artificial intelligence as in some sense, I wouldn't say taking over necessarily, but I would maybe say, how can we think of it as a complement to what we're doing and how to be more creative and think more broadly that way? So we have a lot of projects that have been exploring that. For example, we have one project working with Toyota to try to help their automotive designers be more creative and sort of unlock fixation. And we're using AI there to help them find inspirations from very different fields. Like, how does a crow flap its wings and create vortices that stabilize its flight path?

And can you use that for better mobility in different situations? And so can we use AI to kind of unlock those things that are hard for people to think of, to find, to kind of pull together, and then even pull those into the domain that the designer is trying to solve and think about how we can apply that? What are the trade offs? And so embed that into their design process more? It's interesting, my mind starts to go into these sort of fantastical places when I think about this, because obviously, much of the world right now is focused on what AI can do and this whole idea of a computational brain.

Todd Bishop
And I remember seeing the launch of mechanical Turk and thinking, wow, this is amazing. Like, you could potentially tap the collective intelligence of the entire population to do things that none of us could do individually or even in small teams. And so when I think about combining those two things, one thing that I actually haven't thought about for a long time, and then one thing I've been thinking about and all of us have been thinking about for 18 months, two years, and everything else with generative AI. Like, is it too much to say that some of the world's biggest problems could be solved with the combination of these two things? Am I going too far?

Aniket Kittur
I don't think so. Not at all. That's really exactly what's been driving my own research. I love that you're saying that. One of the things I've been talking about lately, too, is how do we put exactly those two things together?

I'm sort of trying to still come up with the right catchphrase for it. But I'm currently going with LLM, augmented cognition, the idea that instead of replacing our cognition, let's augment that, instead of replacing all of the different cool computational approaches we've developed so far, let's augment those. And instead of replacing all of these rich interfaces we used with just a simple chat of augment all of that together. So is there a way to put those together that can really make progress on those sort of societal challenges? Like you're pointing out, I think in the short term that augmentation is exactly the right approach.

I think. Is there anything in your lab that you could point to that hints at the promise of that long term vision? There's one project that we did with a nonprofit called Conservation X. So they put on contests, innovation contests for conservation causes. So reduced microplastics in the water, or various, let me prevent poaching in Africa.

And we were working with them to help the crowds that would submit contest ideas. Can we help them be more creative in the ideas that they submit? And one thing that we did was we were using AI to essentially match people who didn't make it through every stage of the contest, but they still had interesting ideas and expertise with people who were still in the contest. So, for example, maybe we take someone who's been using drones for kind of identifying poachers in Africa, and we say, hey, maybe, you know, something that could help these people who are trying to, you know, identify microplastics in the oceans. Like, right.

There might be some expertise that might help. The first time we did this, somewhat naively, we just matched people together and it resulted in actually worse ideas. And we were like, oh, no, this is terrible. But then we sort of looked at it, and it turns out there's a little bit of a sweet spot in kind of being outside the domain, but still being aligned in terms of what kind of mechanisms I can help you with. And so when we tuned our AI systems to pick up on those sweet spots, we actually found significantly improved quality of ideas when we did that matching together.

So I think that's one example where you can see that kind of using AI to, in our case, like, find the expertise of people and kind of put them together into a crowd that can think better together. We're actually making significant improvements on some real societal problems. So this is really interesting as the backdrop for your work. And, of course, as I mentioned, I have had experience more recently with a specific project that you worked on. And I've got lots of questions about that.

Todd Bishop
So we'll get into that when we come back. You're listening to geekwire and we'll be right back. Technology moves fast. I need to move faster. WGus competency based education puts me in control of how fast I move through my IT degree program.

Geekwire co founder Todd Bishop
I can accelerate my program by applying what I already know to my courses and focusing on the things I need to learn. Earn a respected, accredited degree that propels your career in the IT field. Learn more at wgu.edu, itserts included.

Todd Bishop
Welcome back. It's Todd Bishop. Joining me this week from Pittsburgh is Nikki Couture. He is a professor in the human Computer Interaction Institute at Carnegie Mellon University where he specializes in crowd augmented cognition. Nikki, I am totally confused because I've been using a tab manager called schema that you worked on with your colleagues, and we can get into what that is and how it worked and what's going to happen to it in the future and where the group is going.

But a tab manager? I mean, this is a very simple thing. Here's my question for you. What kind of lab rat have I been for your group over the past year or so?

Aniket Kittur
Well, first of all, thank you. We deeply appreciate you using it. And, yeah, we are kind of. It is a bit of a probe for us to kind of, like, learn and understand, you know, how people think, like, in our lives today. So.

Yeah, absolutely. Thank you. Let's take a step back there and give a sense for what schema is. So tell me the story of schema and how it came about and what it is. Okay.

So it's probably helpful to understand a little bit about me in addition to my research. I, you know, I'm. I'm the kind of person that is, like every. I like everything in my life to be efficient and optimized, right? So I'm like, you know, I'm going on my commute.

Like, how do I shave off 2 seconds? Or how do I, you know, I'll spend hours finding the perfect Airbnb, you know, for fun, like my daughter Ashi and I, like, we'll watch oddly satisfying videos where they're like, slice things up or weld things, and it's like they just fit together perfectly. And it's like a little burst of joy in my head, right. That whenever that happens. And so I started realizing there's something in my life that I'm using 6 hours a day.

It's very inefficient, it's very unoptimized, and it's very unsatisfying and that was my browser. And it felt like things are kind of getting worse. It's like time was going by and all of my life started to be in my browser along with all of us. All our files started migrating there, all of our apps and everything like that, and especially around the research that we do around how do we help people make sense of all this overwhelming information that started more and more to be what people were doing in their browsers. Schema came about from this idea that, what if we could help you make sense of everything in your browser, help you with all the tasks that you have that you're working on, that you're researching, help you organize and collect all of the information that you have, synthesize it, et cetera.

We had a bunch of different projects that we did along the way. Schema was not the first. We actually had a tab manager for mobile devices that my student Nathan Hahn worked on. It was called Bento browser. We started with mobile because we thought, hey, mobile is really hard.

If we can solve this for mobile, we can do it for Eric for everything. It sort of turns out that's true and false at the same time. Like, you know, it was hard, and we did some cool stuff there, but desktop is like, so much harder. Like you're, you know, you. It's just so messy, and people are doing all sorts of things.

And so we, we ended up taking a step back, and my student Joseph Chang worked on this was we started interviewing people, like knowledge workers and, you know, other people who have lots of tabs. And we started to say, why do you have those tabs? There was a lot of research when tabs first came out on, why are tabs better than not having tabs? But no one had really asked, why are tabs so bad nowadays? It seems like they're a lot worse nowadays.

We started interviewing people. We would ask them every few days. We asked them to explain their tabs to us and say, why is this tab still open? And we started to get some really interesting things, like some fairly obvious ones. Like, we use them as reminders.

I still have to do this task, but we come back and we'd ask them, and then they would start doing interesting things. Like, they would start whispering to us, and they'd be like, actually, I don't know about this tab. I just want to read it. It would be like, okay, so when are you going to read it? They're like, well, maybe someday, but, like, you know, honestly, I'm probably never going to read this, but it's like an opportunity for my future self, and I feel like I'm giving up on my future self if I close it.

So, you know, we started to find these interesting behaviors, and we wanted to, you know, so we sort of went through those, and that sort of gave rise to schema. So we said, hey, people are treating these as tasks. We want to develop a new way of handling your tabs. Like they're actually tasks and kind of, how would you put them away, resume them, prioritize them, annotate them in a way that would better match the way that people were trying to actually do things in their lives with tabs. And for people who want to look this up, and I should note that it's sunsetting at the end of June, if I understand correctly, June 2024.

Todd Bishop
But if they want to look it up, we should note it's spelled s k e e m a. That'll help you find it. But for me, in using it, I would essentially have a bunch of tabs open for me as a journalist covering a story related to a specific story I was working on. And whereas in the past, I might have to go through the process of bookmarking all those things under a folder in my bookmarks, instead, I was able to go to the schema interface, which was available for me on my opening tab in my web browser, and basically drag those open tabs into the designated place for that project and in the process, close those tabs. And it became my way of organizing my thoughts, organizing my projects.

And to your point, there's like, this hang up that I would have had previously, moving something from short term memory into long term memory, both physically on the computer, because that's basically what you're effectively doing in your browser and also in my own brain. And this gave me a mechanism, digitally, to do that biologically in my own head. And so I loved it. And I'm a little disappointed it's going away at the end of June, I have to tell you, which probably is a bittersweet thing for you to hear. Tell me about where the project is headed from here and why it's being sunset.

Aniket Kittur
Yeah. Yeah. So, first of all, I love. I love the way you described it. That's exactly what we were going for.

We sort of call it externalizing your mental model. Right. Or like the what's in your head and being able to take that out and kind of resume it whenever you need to. So I should say that it is a little bittersweet. I think we are.

You know, it's still a little bit unclear what is going to happen to it. So I think there is still some chance that it might stick around. We, you know, when we. You're not the only person who kind of came out and was like, hey, we love this thing. There was actually, you know, it was really nice.

Like it was a really great, a little outpouring. You know, some people have said, hey, we might want to take this over or we might want to help maintain this. So we're kind of working through right now, what are our options with what we do with it? But let me explain, maybe I can explain a little bit why we're making this choice and what we see is coming up. So we are not giving up on this vision, I think is the main point what you're thinking.

But I think this is still a real problem and we still want to solve this problem. And actually, we're a little bit ahead of where we were before. We have a really amazing team that has come together to work on this problem, and I'm really excited about the directions they're going in. There's really two reasons that we are kind of pivoting a little bit on this. So the first is something that we've learned about when people stop using schema.

And it turns out that schema for a lot of people works amazingly well for a while. And then slowly your life changes a little bit. You're maybe working on a few different projects. Maybe these ones you were working on aren't quite as relevant or the things you were using for them kind of went a little bit out of date. And it's not a lot of work to change things, but all of that kind of adds up.

All those little context changes add up, and we found that around four to six months in, often for people, there would be the switching point of just a little too much work to manually organize all of the stuff that is coming at me in this unending stream. And we saw some problems because of that. So what do you do from that? Somewhat obviously, we wanted to help people organize and help reduce the decision making that goes into taking this stream and saving it and making it useful to you. It was turning out to be a lot of effort to change the code.

You know, it was a pretty, pretty complex interface. There were some things that were a little bit challenging from the beginning. When you start to add more and more pages into it. There were some slowdowns just in terms of the drag and drop or just some small things like that. Between the challenge of not making people organize manually and this sort of technical debt.

We said, hey, let's start from scratch and let's take a swing at things. And I think it's actually opened up in some very interesting ways, some avenues we would not have explored before. Are you able to share any of those? Yeah, I'll maybe share one of them that I found surprising myself, which was, we started with this idea, what if we just closed any tabs that you hadn't used in over an hour? And, you know, that sounds crazy.

Like, to me, that sounds crazy. Like, I have so many tabs, I want to get back to them. I'm working on them. I'm, you know, I have a lot of projects in my life, and, you know, that sounds. That sounds nuts.

But when we did that and we just started using it ourselves, we noticed something really interesting, which is that it somehow changed the cost structure of the tabs. So what we would do is we take those tabs and put them kind of on a list for you on your new tab page or on a pin tab that we have. And I think the analogy. I don't know if you heard of the endowment effect in psychology. I'm a psychologist, actually, no, I don't.

Todd Bishop
Know what that is. So it's interesting. The endowment effect is, if I was to offer you an ugly mug, and I said, like, hey, you know, Todd, will you pay me $5 for this ugly mug? And you said, no, like, I'm not paying you $5 for that mug. That thing's ugly.

Aniket Kittur
That thing is ugly. But now I give you the mug and I say, hey, Todd, like, I got $5. Like, will you give me the mug? And you'll say, no, actually, I'm going to keep this mug. Because why, though?

Like, that doesn't make any rational sense, but just having the thing creates this almost like the sun cost for how much we value the thing. And I feel like that's a little bit of what was happening. When the tabs were open, I valued them a lot more than when they would then be in this list. In the list. I suddenly had to change my mind to be like, well, how much value do I get from reopening this tab?

And it sort of changed things in this way that it's almost like we've had some users describe it as a breath of fresh air. Because my tab bar is mostly clean now in a way that even in old schema, often my tab bar wouldn't be quite this clean. Now, I'm not saying that we figured it out by any means. There's a lot of issues still that we're working through around. Well, what if you're in a deep work session and you want all those tabs?

What if there's a group of tabs that need to stick around because you queued them up for later or for a meeting or something? You need to be reminded about those tabs. We're still figuring out how to put together all of these different jobs that tabs are doing for us in a way that's intuitive. But critically, in this new direction, we are not going to make people work. We're going to try to help people be organized, but not spend their time organizing.

We're still figuring out with the right interaction techniques, with the right AI and intelligence, and kind of the right ways of thinking about how people actually work, how to do that. I'm talking this week with Nikki Couture. He is a professor in the human Computer Interaction Institute at Carnegie Mellon University in Pittsburgh. And I have a few more questions for you, so if you can stick around, we'll be right back. This geekwire podcast is sponsored in part by Yale University Press.

D
Are you concerned about the rise of AI and how it will impact our society? Every day, artificial intelligence presents us with urgent ethical challenges. How do we harness this extraordinary technology to empower rather than oppress? Nigel Shadbolt and Roger Hampson have written a how to for building ethical machine intelligence. Their new book, as if human Ethics and artificial intelligence is now available wherever books are sold.

Todd Bishop
Welcome back. It's Todd Bishop from Geekwire. I'm geeking out this week about human cognition and crowd augmented cognition with Nikki couture. He is a professor in the human Computer Interaction Institute at Carnegie Mellon University. So, Nikki, as you mentioned, you are a psychologist and also involved in computer science.

And you mentioned earlier that you like things. I don't know if I'm paraphrasing this correctly, but I got the sense that things that are orderly and logical and snap together correctly in your life. Am I on the right track? In theory, I like those things. Exactly.

Aniket Kittur
I'm not sure my life is like. That, but yes, yes. Yeah. That you're drawn to those types of things. And I am right there with you.

Todd Bishop
Like, one of my greatest accomplishments, I feel like in my late forties, was realizing just how important folders on my desktop are. And this whole notion that rather than having a cluttered desktop, if I was just maniacal about getting those things, those files into folders, it just made my life so much easier. And so I've got this whole structure and these patterns that I do, and I talk to my colleagues, or I look over their shoulder at their browser tabs or their desktops, and I go, oh, my God. How do you even exist like this? One of my colleagues had so many tabs open, and I swear it took three, four minutes.

And I finally said, I'll come back. Let me know when you find it. What is it? It's almost like a Rorschach test, right? Why do some people seem to feel it and others not so much?

Aniket Kittur
Yeah. Yeah. What is that? So, first of all, you're absolutely right. We've actually done a survey on this, and it's about 50% to 60% of people who report that they feel this kind of tab overload and the pressures around that.

And so you're right. Like, it's not everybody, but it's a good chunk of people. It's almost like splitting people right in half. We've looked at, like, what is it? And does it correlate with other things?

One of the things that it seems to correlate with relatively well is something that's known as the satisficer maximizer scale. So, basically, a maximizer is sort of what I describe myself as. And it sounds like you are, and we're all a little bit of this in some aspects of our life, right? Like, we really care about optimizing some particular things that are important to us. But some people care more than others, right?

Especially in some area. In more areas. And the opposite of that is what's kind of being coined a satisficer. Someone who's like, I'm fine with just the minimum thing. The first thing I come across that satisfies my goal.

And so it turns out there's a. A set of questions. Like, there's a questionnaire, a survey you can give people that has questions around this. And we've actually found that we. I think we can get it down to something like three questions or something like that.

You know? And they are things like, you know, originally there were even things like, I mean, this will date me, but, like, you know, do you change the radio to find, you know, like, check for, like, what's the best thing on there? Or, you know, you say, I never settle for second best, or things like that. And that, like, surprisingly had a surprisingly high correlation with the people who would find our systems more useful. So the kinds of things that were around tap management and kind of making sense of information and stuff.

Todd Bishop
So maximizers, people who are looking for ways to make things efficient, and productive and really get them most out of what's around them. It's too bad. As you were starting to explain that, I was really hoping you were going to say that it correlates highly with charm and wealth, but maybe that could be a future study. Yes, yes. Let's just use the two of us as a sample size of two for your next study.

There we go. I'm fine with that. So, connect all of this around back to your original work, if you could. And that is one of my questions for you. Was schema kind of an alley, a blind alley that you went down that was an offshoot of your larger work in crowd augmented cognition, or is there a connection to your broader body of work?

Aniket Kittur
Yeah, absolutely. So I guess the vision in my work is how do we create a force in the world that can start to stitch information back together, stitch it back into knowledge that can, you know, we can constantly be improving and learning faster, being more innovative, solving the problems we need to solve. Like, how do we create a force that starts to do that? Schema is, I think, one step in that direction. Right?

So the idea that, you know, schema right now is sort of a very individual thing where you're taking sort of the fragmented information in your browser and trying to pull it together and make it useful for yourself. But now imagine that we can start to connect people who are pulling this together, right? And, like, maybe I can build on other things. So maybe, you know, Todd, you've gone to on a trip to Hawaii, and we might be going somewhere to. And, you know, there's probably a thousand, you know, 10,000 other people like me who've gone on this trip already, probably have done a bunch of work finding good sources and figuring out how to know, put those together.

Like, why can't I build on that? Why can't we build on the work that each other have been doing instead of starting from scratch every time, right? Sort of stand on the shoulder of giants, so to speak. And I think that's important, actually critical in every aspect of our lives, you know, whether it's science and understanding, scientific, like, what's happening in all the different threads of research, so that we can kind of keep advancing, important in making good decisions as consumers or as voters. And it's important in industry and work as well.

So I think what really schema is one small step towards, is trying to see how do we capture what is in our minds kind of externally, like you had said kind of earlier. Right? So we're taking this, this information we're turning it into our mental models and then kind of if the machine can understand our mental models and start to aggregate them and start to learn from them, suddenly I think that can unlock a lot of innovation and problem solving that will help us as a society. It strikes me that part of what you're describing sounds a little bit like the process of training large language models. Am I on the right track?

Well, large language models, exactly. So large language models are one way of taking all of that knowledge that is out there on the Internet and crystallizing it in some ways. Right. And I think they play a really important part in creating this kind of force for stitching things together, but I think they're not kind of sufficient in and of themselves. Like one of the greatest analogies I heard for for that, that I think resonates with me a lot is it's kind of like a blurry jpeg of the Internet.

And so a lot of our work recently has been looking at how do we put language models together with other things, like human minds are in the right pipelines and stuff to kind of support that goal, but I think that's right. Like there's constantly new things coming out that we can use to help further us towards that goal. And maybe the analogy that I like to think of is like, actually the human brain itself, right? So if you look at the brain, it's not made up of just one type of thing. It's actually a collection of lots of different components.

Like our memory works in a certain way. It's super parallel, but our ability to reason is super serial and can only hold a few things. And we put those together along with vision and kind of all these other modules that work in different ways. So why don't we think about this as kind of at a larger scale, an information processing problem, and let's find all of these different components. And LLMs are definitely one of those that have different characteristics that we can put together to kind of help us with our human goals.

So that's sort of the way that I would think about that. But I think you're exactly right. Well, I would just leave you and your team with this as you're thinking about the future of schema and what to do with it and where to go next. When I heard about the sunsetting via the email from one of your colleagues, I went searching for alternatives and I found something that's in the realm. I actually don't like it as much as I like schema.

Todd Bishop
It's called work ona wrokona and I'm paying for it. I am paying for it. Wow. So that will give you a sense for the value that I placed on the habit that you created in my life. And I would definitely pay for schema, some version of it if you brought it back or did something new with it, as you're alluding to here.

So take that for what it's worth, from one, again, from a focus group of one here. Thank you, Todd. That's. That's great to hear. And, you know, if.

Aniket Kittur
If anybody is listening and wants to be involved in that future and make Todd's dreams possible, please, please contact me. I, you know, as a professor, like, I definitely have different things that are calling out my time. So I think, like, you know, definitely looking for someone to help with driving that forward. I think it'd be great if we could do that. That would be wonderful.

Todd Bishop
Well, this has been a lot of fun. I really appreciate you talking with me. Thank you. It's been great to be on here. Appreciate it.

Nikki Couture is a professor in the Human Computer Interaction Institute at Carnegie Mellon University. I will link to his bio from the show notes just in case anybody does want to get in touch with him, in addition to other information related to his work. All right, thanks for listening, everybody. Our show is produced and edited by Kurt Milton. I'm Geekwire co founder Todd bishop.

We'll be back next week with a new episode of the Geekwire podcast.