ELISE HU: On today’s show, John Maeda. John Maeda is a Vice President of Design and Artificial Intelligence at Microsoft, and in his richly varied profession, he’s also been a professor, an writer, a university president, and a business executive. His digital artwork, books, lectures, research, and teaching have explored how digital technology can empower creativity. So we’ve got a wide-ranging chat today about this moment that we’re in for AI. So without further ado, my conversation with John Maeda.
ELISE HU: Thanks for coming on the show.
JOHN MAEDA: Glad to be here.
ELISE HU: And you lately made this big profession move to affix Microsoft.
JOHN MAEDA: Well, after I was in highschool, I attempted to use for an internship at Microsoft and I didn’t get in. So luckily they didn’t ask me the identical questions many years later, and I’m in.
ELISE HU: Well, welcome. There’s a lot to discuss with regards to AI, especially recent breakthroughs in large language models. It’s being called an inflection point. We’re hearing that quite a bit, or a Cambrian explosion. So why?
JOHN MAEDA: Well, I sort of chuckle after I keep reading things like inflection, Precambrian, or whatever. All these giant ways to say the entire world has shifted. I feel it’s just the right example of the Moore’s Law effect, that the concept of doubling doesn’t appear to be a giant deal when it’s like one becomes two, two becomes 4, 4 becomes eight, eight becomes 16. However the iteration, 30 or 40 of a Moore’s Law construct—it’s like ketchup, the old sort of ketchup within the glass bottle where it just all plops out and also you’re like, Whoa, where did this glob ketchup come from, since you’ve been holding the bottle over your head. The doubling feels very big.
ELISE HU: What are the implications?
JOHN MAEDA: Well, the implications are exciting because this technology is definitely sort of useful. I feel it introduces a brand new sort of scratch-your-head moment. The whole lot was command line based within the seventies and eighties: type in text and it does something for you. After which there was this graphic user interface boom, where suddenly you were capable of use a mouse and use a pc. It was democratizing. Satirically, this implies they’re going back to the command line, which is so interesting. But that is something that has been long foreseen, already a quite common user experience pattern in China, as an example, with a WeChat world. So I feel it was inevitable that we’d find yourself here.
ELISE HU: And once you mean that all the pieces’s sort of returning to the command line, are you able to talk a bit bit about that?
JOHN MAEDA: Well, I spent six years writing a book called The best way to Speak Machine, and your complete thesis was it’d be really good for individuals who don’t understand how computer science and AI works to know the mechanics, the physics underneath it. And at the top of the book, I spotted it wasn’t about the way to speak machine, but the way to speak human. Now we speak in natural language, English or whatever language you want. We’re speaking human to the machine.
ELISE HU: John Maeda, Wired magazine has said that Maeda is to design what Warren Buffett is to finance. I’m not going to ask you to must, you recognize, reply to that individual quote, but I’d like to know, because you might be so deeply embedded and regarded an actual leader within the designer community, how is the larger design community occupied with the potential and pitfalls of AI?
JOHN MAEDA: I feel that design today goes to play a very important role on this LLM AI world, with the attitude on ethics, what matters. Trust. These sorts of ideas, which have been embedded in great products are actually going to must be higher than ever with regards to this recent sort of AI. In the event you consider the Triangle of Engineering, product and design for technology products where, you recognize, product really has to hold that business role, has to make cash, has to grow, preferably. And engineering is playing the role of, does it work or does it not work? Does the bridge stand by itself? Okay, it worked. Design tends to be stuck in a job where, like, is the bridge pretty enough, which is typically pretty vital once you’re competing against other bridges. It also plays a very important role in, does it appear to be it’s not going to fall down? And or, you recognize I just discovered that a certain sort of stone really will not be good to take from the earth. Is that this bridge manufactured from that sort of stone? Then I actually don’t need to cross it. And I feel that design cares about these dimensions. Not only the aesthetics, the wonder, however the aesthetics of the ethics inside any experience you encounter, in a way that a product person doesn’t must care about as much and an engineer doesn’t must care about as much. They care about it, nevertheless it isn’t of their ‘jobs to be done’ list.
ELISE HU: Huh. Well, let’s discuss a few of these ethical concerns. What would you say are the questions that researchers, designers, firms grappling with AI and its potential—what must be worked out still most pressingly?
JOHN MAEDA: Well, there’s so many levels to that. , like, I’m creating the brand new design and tech report for South by Southwest, and I look back on the last nine years.
ELISE HU: Yeah.
JOHN MAEDA: In 2017, I noticed that Microsoft was really high-centered around responsible AI, inclusive design. And there’s one value that’s fairly easy but vital, is the worth of transparency, not like just see through, but do I understand it? And I feel at a really basic level, understanding large language model AI, the way it actually works, scientists are still attempting to figure that out. But even for the final person to assist them understand how it really works is a very important thing for design to do.
ELISE HU: How will people give you the option to make use of, beyond just these chatbots without delay, but other programs to extend their creativity and their productivity?
JOHN MAEDA: Ah. On this age of AI, there’s an easy method to be less terrified of it. Ask yourself, What do you not actually like doing in your job? Like, gather all that information right into a chart or summarize it for my boss. Versus, What do you must really keep? There are things that I enjoyed doing—occupied with the strategy of something and the way it would unfold. Think of the way to give you the option to do things 10 times faster than I ever thought possible, due to this fact, I can actually do 10x more. So on one hand, greater productivity since you’re doing what you might be most efficient and enthusiastic about. And in addition productivity, like, hey, I didn’t want to try this thing in the primary place. So it’s all gone.
ELISE HU: I understand you have got a metaphor you’ve been using, a scissors metaphor, to discuss AI. What’s it?
JOHN MAEDA: Oh, well, you recognize, I held on to this thing from my early days of trying to know artificial intelligence within the eighties. This work, from an individual named Herbert Simon, he’s a Carnegie Mellon AI legend, but interestingly, he received a Nobel Prize in economics. And he had this phrase that at all times stuck with me about how the method to consider intelligence is, it’s two blades of a scissor. One blade of the scissors is cognition, and the opposite blade is context. And once you slice, slice, slice those two together, rub them together, it creates what seems like intelligence, which is what’s happened with large language model AI.
ELISE HU: It’s not only cognition that computers can handle now, it’s context.
JOHN MAEDA: Well, this amazing cognition blade arrived. And now we are able to just, like, rub context against it. Like, I could take the last eight things we said to one another—the context—rub it against the cognition blade and say, Hey, what did we discuss?
ELISE HU: Yeah, sum up the themes of our conversation.
JOHN MAEDA: It does that. A cognition blade is like, able to go, boss. And the context is just pouring our information on top of it. And voila.
ELISE HU: Is AI able to creativity itself, or does it just facilitate human creativity?
JOHN MAEDA: One of the simplest ways I’ve heard this technology described is, it’s like a parrot, nevertheless it’s an awfully good parrot. It doesn’t just repeat back belongings you said to it, it could repeat back things that numerous people on the earth have said. So is it creative by itself? No. Can it make you more creative? Well, the reply is, each time you expose yourself to recent information, do you get more creative? Yeah. So it’s a method to speed up your personal creativity.
ELISE HU: Well, we’re asking a wide range of people such as you, experts of their field, in addition to civilians, how they’re using AI in only their on a regular basis lives. So what’s it for you?
JOHN MAEDA: Well, as you discover the way to leverage this odd technology, you discover that, wow, that’s easy. Like, I at all times use Python, the programming language Python, to do things fast. Like, oh, I’ve got to sort this document this fashion, I’m going to jot down a Python code or whatever. Now, I just give it to the model and say, Hey, that is all of the stuff I even have, the context. Are you able to now categorize this stuff? And it’s like magic, voila. Or I’m attempting to figure this thing out and I would like 10 different perspectives, so are you able to be someone who’s a botanist? Are you able to be someone who’s a shopkeeper? So it’s like running user research studies in a short time.
ELISE HU: Yeah.
JOHN MAEDA: With fictional people, they’re higher than a persona, actually. You may talk with them.
ELISE HU: Oh, that’s interesting. Do you have got sort of a dream scenario for where things look two to 5 years from now?
JOHN MAEDA: I feel that we’re already seeing elements of how this model-based work that we do, whether the model is language-based or it’s image-based or interaction-based, it’s going to affect how we do things. Once we make images or image with text or video, mainly all the pieces we do to speak, I feel it’s going to make it quite a bit easier for us to do the part that we normally only do if we’re not drained, you recognize? I mean, what number of things have you ever made where you’re like, Oh my gosh, all this planning, here I’m going, do it. Okay, I did it. Well, I’m really drained. I don’t know what it’s going to be like, but I feel like I’m going to do the part that I actually thought I needs to be doing on the very end, but I got too drained.
ELISE HU: I feel prefer it could increase our body of data too, right, to give you the option to see so many things in alternative ways or go searching the corners that we were too drained to go searching.
JOHN MAEDA: Oh, 100%. This whole list of things that we are able to do higher, that I keep asking myself, What do I not wish to do now? What can I Marie Kondo out of my brain? And now what if I had more time? What would I do as a substitute?
ELISE HU: Yeah. Okay, so let’s talk a bit bit about leaders of organizations and leadership. What should leaders, or what could they do, to harness this potential of AI, not only for themselves, but in addition for his or her teams?
JOHN MAEDA: I feel what’s really powerful for leaders is the power to listen broadly. Since the only way for leaders to listen without delay, generally speaking, is thru one-on-ones, which don’t scale.
ELISE HU: Those are only their lieutenants, though, right? It’s not a foot soldier.
JOHN MAEDA: Well, you recognize, the nice leaders skip levels and really break the principles and, like, discuss with everyone. I like those sorts of leaders since it creates individual bonds of trust, which suggests the organization can normally move faster due to that. Nevertheless, it takes numerous time. So, ultimately, you have got the opposite selection, which is surveys. As we all know, the very best a part of those surveys is the fill-in-the-blank part. Previously we only had word clouds, but now, bosses can discuss with all of that feedback and say, Tell me in regards to the time I allow you to all down. Tell me in regards to the time that you simply felt really proud to be here. So it’s like doing Q&A, 24/7.
ELISE HU: Yeah. And the potential for having the ability to take those learnings and apply them, or change direction or provide you with a brand new vision, are really limitless.
JOHN MAEDA: It mainly lets them save time to do the part that they probably were hired to do, but they might never do since the logistics of having the ability to communicate through a hierarchy are tremendous, as you recognize.
ELISE HU: Okay. More broadly, John, you have got spoken quite a bit on what corporations and company leaders can learn from entrepreneurs or more scrappy start-ups. What can they learn?
JOHN MAEDA: I felt that there are these start-up firms and there are the grown-up firms. And the irony is that start-ups need to find yourself just like the grown-ups, but, you recognize, the grown-ups are at all times like, Gee, I wish I used to be a start-up again. So I feel that each can learn from one another. But the largest thing one can learn from an entrepreneur is proximity to the shopper, since it’s like a automotive with no partitions, barely wheels. It’s got a jagged steering wheel. It’s like, Ouch. And the shopper’s like, hey, I don’t like this, on a regular basis. Whereas should you’re in a big corporation, you’re sort of like in an SUV or a bus or a jumbo jet. And so you actually can’t feel the shopper and the way they’re experiencing what you’re providing to them. So, learn from entrepreneurs the way to take heed to the shopper, and that goes to the fantastic thing about these recent LLM AI systems. It signifies that the CEO or any different level of a company can actually begin to speak with customers, effectively, 24/7—understand what they’re pondering from all the shopper support data that they get, which if I were a customer support skilled, I’d think, Wow, thank goodness it’s not only me hearing this. It’s my boss, my boss’s boss, my boss’s boss. Entrepreneurs are great with customers and that’s where they’ll learn.
ELISE HU: Okay, so for the listeners on the market who’re excited in regards to the potential of AI and numerous the things that we’ve got talked about, where should they begin?
JOHN MAEDA: Well, they need to first start by trying these items out. I feel that I even have presented to a wide range of audiences of all sizes, and I’ll ask, Hey, you recognize, who’s used this thing, ChatGPT, before? Who uses it every single day? Like, who’s never heard of it? And ultimately, there are those that haven’t heard of it. The second thing is to interrupt that transparency barrier, because what persons are afraid of is that they don’t really understand it in any respect. I wish to indicate that the one letter you have got to care about in C-H-A-T G-P-T is the P. The P stands for pre-trained. So what meaning is you’re getting out-of-the-box, powerful machine learning. As you recognize, within the old days, the one method to get AIML was to have numerous data, since you had to coach it. What’s different this time is, it comes pre-trained. It’s like a puppy that arrives, like capable of do all the pieces. And so that you’re freaked out. You’re like, Whoa, this AI comes pre-trained? After which when you recover from that cognitive hurdle, you discover it could do numerous belongings you didn’t expect. And so, try it out. Learn from it. Find out how prompts work, learn the way context works. Take the scissor blades and begin snip, snip, snipping. I feel the opposite thing that’s actionable is to assist everyone of their organization understand that change is at all times a scary thing. And it is a change that actually is a big blob of ketchup coming out, possibly the entire bottle got here out unexpectedly. And so the subsequent response is like, Hey, I don’t like ketchup. Ketchup will not be good for you. , that sort of feeling. And so every organization should ask the query. Let’s first understand it. Let’s try it. Let’s learn what the cons list are, like, pros and cons. Let’s take a look at the professionals and just sort of adapt as quickly as possible to what we would like to make use of and what we don’t need to use. Because this technology is very similar to the world wide web’s emergence. I’m undecided should you were like me, but when someone showed me a homepage, I used to be like, Nah, never going to take off. Like a month and a half later, well, gotta construct a homepage. So it’s like that, I feel.
ELISE HU: John, you mentioned that you simply are neuroatypical, and so many people on the market are. So I’d like to know what potential you see for AI and accessibility.
JOHN MAEDA: Well, I just like the proven fact that I can discuss with it and share things, and I can ask it, Hey, are you able to make that more sense to nearly all of people? And I feel that it is a superb translator and interpreter of things. I’m also high on the autistic spectrum, so sometimes I can’t read emotion thoroughly. So I can ask it to inform me, like, what does this mean? Like what’s the downlow? And that’s extremely helpful.
ELISE HU: I like that. Okay. Thanks a lot.
JOHN MAEDA: Well, thanks for having me.
ELISE HU: Thanks again to John Maeda. I loved that conversation. And that’s it for this episode of the WorkLab podcast from Microsoft. Please subscribe and check back for the subsequent episode. In the event you’ve got a matter you’d like us to pose to leaders, drop us an email at worklab@microsoft.com. And take a look at the WorkLab digital publication, where one can find transcripts of all our episodes, together with thoughtful stories that explore the ways we work today. You’ll find all of it at Microsoft.com/WorkLab. As for this podcast, please rate us, review, and follow us wherever you listen. It helps us out. The WorkLab podcast is a spot for experts to share their insights and opinions. WorkLab is produced by Microsoft with Godfrey Dadich Partners and Reasonable Volume. I’m your host, Elise Hu. Mary Melton is our correspondent. Sharon Kallander and Matthew Duncan produced this podcast. Jessica Voelker is the WorkLab editor. Okay, until next time.