Jaime Teevan on the future of work

As we're thinking about the future of work, even in the context of AI, you know, we talk a lot about how we can start automating the drudgery or sort of the repetitive parts of work. Sometimes those pieces of work are actually important for human attention. It's not just about like, okay, raw, what do the large language models do well? How do we bring them together to make it better? But it's like, how do we set people up to contribute the best to think well, to see things in new ways? Hi, I'm Reid Hoffman. And I'm Arya Finger. We want to know what happens if, in the future, everything breaks humanity's way. We're speaking with visionaries in every field, from climate science to criminal justice and from entertainment to education. These conversations also feature another kind of guest, GPD4, OpenAI's latest and most powerful language model to date. Each episode will have a companion story, which we've generated with GPD4 to spark discussion. You can find these stories down in the show notes. In each episode, we seek out the brightest version of the future and learn what it'll take to get there. This is possible. Obviously, 2023 being the year of AI, one of the primary things everyone's talking about is what is the future work. That's why we're so excited to be talking to Jamie Tavon. Jamie is chief scientist and technical fellow at Microsoft, where she was responsible for driving research-backed innovation in the company's core products. Jamie is an advocate for finding smarter ways for people to make the most of their time. She leads Microsoft's future of work initiative, which explores how everything from AI to hybrid work changes the way people get things done. And part of the reason, of course, you know, doing the book in Promptu was to say, actually, in fact, we're going to have these all-ha moments, e.g. the amplification of human ability, or AI as amplification intelligence versus, you know, kind of artificial intelligence, because people are always talking about replacement versus amplification and augmentation. And I think Jamie is going to be great to talk to about this because, you know, part of her research for years, the things she's been doing at Microsoft, I've talked to her a number of times, where she's bringing a kind of a clear kind of a scientist. And I, let's look at the data. Let's study this about like, what are the things we could do to help people work? One of the things that make them more productive, happier, you know, more connected, more creative in the things that they're doing. And so this discussion about the future work will be, you know, kind of grounded in real lenses of the future versus, you know, what you normally get, which is just people's fears or uncertainties as they approach the topic. And I think everyone is, of course, so interested in this topic because it's so personal, where you work and what you do, everyone says, you know, you spend 40, 50, 60 hours a week doing this work. And so what is it going to look like in 10 years? What is it going to look like in 20 years? Like, that's what we all want to know. And I think also, like personal productivity, it's like, how can we all be better at our jobs? I'm also so excited about this episode, because the future of work affects everyone. So Read and I reached out to a few key people in our network to hear their hottest takes about what's to come in the world of work. You're going to hear voice mails from these special guest stars throughout the episode. Here is our conversation with Jamie Tupon. Jamie, it's a great pleasure to be doing this. One of the things that I have learned from our years of working together in the Microsoft thing is how thoughtful and kind of data and truth oriented you are in these things. It isn't just kind of like an evangelism. It's a, it's a, no, no, we're studying how to make this stuff better because it's not just a question of opinion. Is it going to be like, good, bad, ugly, you know, wonderful? You know, did it up, but actually, no, no, here's, here's what we're learning. And here's how we're, we're proceeding very intelligently. So welcome to the possible podcast. Thank you. It's my pleasure. Jamie, I'd so lovely to meet you. I was saying earlier that I think of myself as sort of like a superhero because I'm the mom of three boys. But then I read that you're the mom of four boys. And I was like, oh my god, I need to learn everything from this woman. And then actually that moves into my first question. It's like everyone has to, but especially moms, dads, working parents, it's like we're on this constant quest for productivity. Like how can we be more productive? How can we leave more time to the things we want to do? And you've done so much with micro productivity and micro tasks. Like do you think micro tasks are the way of the future? I would love to hear your thoughts on them. Oh, that's such a good question. And actually my kids, I strongly believe have made me more productive. Like there's this forcing function that moms have like in parents, like just you have to use your time and it up. It forces you to use your time thoughtfully. Like when I'm working, I'm focused on work and I'm not going to screw around it work because I know I have other things that I could be doing if I'm not. And I would say, you know, we're talking, we're talking a lot about sort of how work is changing and in particular how technology is changing, um, she can change work. It was interesting for me, even just to reflect on how much has changed since my kids were born. So Griffin was born in 2004. And when he was born, Facebook didn't exist. Twitter didn't exist. iPhone didn't exist. Some important things that really have opened up a lot of the AI revolution to like image net didn't exist. Crowdsourcing wasn't a thing. And so much has changed. And I would say the research that I did related to micro productivity was really inspired by those sort of technological transitions. In addition to sort of my family and personal transition. Here I am. I've got four babies. And it's overwhelming and it's so much work because I never know when somebody's going to wake up or need something from me. And so I got really interested in how we could use the phone, use the small little bits of work that we'd, I mean, in some ways, what you do on Facebook or Twitter or even in Crowdsourcing are little bits of work. And we were getting, you know, as a, as a scientific community, we were getting very smart about how to take these little bits of work and stitch them together into something bigger. And I got interested in how I could do that for myself so that while they were napping or doing something else, I could be doing something that was productive and valuable to me. And I feel like so often we think about that as interrupting our work. It's like, oh, you can only do work if you need three hours of time to like sit at a keyboard and do three hours in a row. But you found that we never have that. Can you be productive in the short bursts? And what did you find about that? There's value to focused work and taking a lot of time. So I'm not trying to discount the value of that work. But there's also value to these small little bits. And it can even be intentional. So we've done some research, for example, that shows coming at the same problem from different perspectives or different times or different locations actually inspires creativity. So you may have, you may have an email that you want to respond to. And it's hard. And actually, you know this like, this is a comment like, I get emails and like, I'm like, oh my gosh, what a jerk. Why are they sending me this annoying mail? And we all know that you shouldn't reply to that mail immediately. And that ability to sort of look at the mail, internalize the mail. Think about what matters. Think about the different kinds of points you want. Right pieces of it come back to it. Actually allows you to internalize the content that you received, see what's valuable, you know, to get to get over sort of your initial range reaction or response and and see what's valuable there. And then and then respond to in a meaningful way. So there's all sorts of different things that you can do with these little bits at time. Say a little bit about how these like micro tasks and the work is evolving given this M365 co-pilot and what the work that you've done around like, like how gets crafted to really be helpful and really on target for making people more creative and productive. All right. So Microsoft announced a new feature called co-pilot that integrates AI technology into its Microsoft Office software. Co-pilot combines the power of large language models with your data in the Microsoft graph and the Microsoft 365 apps to turn your words into the most powerful productivity tool on the planet. Microsoft's promise is that this will effectively allow you to start interacting via natural language with your computer. A lot of the early micro productivity work was sort of looking forward to this moment where AI was able to be a real kind of participant in what you're doing and thinking about that. One of the things that's happening with things like the Microsoft 365 co-pilot is that we're now able to use language to engage with the system, but more than just using language, we're able to use conversations. So it's like that back and forth and that ability to iterate. And so when we were just talking about what it means to look at something from a new perspective, it becomes really useful if the system can actually be helping coming up with different ideas and different perspectives and then you have the space to respond to that. It's so funny just today. I just got an email from a friend who's she's a one woman social media agency and she was saying that she's been using Chad GPT to be her partner. She's like, I'm only one woman. And so Chad GPT gives me feedback and gives me ideas and I bounce out ideas off it. It was so interesting that she was saying this is my co-worker for the time being because because I'm starting out, it's a startup and it's just me. So your point, that's how she's using it as her co-pilot. When you think about language generation, one of the obvious things is content creation. Of course you can say, "Oh, I have some ideas, make it into a lot of content." Another place where it's sort of obvious is summarization. How do you take a piece of content and then summarize it? But exactly as you're saying, "Aria, it can do so much more than that." And it becomes really interesting when it can give you feedbacks. I've actually found when I've had, for example, hard emails to write or something that I'm not exactly sure how it's going to land. And I would guess this is great for somebody who's a social media person as well. You want to see the different perspectives people might have. What are the things that might raise red flags for others in this so that I can respond to that first? Hey, Aria and Reid. This is Jessie Hebel. I'm host of Sleet & FliveHit Podcast, which is called "Hello Monday." And at "Hello Monday," we spend our time thinking and talking about the future of work specifically and how that work is changing us. And so you ask, "What is your hottest take about how we will work in the future?" And I have one fairly clear thought about that. The work of the future for us is relationships. This becomes no much more clear as A.I. comes on the field and takes over so many of the aspects of what I had always thought of as the work that we needed, specific education to prepare for. Now, A.I. can do it in minutes, right? But the thing that uniquely is ours and the thing that we will need to lean into as all of the institutions around us to be invented, as all of the norms we have come to depend on, our reexamined, is the ways in which we can be present with each other, in which we can support each other, in which we can know each other, our relationships. You know, some of your work a couple of years ago was also when we was going on with hybrid and COVID in the pandemic, Microsoft released a report, not having what work affects collaboration. What do you think were the kind of key learnings going through the pandemic relative to hybrid teams, hybrid work, remote? What should people kind of have learned from that going forward? Yeah, and it's fascinating how intertwined, even if we maybe think of these two major disruptions to work that we're experiencing right now, one, the rapid shift to remote and then hybrid work and the other, A.I. how much they're actually intertwined and related. You know, in many ways, I think of the shift to remote work as breaking down spatial boundaries. So all of a sudden space was something that we stopped thinking about in the same way I can work remotely. I can collaborate across different time zones, across different locations. But in the process, that broke down a bunch of other boundaries because space actually affords temporal boundaries. And when you go into the office, that's when work starts and when you go home, that's when it ends, it broke down a lot of our boundaries between work and life. It forced us to rethink of the technology of space in a fundamentally different way. It's basically a technology we've been using from millennia to get things done. You know, if you want to transact, you go in person. If you want to bump into people or have brainstorm or have new conversations, you go and do that with them physically co-present. And now we're thinking about how you can do that in a technologically mediated way. The fact that so many conversations are now technologically intermediated, A.I. now has this real opportunity to help us figure out. Like we've sort of like threw up in the air, what does time mean, what does space mean, and we need to make sense of that. And we have all this new data and all these new surfaces on which people are interacting. And we have this opportunity then to bring A.I. to help make sense of that and figure out where it's going to go. Well, I'm looking forward to when we're doing, you know, kind of remote podcasts like this one because one of the things it does all it was to do is relatively easily assemble with Nari in New York, you and I in Washington, you know, it put it in different locations to a semilist, but to have the A.I. co-pilot going, oh, you just said this. I ask her about this, this thing. Right. I'm looking forward to today. You don't have long to look forward because we actually, I mean, that's we now in teams can can actually help you real time in a conversation and be like, and especially be like, oh, wait, read. You meant to talk about this or you seem like you maybe disagreed about this topic and can start tackling that as well. Talking about like what the difference in work is going to be in five, 10, 20 years, I remember when I joined the workforce, my boss had a big paper calendar on her desk and the only way you could get time on her calendar was by going into her office and writing, you know, 4 p.m. today and then we erased it as something changed. So the idea of paper calendar seems really quaint now. What are the things that we're doing today that you think in the future are going to be quaint? It's just we're not going to do that. It's just going to be a different world of work if we look down the road. Certainly the idea of not being able to speak to your computer will be weird. I suspect the kind of artifacts that we use for communication are going to change in a fundamental way as well with conversations becoming where knowledge is embedded and where where. And so really our focus is going to be on how do we, how do we talk? Do we communicate less on how do I put this in a document and present it separately? Let's turn to what GPT4 posited that the world of work might look like in the future. So the first AI story, it was about Anna. She was a 35 year old woman in Mexico City and she uploaded all of her proprietary data and she created LLM clones of herself. Anna enjoyed her work at her startup, Lingo, but she also had a personal project that she pursued in her spare time. She wanted to create an LLM clone of herself that could act as her alter ego in the virtual world. She saw it as a way of expressing herself, expanding her horizons and as a potential source of income since she could use it to take on more work opportunities without sacrificing her own time and energy. She started by collecting and organizing all the data that represented her identity and experience. Her projects, notes, intellectual property, social media posts, emails, chats, photos, videos, audio recordings, biometric data, and more. She stored, encrypted, and backed up her data to ensure that she had full control over it. She then used Lingo's LLM as a base model and fine tuned it with her own data using 100 gigabytes of cloud storage space. She also added some features like emotion recognition, voice synthesis, face generation, and personality profiling. The AI story positive that instead of doing the work, she actually became a manager of four or five different cells of her that were doing the work for her. So just I would love your reflection on that story. What did it get right? What did it get totally wrong? Like where do you see that as part of the future of work? It was a fun story. And probably the biggest challenge I saw with it was it really personified the AI. It was how does AI become a replacement for Anna. And I think it's much more interesting to think about how it can make Anna better or do things differently. So I actually used GPT-4 to ask some questions about that. I love it. The first one was really to help me think through it. Actually, the first thing I did was, and I've been doing this for all my documents, you get, you go to all these meetings and you get pre-read documents. My favorite thing to do when I get a pre-read document now is to get it summarized, but not just raw summarized. I like it summarized as a poem. For some reason it's a lot more fun for me to read it as a poem and brings a little bit of joy into that, into that, into that, and I just pay attention and I'm like, "Oh, that rhyme didn't work." So I summarized it as a poem, but the thing I liked best was actually encouraging the model to help me think of ways to think further. So I gave it the prompt. I'll read the prompt. I said, "As good as LLMs are, I don't think using them to clone people is the right thing to do. It doesn't seem like the best way to capture their potential." And what other things might, on a ton with the technology that are really different and novel ways of imagining it? And it had a bunch of interesting suggestions. One that I really liked was thinking about it as a digital mentor coach to help her grow and think differently. Another one was to think about using the LLMs as a digital artist or storyteller. And like, you know, I think of all the work that we do just to capture our lives and think things through and how a special live would be to have that captured as well. No, I think about the coach all the time, because I feel like in so many organizations, you know, the CEO, the executive team, of course they have a coach. Of course they're talking to someone outside to make them better. And, you know, we would love to have a human coach for everyone in the organization, but it's not going to happen. And so if we can have an AI coach to help everyone at every level, I mean, that's a great question. we've just unlocked so much productivity. So that's awesome. I love they came up with it. It's a little bit of like, and then she creates a manager to manage her clones that she creates and it's a little bit turtles all the way down. So I actually asked about that as well, and it came up with sort of this story of imagining clones of clones. And I thought that was sort of a fun one. It says, "Anna created an LLM clone of herself to work for four different employers at the same time." And she was so successful that she decided to create another clone of herself to manage her clones. And then she was so busy that she decided to create another clone of herself to enjoy her life. And she was so happy that then she decided to create another clone to share her happiness. And then she was so cloned that she decided to create another clone to help herself remember who she was. - Yes. And now the phrase won't be turtles all the way down. I've been clones all the way down. - It's all the way down. - I think there's a couple very good highlights. One is, it's a co-pilot. And it doesn't mean that what it says is necessarily true. It isn't what it says is the complete story. It's ways of kind of elevating your game. I also, whenever it was like, I think the clones thing is particularly useful. And the real thing is how is this tool really catalytic? It can help with the blank page problem. It can help with, I don't think we're gonna get to when we're doing a possible podcast. We'll say, "Well, my clone will talk to you or clone." (laughs) And that'll be done. But it can help us elevate our game. (electronic beeping) - Hi everyone. My name is Papia DeBroy and I'm with the organization Opportunity at Work. Make our communities as citizens competitive and resilient to what are now constant changes in our economy as the future of work is evolving. We have to activate the potential of the entire US workforce. That includes more than 60 million workers who have a bachelor's degree. It also includes some more than 70 million workers who are skilled through alternative groups or starts. And they face incredible misperceptions. The majority of managers in this country believe the majority of the workforce has a bachelor's degree. This impacts their perceptions of who should get jobs. It also impacts how managers invest in their current workforce. These are invisible barriers. They come at every turn for workers without bachelor's degrees. Yet, our analysis actually suggests more than 30 million stars have the skills to make transitions to jobs that pay on average over 50% more than what they currently earned. It's time we rid ourselves of that invisible barrier that we tear the paper ceiling and see the world beyond it. - Before we go to some more of the AI stories, I think one of the things we should address is obviously the general dialogue around AI, generally in the press is a worry about job replacement. And what is your research currently leads you to think about how our job's gonna be transformed? What is like, you know, a lot of people listening to this will have like, what about my job? You know, like I even go to an investor conference and I have investors asking me like, well, what is the AI gonna mean for my job? And you know, like, well, you're gonna be able to invest in a lot of really cool AI things. That's one. But anyway, so what's your current job? What's your current lens on this and how would you talk to the kind of the general dialogue on this? - Yeah, no, and it's terrifying. It's really scary to have this new powerful AI tool to work with. I get scared too. Like I have a lot of empathy for that. These large language models are a really, really powerful tool. And we're going to have to figure out how to use that tool and it's going to open up a lot of opportunities. It takes time for our imagination to really get to work and figure that out. You know, anytime you think about a new technology, sort of the most obvious uses of that technology are substitutionary. Like I even think about something like very simple. Think of GPS technology and how we initially use GPS type of technology. Like, oh, it's great. It's a map. I don't need to know where I'm going. I can just go follow my phone and it will tell me where I'm gonna go. And it's been great. That way, and that has changed. I no longer mean maps. I actually don't pay as much attention to where I'm going. People do a little worse with navigating and locations. But what we didn't see the time that was hard is all of the new things that it was gonna bring. I can now quickly find places to eat near me. I can find the, I can know where my husband is on his way home, but it can do so. But then we can start looking at that data and start thinking even more creatively. We can see traffic because we know where people are traveling. We can do better road planning as a result of that. And so all of these complementing uses of the technology are going to emerge and are starting to emerge, but they're hard to see. So it's not, like, I think it makes sense that we find it scary, but there's a lot of opportunity to do things that are pretty amazing and interesting. - There's relatively smaller number of jobs. It'll just be completely eliminated, not zero, but completely eliminated or reduced and substantially in count. I think what's jobs will be transformed. So what the work you were doing before, you know, now or before will be different than the work you will be doing. Like one of my big hopes is as a lot of the equivalent of form entry part of this will now be all much easier and people can focus on other things than form entry part of the job. It's because we almost all have some form of form entry as part of jobs. Some people more, some people less. But like one of the other AI stories we talked about was new occupations that didn't exist today. Did you see that little amusement of that AI story? - Here are 10 common occupations in 2053. Lunar minor, bio printer technician, climate engineer, cyber security analyst, virtual reality therapist, nano-medicine specialist, blockchain developer, augmented reality designer, gene editor, space tourism guide. - Yeah, I loved all the outer space stuff. - Yes, although by the way, when I read it, it was like, well, I was clearly an instance where there wasn't, you know, you would have failed in an interview test for me because it's like, why would you send a human to do litter of mining versus a robot? - Yeah, right. - Like, then this is not actually thinking it through. So what did you think about those occupations and which other ones do you think just to help people think is like, well, here are new things that are gonna be coming because in this and obviously it's always, you know, through a glass darkly and a complicated, very fast moving environment. - So GPT-4 was that help here as well? I thought that the jobs were a little too predictable. So I asked for help coming up with sort of bigger changes or jobs that looked more drastically different. And the suggestion was to, you know, how much jobs are gonna change according to GPT-4 was really a function of three things. One, the degree of complexity on certain to your novelty involved in the job task or problem. And I think that the idea there is it, you know, we're gonna want those jobs that are really uncertain or new are places where people are gonna be leaning in. Another was the degree of human empathy or emotion or ethics required for the job. And the last one was the degree of regulation, resistance or risk associated with the job and its impact. So to looking at the jobs that were suggested, that then sort of would put things like virtual reality therapist or gene editor or even space tourism guide at sort of the high level of jobs that are gonna be important in the future. And then as exactly as you were saying, read things like lunar minor or cybersecurity analyst down at the other end of potential jobs. All right, I'm glad I'm not gonna have to go to the mood. I'm gonna lie. (laughs) (phone ringing) - My name is Simone Solvoss and I'm the author of the book, "The Good Enough Job" reclaiming life from work. The most interesting stat about how we work today is that 40 years ago, the average American and the average German worker worked the exact same number of hours each year. Today, the average American works 30% more than the average German. My hardest take is that by the year of 2043, universal healthcare is going to become a reality in the United States. Part of the reason why our relationship to work is so fraught is that the consequences of losing work are so dire. The United States spends a whopping 40% more of capital than any other country in the world on healthcare. In the next 20 years, our country will radically redesign the way we offer healthcare and decouple our basic human needs from our employment status. It'll be one of the most revolutionary shifts in cultural opinion on the par of the federal legalization of gay marriage or states legalizing cannabis. (phone ringing) - So another part of your research that we were excited about was your team did some great research, putting EEG caps to monitor activity on people's brains, to see where people's brains lit up, to see where people were stressed out. And I would just love to hear like what were some of those specific findings and how can we use them to make our current and future jobs even better? - Yeah, so we did a bunch of studies to try and understand the impact of remote and hybrid work on people and we studied it in a number of ways [BLANK_AUDIO] EG was one of the ways that we studied it. And part because one of the things that you can measure fairly well from brain studies is stress. You know, so we ran, they said, "These are small-scale. We're not, you know, this is, we bring people into the lab and are studying them." They're looking at the stress people feel working in back-to-back meetings. Well, remote, you would see, for example, that the process of going in back-to-back meetings increases your stress over the course of the day. But if you do something like take five minute breaks between the meetings or even better step outside and look at nature for a minute, that that significantly reduces the stress and then you can show up better at those meetings. And I think it makes a really important point as we're thinking about the future of work. Even in the context of AI, you know, we talk a lot about how we can start automating the drudgery or sort of the repetitive parts of work. You know, I think about taking, doing the dishes and that sort of meditation for me. I even think about how I start working on a PowerPoint document. I often like to sit in futs with like the bullets in the formatting before I get into the deck. And we've done a lot of research on studying the transition of attention as well. And I think what the studies really suggest is how it's not just about like, okay, raw, what do people do well, what do the large language models do well, how do we bring them together to make it better? And that's not always just rawly like, okay, come up with a great idea. Tell it to me now. You know. And actually one of the things with the brain studies that I think is going to be really fun. You can also like joy is another thing that you can see pretty well. And I just think of how much fun it is to work with these models and see things in new ways. And I think there's a real opportunity to lean into that joy. Do you think it's going to be more, you know, like you're going to discover a certain number of people are going to be trying to summarizing their documents into into poems and sawdates as part of the joy of that? You're going to be curious like, what do you think your particular way of actually in fact using these tools to re-engage your own, you know, delight as a way of being, bring mind and focus and attention? Do you think that's going to be a broad, broadly adopted thing or maybe not all poetry but a variety of them? Yeah, absolutely. And joy and magic has been really at the forefront of what we're building. One of the things that was interesting that we learned from GitHub Co-Pilot, for example, as we were going through it, they're sort of all these standard metrics you can use to determine the quality of the suggestions you're making to people as they're programming. And, you know, we like to say, okay, we want to help people save as many key strokes as possible. And, you know, we're going to have a lot of effort into like, okay, how often are people accepting the suggestions and how many characters have been saved? And it turns out actually the way to optimize that is to get quite short with the coding suggestions that you're making because if somebody's going to, you know, then they'll accept it and they won't change it and they'll move forward with it. But when we did that, even though our like metrics went way up, we got a lot of fuss from people who were like, wait, where are those really magical long suggestions that were showing up. And we figured out that like, that's an important part that being able to see things holistically and think about that was important for people to get things done. And so like, really leaning into that magic and what does that mean and how are you seeing things differently is important. Well, one of the things we're definitely going to have to do is, you know, kind of feed this transcript or some portion of it into GPT-4 and say, give us a poem or a sonnet, summary or something. Feed, I can help with that. In work, there's value in taking time, but also in short, intentional bursts. New perspectives can inspire, we find, by changing places, problems seem reversed. The rapid shift to working from afar and AI's growing role are intertwined. For boat work breaks down boundaries by far and leaves old definitions redefined. Summarizing texts as poems, a new trick, can bring a little joy to meetings dry. With prompts to guide us, works finer and quick, as we learn how to ask and what and why. Work life balance must remain in sight for wellness matters in the work we write. I mean, I really do think that read has created more poems with GPT-4 than any poet has created previously. So it really can bring out people's creativity and magic and winzy. Like the future of work doesn't just have to be to your point about like brute force data. Like how we improved it 1%. Oh, we improved people's joy 10% or we improved there, you know, enthusiasm 10%. Like those could be important numbers that we're looking at. Well, and this is so important to you. All right, one of the things that's really interesting is what we want as people isn't to do less. We want to do more meaningful stuff. And I actually was remembering during COVID. I needed to take a vacation, so I took a week off, but you can't travel anywhere. So I'm laying in bed and all I did for the first two days was play on my phone and watch Netflix. And I was miserable. Like I was in tears. I was literally, and I was yelling at everybody, yelling at my kids, yelling at my husband. And then I decided to clean the house. And I went, like I took one day of the vacation and I did one room at a time and like just totally tore it apart and cleaned up. And it was so facility and fun. Like I really meant a lot to me to be able to do that. And so I think that that ability to, you know, it's not about doing stuff for us. It's about leaning into the meaningful work. That really transitions well to my next question because when we had Trevor Noah on the pod, his dream was to have four hour work days. And then we've just seen the research coming out of England. They had that small pilot where they saw that four day work weeks increased productivity. And a lot of my questions would be about, you know, was that a short term that it increased or you know, kept productivity at the same levels because it was novel and exciting. And what would that look like over time? And I would love to hear what you think. Like are these four day work week ideas smart, not smart, too small scale? How would you think about that? I actually think a four day work week is not the right thing to aspire towards, but like getting stuff done. And like sometimes that's going to mean I'm working 24/7 because it's super exciting. And sometimes it means I'm at the dog park with my puppy all day. And that flexibility, I mean, and then it's one of the things that the pandemic really get created a ton of flexibility in how we get things done. And now we need to start figuring out how to use that flexibility properly. Well, I think it also highlights that different people might need different things. Some people need flexibility because they have a dog. Some people have kids. Some people might have a disability that means that they need to work a different time of day. And so how can AI help us smooth that out? Because there are some problems with asynchronous work. There's benefits in us all being on the computer at the same time. To your point, that flexibility can help us a lot. It's not about working last or only working four days. It's figuring out what is right. Like for you, Jamie, which might be different from Reed or someone else. Correct. But if we weren't all on the call right now, we wouldn't be talking. So you're exactly right. Like the flexibility matters. But we're social animals. We live, we collaborate. We work together. And that matters as well. And that's been one of the real challenges is figuring out how to balance that flexibility. So I think it really starts becoming about setting us up to succeed when we're together. And AI can help us with that as well. We're doing a lot of research in different, for example, how to co-optimize schedules. So that you can both maximize your own personal preference, your own flexibility, and maximize our joint needs as well. We've talked a lot about, again, the sort of pandemic. How short in these questions, AI, how short in these questions, what's it aspect about the future of work that people are talking about? Do you think they should be talking about more collaboration? And I think we really should be thinking about collaboration more. And that was a big part of the pandemic. But a lot of remote work, a lot of what you do, while your remote is actually your individual work, you're very good at getting your own stuff done when you're at home. Going into the office is about collaborations, but unlocking other people, getting new ideas and so in some ways that is a me versus me balance going on there. And actually, I remember this when I first went back to the office a couple of days a week. Like, I'd be like, oh my gosh, I got so much less done. I've got a pile of emails I haven't responded. I didn't do anything. And it turns out I did a lot. It just wasn't the same kind of work and I had to figure out that balance. And transition again from like, oh yes, I was super productive me to also thinking about the way. And I'm sure likewise going to do that with AI, like how am I making it super productive? And how am I making it super productive? Well, one of the things that I think we already touched on some, but I want to come back to a little bit is, you know, most often people tend to think about work entirely on the outputs and efficiency, which of course is an important part. but they don't tend to talk as much about like the joy and the engagement, you know, because by the way, those have longitudinal, they have the endurance, they have, you know, staying in a job, you know, having those, those weeks that are six day weeks versus five day weeks, or seven day weeks, because you're just like, it's crunch time and you're in it and you love it. I'm just coming off a lot of like seven day weeks. Exactly. Well, thank you for, for, for spending your, your dog park time with us. You know, what are some of the things that are kind of is, as people think about the kind of architecture work and so worth is what are the being going to be some of the angles to think about these broader variables? What do you think some of the other variables that are going to be important for us to think about, you know, kind of work design and and team design and, and you know, that kind of stuff. Good. That's a great question. And actually you make me think very much of measurement and how we understand what people are doing and then you sort of started that out that way. We at the moment have very naive measures of productivity, you know, the best ways we measure productivity will be like, oh yes, number of keystrokes saved or the amount of time saved or maybe the number of emails you sent, which is certainly not something we wanted people to be like trying to increase or optimize. What really matters is getting useful stuff done and our ability to get richer, better measures of what we're doing and it relates directly to the goal directed AI to like sort of getting, translating from sort of easy to measure numeric outcomes to meaningful outcomes and large language models are amazing for that. And actually, we think of some data that we found in a recent work trend index that we have where we found people think they're being very productive. This is particularly during the pandemic. People are super productive and pay a lot of attention to the remote work, you know, to the work that they're doing. Whereas employers actually are concerned about that and and, you know, I think it was 87% of employees reported that they were being productive at work, but on the opposite end, 85% of leaders said that the shift to remote work was making them question whether people were being productive. And I think that what partly has to do with the wrong measures. And it may even be like, oh, I'm being very productive. I'm stressed. I don't know what's going on. I'm doing lots of work as I'm working hard, but I'm not producing the important outcomes. And as we beat, as we increase in their able to understand and measure those outcomes, we're going to be able to be more productive and get things to matter them. It's Ryan Roslansky from LinkedIn. For me, the most significant fact about today's work environment relates to the rapid pace of change. LinkedIn data shows that on average, if you look at the same job listing in 2015 versus 2023, 25% of the skills needed to do that job have changed. And it's apparent that AI will accelerate this even further. So even if you aren't changing your job, your job is changing on you. Thus, success for individuals and companies over the next decade hinges on adapting to technological advancements like AI and embracing a skills first mindset. By understanding our existing skills and identifying those needed for the future opportunities, we can navigate these changes effectively and better thrive in an ever shifting economy. The pandemic opened up so many types of work that some of them were positive. We realized how many types of work were knowledge work. We realized how many things could be done remotely. We gave people flexibility. Like there were some silver linings that came out of this global pandemic, but we also saw so much burnout. And so it seems as if that burnout has stayed with us. I think it's like an all time high 42% of workers globally, a report feeling burnt out. Like why hasn't that abd like what could we do about that? Like do you see anything about presenting burnout or sort of returning to where we were before? Yeah, well, things have been changing so much. Like I'm like, didn't we just deal with a crisis and work with the mood to remote work? And now we're dealing with a major reimagining with AI. Like it is tiring personally. The thing that that I found was actually was not just as we was leaning into the work and the opportunity when when I think about like all that's been happening, it wears me out when I think about the opportunity with AI. I'm like so excited. We started building all of these cool AI and all of a sudden I'm working harder than ever and like really excited about what I'm doing. There is real challenge and we have to figure out how to deal with all the disruption. Something like 53% of employees care about their health and wellbeing at a higher level than they did before. Like that's a strong signal. Just like the data you write, like there's a strong signal that we need to be helping people figure things out. We're not our best selves when we're burned out. We're not our best selves when we're stressed. It is when we are relaxed when we are safe when we understand the world that we're able to do great things. And one of the things I think we're going to see through these various chat pots is helping on those variables too. It isn't just going to be the filling the forms or you know kind of like help write the memo of super user or taking the meeting notes or remember the action items or get the for information or so and so. I think this is all in the in the thread of it's not just the productivity. It's actually in the human engagement in the in the delight and the fun. I think the summarizes this as a poem for example. You know making it as a rap song or put this in I am big pentameter and it can be like the Odyssey or the Iliad. You know one of the things that you know I'm doing is part of the and you know you'll get this to this part of the personalized versions of the impromptu book are like epic poems that are kind of like you know you me and I as a way of doing it. So let's do a few of the rapid fire questions. Is there a movie, "Solar Book" that fills you with optimism for the future? So I reread and rewatch things over and over again. So my answer is going to be somewhat trite but something that I've seen a lot is probably Star Trek the next generation. And I just restarted it again. I'm on episode five now. It was something I watched early in the pandemic and that I'm watching now for slightly different sort of helping me process things differently during the pandemic. It was really cool to have this like you have the small little world with all the same characters. It's you know very much and they're always like playing playing concerts and putting on plays and doing because it's made me think of being stuck at home with my four boys and my husband. But then you're traveling the galaxy and seeing all of these other places. So I thought it sort of fed my pandemic need. And then now where they are, it's just it's really interesting in the obvious ways. You know you've got data and you've got the computer and it's funny to be watching it again now where when I was watching it not the long ago. I'm like oh my gosh this is you know oh yes Star Trek actually has all sorts of interesting human computer interaction affordances that they sort of are bringing but the computer stuff is all made up to really be rewatching and be like oh and we could do better than that if when you get your impromptu book a Star Trek episode is also a prompt to yes and prompt in there. One of the things I did is when I was in undergraduate Stanford is actually taught a one particular like one class not a quarter class but a one class around one of the Star Trek episodes the measure of a man which is the one with data on you know what it did and he had to follow an order that was about dismantling himself and I thought it was a it was a great encapsulation of a whole set of different issues so I literally you're the asynchronous work right go watch it and then we're gonna pose some questions and talk about it so awesome I love it all right so question two obviously future of work AI super top of mind for you but where do you see progress momentum outside of your industry that inspires you for the future so nothing is outside of that right now well shit where when you man is true that being said I'm probably I'm really excited about education right now and what that means and you know it's because the questions that we've been talking about this whole time there's the fundamental questions of like what does it mean to exist in the world what does it mean to make meaningful contributions and I think that's really interesting and and certainly education was deeply impacted by remote work and it's going to be deeply impacted by AI as well and I think we're going to need to do a lot of deep thinking as well and about these new skills that we want the next generation to learn in ourselves to learn obviously you're paying a lot of attention to the technology and like AI and impact on the work and impact on research and impact on all the rest is there any technology and maybe it's also AI is the is the answer but is there any technology you're watching to make sure it stays on course to make sure that you know the guard rails are there and so forth and and maybe AI is the answer in which case what are the what are the initial and the initial amplification where the guard rails are paying attention to I think the thing that I'm particularly excited about anxious about and paying a lot of attention to right now marries too deep interests of mine so I have a PhD from the MIT AI lab but I actually got it studying information retrieval and right now there's this really interesting opportunity that we talked about a little bit to like take these reasoning engines and couple them. with grounding knowledge and building on that. And that's actually, that's really leaning into the language models where they're different from us. Like that ability to go and search engines are amazing and the ability to index all of the world's knowledge and all of your emails and all of the information that I've access to in my guess, we start capturing more index and surface that. Use it is really interesting. And so I'm excited about that. - Well, we'd love for you to leave us with one final thought. So we always say, you know, what is possible to achieve if everything breaks humanity's way? So in the next 15 years, like, where do you think we can be if everything breaks our way and what's the first step? How do we get there? - Oh, well, I hope to be exploring the universe, Star Trek style, although it doesn't need to, it can be, I can just be on this planet. And I think the first step is really leaning into what makes people think well? - Keep that curiosity off and we'll be okay. - Well, I think those are amongst the index of skills, you know, we're ranging from empathy to, you know, kind of all the other ones we were talking about, you know, collaboration, keeping your curiosity up is gonna be one of the skills that our AI co-pilot's hopefully gonna help us with. - Yeah, for sure. So far it is. - Yes, Jamie, as always, thank you and thank you for after many seven day weeks coming on to talk with us. - It was my pleasure. So what I loved about talking to Jamie on this auspicious day in particular, the day that, you know, Microsoft launched M365 and like is literally showcasing how AI can be a co-pilot for your job, for your profession, for making things better, is she is a creator, but also a consumer. I mean, throughout the pod, she was like, "Oh, you asked me that question. I popped it into GPT4. This is what they said." You know, she talks about how every email or every document she's getting it, she's getting bullets, she's getting a summary, she also mentions she's getting it in poem form. And I love that she was using her own products to actually enhance her own productivity, but also didn't forget about the whimsy and the fun and the joy, which is often forgotten. So I loved that she mentioned that as well. - Typically when you go to a research scientist who's been researching work, you're thinking you're gonna get productivity X or productivity Y or like, you know, you find that you're all you intervals at eight minutes, it's followed by, you know, one minute rest breaks. But the fact that it was like, well, actually, in fact, here's how I keep myself creatively engaged and here's how I bring the light to it and we need metrics for those things too. And the way I do it is yes, we've just launched all these really cool co-pilot summary features, those things, not just were important kind of imaginations in the future work, our imagination and, you know, the use of GD4 and tools, but also the things to do, both as individual workers but also as designers and kind of creators of the future. And I think that was part of the unexpected delights of this interview. - The two of you also spoke about, which I thought was so good, was moving from the like, I and the me, like, my work is gonna be enhanced, my work is gonna be better. Well, what about our work? What about the work we're doing collectively? Like, I just remember when you're, you know, you're a junior in high school and you get assigned to group project, you're like, oh, this is the worst. I have to work with them, like, I'm so much better and you don't quite realize that you're, no, you're not good unless you can work on a team. And so that's just so critical to the future of work and how can AI enhance that teamwork and not just you as a solo contributor and that's so critical for what we're doing? - Yeah, it was one of the things that was reminded me of my first book, The Startup Review 'cause it's like life is a team sport, not an individual sport. It's actually one of the things I think the educational system through it's, it's like, you know, how do you measure individuals and so forth, gets fundamentally wrong but actually in fact, every class should have group work components because almost all the work, even writing a novel is a group work thing, right? So that group work and the fact that look at the tools not just as individuals like, you know, I am here with my trusted co-pilot sidevig. So we are here with our co-pilots and it's the we about how we're working that is I think such a valuable lens into the future work. - Yeah, and it's like how can the AI be a trusted team member and co-pilot for what we're doing? But again, for the whole team and maybe you have a different LLM for what you're doing personally, you have a different one for your group, you have a different one for the group over there and so again, just remembering that I think is really great. (upbeat music) - Possible is produced by Wonder Media Network, hosted by me, Reed Hoffman and RA Finger. Our showrunner is Sean Young. Possible is produced by Edie Allard and Sarah Shleed. Jenny Kaplan is our executive producer and editor. Special thanks to Caitlin Cummings, Katrina Zuccaro, Lauren Cole, Surria Yalimoncelli, Saeeda Sepieva, Ian Alice, Greg Beato and Ben Rallis. And huge gratitude to Papia DeBroy, Simone Stolzof, Ryan Erzlanski, Jesse Hempel, and everyone who called in with their thoughts on the future of work. Thanks so much. (upbeat music)