384: When AI Creates Art, What Stays Human? with Jake Saper, General Partner at Emergence Capital

We did analysis on companies in kind of the B2B software realm in 2022. It took the top quartile of those companies 18 months to go from zero to a million in ARR in 2025 is going from zero to one million in that quartile comparison is shocking. So it is the most exciting and perhaps the scariest time to be both investing as well as building. Good morning, good afternoon or good evening, depending on where you're listening. Welcome back to AI in the future of work, episode 384 on Mirrorhost Anturche and CEO of PeopleRain, the AI platform for IT and HR employee service. Our community is growing. If you're not yet subscribed to the weekly newsletter, why aren't you? Publish about 90 of them and they contain some additional fun facts and some clips from the episode that don't always make it into the full weekly conversation. We will of course share a link in the show notes. We're all in this together. If you like this conversation or any other, it is your duty to share it with at least one friend. They'll know you care about using AI to make humans better. And you know what, it's good karma. If you leave a comment, I just may share it in upcoming episode like this one from Tamra in Mill Valley, California, just up the road from where I am now. Whose favorite episode is number 25 from the Waywayback Machine with Bill David Owl from more David Owl, one of the founding fathers of venture capital, talking about the autonomous revolution. We learned from AI Thought Leaders weekly on the show. The added bonus of course, you get one AI fun fact. Here it is for today. The news team at Emory University writes about the future of creativity at the age of AI. From Taylor Swift's breakup songs to Rumi's poem celebrating love and beauty, human creators draw actions, producing works of art that unlock self-expression connection and healing. When asking if AI models can be similarly creative, the reductive answer seems to be no. After all, a machine doesn't have child and memories, love affairs or a physical body that feels pain. An AI be truly creative. If it lacks the very things that makes humans human, many creatives wonder how far AI will encroach on their livelihoods while it simultaneously trains on their novels, drawings, animations and music, Gwendolyn Reed, an English professor, Amory says writing is a deeply human act and it will be challenging even painful. New writers don't realize these moments of doubt are normal and may fear something's wrong with them, but they're necessary to produce something unique. It's a great read. Of course, we will share a link to that full article in the show notes. My commentary, all art throughout history has been influenced by external sources. Michelangelo was inspired by Greek and Roman sculptors Paul McCartney was inspired by Little Richard and Elvis Presley. New works always derive from some combination of old works. It's always hard to reveal the hidden layers. AI accelerates the process of artistic creation and destruction, but how it makes us feel, regardless of the offering process won't change. We experience art when a song makes us joyful or sad, a meal reminds us of childhood or a photo helps us explore a new place. As humans, we get to judge how all forms of art connect with us, give us meaning and inspire us. AI is pattern matching at scale and it operates an unimaginable speed. Let's use this complicated phase in human history to rethink our relationship with art. Let's be open to judging all art based on how it makes us feel, rather than by who or perhaps even what created it. Now shifting to this week's conversation, I've so been looking forward to this one. Jake Safer is a general partner at Emergence Capital, one of the most iconic and successful venture investors in the world, having backed legendary companies, including Gusto Viva, Together AI and Zoom, among others, Jake led Emergences 2024 Investment in Mechanical Orchard, Pioneering AI Services Company. In 2016, he co-authored Emergence's thesis on AI coaching networks. Jake started his career in venture at Clienter Perkins. He holds a BA for Meal, Go Bulldogs, Animasters in Environment and Resources plus an MBA from Stanford, Go Card outside of work. He's a proud girl dad, times three, and a musician with a passion for exploring the intersection of creativity, art and AI. We had a version of this conversation, a few weeks back at a dinner event and it so inspired me that I insisted Jake come and we continue the conversation with this with this listener audience. So, Gus Jake, welcome to the podcast. Without further ado, let me share a little bit more about your background and what got you to where you are today. Thank you Dan, thank you for the kind intro and I am grateful that you invited me to follow up this conversation. So, my background, I grew up in Austin in the sort of early tech era of Austin when kind of Dell was getting off the ground. My parents are serial outconeurs, they started a bunch of companies together which meant that I sat around talking about startup strategy around the dinner table when I was in middle school. So, I come by this venture capital role, this company advising role quite naturally. I was a singer growing up, I still am a singer but I kind of found my passion for it as a kid started classically singing when I was a kid and then saying in musicals, I sang in acupeller groups, I have performed in a bunch of different bands. I think I have performed in like most possible genres. I did a metal show two weeks ago, I've sung a lot of country, I did blues a few months ago. So, I am an intrepid explorer and experimenter along the musical domain. I play a little guitar and a little piano as well but I'm mostly a singer. But on the career stuff, I did consulting at a college, I really like frameworks, consulting was a very good fit for my brain. And then I did an entrepreneurial journey, I helped run a startup that was building solar power plants in India and Africa. So, a very exciting and difficult challenge in the late 2000s or early 2010s based out in India and then as you mentioned I came back to California for grad school and then started my career in venture and it had been an emergency now for going on 12 years. I feel very grateful to do this job. The very first deal that I led, Diligence Song, when I joined the firm back in 2014 was our investment in Zoom. So, I'm glad we're using the product today, big fan of the product obviously. And also taught me a lot about the importance of not just being right but being lucky. And I think there's an important amount of humility necessary in the venture industry that isn't always present and having that experience so early in my venture career. I think it's really ingrained that realization in me. I have one important ask, this is way off topic, but ordinarily when we publish the episode, it will contain some syrup, the professional headshot of you. Can we please instead replace it with a picture of Jake as a head banger in any chance makeup from two weeks ago. You can take it from that show. You can take it however direction you want. For that show I did wear eyeliner. I had my wife apply eyeliner to a much to Hersha Grin and my daughters, two of my daughters came to that show and were very confused. My mom was putting makeup on data going into that show. Oh fantastic. Well, if you can jibble fine image then I'm sure you can kiss a fine image. Exactly. Exactly. Brilliant. Well, I think I'm somewhat obligated by the cause most to ask you a question about venture capital and AI investing. But as I mentioned when we were chatting before recording, we'll quickly move off from that topic. But genuinely, I'm interested to know. I shared that you authored an important thesis for Emergence back in 2016. Does AI first, everything make venture investing kind of venture formation company financing harder, easier or is it more nuanced than that? Well, from a company's formation perspective, it's much easier to build a demo. It's much harder to build a mode. Right. That it just collapse the cost of building something that looks good. So we're seeing a lot more me two products, a lot more noise. You know, used to write code. Now you kind of orchestrate APIs and agents in natural language, which is awesome. But it does change investing because a slick demo is cheaper and it's therefore less informative. That being said, obviously, this is a transformational moment. There are not just new companies but whole new business models that are being launched. You mentioned Mechanical Orchard, which is an AI enabled services business. So they're selling an outcome versus software. This is an entirely new business model that's being born. And just like SaaS came out of the last technology wave of on-prem to the cloud, there are new business models that are going to be created by this technology wave of the rise in AI. And I find that very energized. And very energizing. But yeah, my job is to sort through that and founders job is to figure out how you can differentiate yourselves. I think one other important thing to note about this kind of generation post-genai of company startups and startup building is the rate of growth has changed meaningfully. So we did an analysis on companies in kind of the B2B software realm, which is where we tend to be focused. In 2022, it took the top quartile of those companies 18 months to go from zero to a million in RR. In 2025, the top quartile of B2B software companies is going from zero to one million in four months. And this correlates with what we're seeing anecdotally. You know, we right before we went on break, we had three companies coming in on Monday and present that it all gone zero to three million in RR in a year. Four years ago, that would have been mind-blowing growth. But now, in part because there's so much demand from buyers, in part because the technology then locked things that weren't possible before, a lot of these companies are able to grow faster. And there's lots of open questions around gross margin, defensibility. One of the downsides of the fact that these companies are growing so quickly is that when they go to raise their series A, they often don't have renewal data because they haven't been selling for over a year. So if you go zero to three in a year, but you haven't actually gone through an annual renewal cycle, it's tough for the investor and it's frankly tough for you as the CEO to know how to pile up the ship before you've got, you know, hard data on whether or not customers are going to renew up sell, etc. That quartile comparison is shocking. That would imply that not only are there an increasing number of companies, because what you said about venture formation and rapid prototyping, but also the number of quality companies is increasing if you base it on time to, time to pursue an annual 2 million. Well, what's interesting actually is we also looked at the number of companies that are raising series A's and in 2021, which was kind of the last peak prior to the interest rate reset, you had lots of companies raising at high valuations and actually a higher number of companies raising series A's on a just absolute quantitative perspective. Over the past 12 months, that number of companies that are raising looks more normalized to kind of where it was in the 2019-2020 era. So what that means is that there is a lot of venture dollars that are chasing a relatively smaller amount of pool of companies that are actually able to successfully raise the A than there were back in 2021. So those numbers I shared, I was talking about the top quartile, what it means is that the delta between top quartile and bottom quartile is actually grown in this post gen A era and as a result, the number of companies that are successful raising series A's is actually been static relative to where it was kind of pre-COVID. Annic, totally. Are you finding it harder or easier to get to source high quality deals? Well, this is interesting because it gets into how AI plays into our job. So sourcing is obviously an important part of our job and we are using AI to do that. So our coverage ratio, in other words, the number of companies that we see and meet with in our domain before the series A is at an all-time high. I don't know that lay this number but I think it's somewhere in the 80s. So we see a very high percentage of them. Part of that is the improvements we've made operationally internally. We have now someone who is in charge of investment operations for us. The same way you have a rev-ops person in a company, we've actually created investment ops function to make sure that we are excellent at this. And we're using all sorts of data tooling and AI tooling to make sure that we're really on top of this. So I think we're seeing more than we ever have. I think the question really comes down to taste, which is this qualitative thing. But I think it's an important part of our job that I think will likely endure for some period of time. The reality is a lot of the static parts of venture capital will be automated away by AI the same way that many other parts of other jobs will be as well. Satic parts of other jobs. But the judgment part, the trust building part, the taste part. In some ways, the importance of that goes up because the alpha you get on sourcing likely declines over time. So the fact that you're seeing north of 80% of the deals that are, let's say, a stage fit, emergence. I would imagine leads you to form maybe more conviction or kind of solidify your investment thesis faster. Is there one pitch if it walked in the door today with the right combination of team and kind of ability to execute? One pitch that you would fund before you let them walk out the door. Well, I, as I mentioned, have become a big believer in this AI enabled services concept. I think it's a disruptive new business model that can capture a much larger tam than software because you're explicitly going out to labor spend and not your software spend. There are lots of open questions on how to build these businesses. I've become really excited and become a student of this new model. I wrote something called the AI enabled services playbook, which is attempting to diagnose what is working and what's not working in these types of companies that we've invested in and the ones that I'm just studying that we're not on the board of. So in general, I'm always curious to spend time with these businesses because there's like an academic interest, but also obviously trying to show that learning with new companies. We have made AI enabled services investments in IT as you mentioned mechanical orchard moves main frames into the cloud using AI. We have done some AI enabled services investing in the finished services space. I'd say like areas that we haven't yet invested in, but I think are probably right for AI enabled services businesses. One would be accounting. I'm very curious. There have been a number of attempts, but I don't think anyone's really cracked them not yet on building the Accenture Deloitte accounting function out using an AI enabled service. There are some attempts at doing this by sort of a private equity style model of acquiring a bunch of companies and then trying to push AI into those companies. I tend to prefer the more organic approach where you start Denova with a tech platform and you build a book of business yourself and maybe you acquire something down the line, but once the tech is truly established versus trying to kind of smother it onto something that already exists. So I would say to answer your question in a shorter breath, if there were a team where you had someone that was hyper-critical credible from the accounting space, someone that was a hyper-critical credible kind of go-to-market. It was a person who had sold in this type of way before and then you paired that duo with someone who was world-class and understanding how to capture the greatness of all of them. I'd be quite interested in that team. Okay. Entrepreneurs listening. You'll notice. That's the playbook for getting a meeting with the great Jake Safer. But the conversation is just starting. Jake, I shared that fun fact about AI and creativity and just intentionally because it was a bridge to the conversation you and I started in person recently. Just starting open-ended. What is art to you? Art is compressed human experience. It is an attempt to transfer a way of seeing, a way of experiencing life generally or historically from one nervous system to another. I think that the act that art is the act of making meaning shareable. It's not just the artifact of song, the painting, the film. It's this relationship between the creator, the context, the audience. The same melody can be a jingle or it can be a prayer depending on intent and who hears it and when they hear it. There's a second layer. Art is also a form of identity and belonging. We don't just consume it. We use it to say, "This is who I am or this is who we are. I am as 50." Or I am definitely not as 50. In my case, I am. That's social layer is why provenance and story matter so much. I think about this a lot as a singer, like as a musician. The notes aren't really the point. The point is, "Did it move someone?" Was there some connection created between you and the audience or between the audience and whatever I co-created? That definition you can't separate Jake the artist from Jake the parent or Jake the venture capitalist because part of you that radiates humanists and makes connections that transcends you on stage versus you and a sweater vest. I hope so. I try to think about living my life holistically because I think that I want to build a connection and build trust with an audience and have communion there the same as I want to do with founders, I want to do with my kids. It's really just like a different instrument. I guess that I'm using in those different contexts. Yeah, that's. I was kind of paving the witness, but yeah, that's how I obviously. You know, that's how I think about it as well. So, the question arises, this one's kind of, kind of an obvious, but got it, got it thrown out there. Given how we both define art around how it makes us feel as humans, can AI produce our. It's a really hard question. I would say that AI will democratize the creation and commoditize the execution of art. So, the scarce thing, actually interestingly, thinking back to venture, the scarce thing will become taste, intention, trust. AI collapses the cost of iteration. It can turn, I can't draw, I can't write, I can't, whatever, into, I can direct, which I think will unlock a lot of latent creativity inside a lot of people. But it also creates this flood, or you could call it a sloppy flood, if you want to use the term to sure, when output is cheap attention, it becomes the bottleneck. So, I think the enduring impact is a shift from raw production to curation and point of view. The most valuable creators will be the ones with distinctive taste, authentic intent, and a relationship with an audience, not just the ones who can execute. And it ties to your earlier point, which is that since humanity started creating art, it's always been derivatives, right? But the value has been in the Beatles interpreting what some Blues artist had done and curating that and injecting whatever their point of view is and then sharing that with the Beatles audience. Today's episode is brought to you by my company People Rain. As you hopefully know, People Rain is the AI platform that automates work. Tens of millions of employees around the world get back four to six hours per week, because People Rain automates away the friction. Imagine a world where your teens never have to wait for technology upgrades or have a human reset their password or get added to a distribution list or get help with payroll, PTO, or benefits. Inbox zero for the help desk, zero friction for employees, that's the future of work. That's People Rain. We're offering all members of team human. And hey, that's you, our listeners, an exclusive discount on People Rain just for being part of this amazing community. Get the discount code at peoplerain.io/podcast. That's People Rain. Like the Rain of a King or Queen.io/podcast. Now back to this week's conversation. One of the things that is fundamental for the human condition is figuring out what how we drive meaning and purpose. And certain segments who fear the Bata-Pop lips or fear what AI might do are concerned that AI may extract the humanness out of how we drive meaning as a species. How concerned are you that we're both girl dads, U-Temps 3 and U-Temps 2, and we think about the world that our kids will inherit? Is AI likely to sat meaning that maybe you and I enjoyed as the ages that our kids are? My hope is that AI will help create space for deeper meaning. Graphs in rare moments, if that makes sense. I think that, let me unpack that a little bit. So in one case, if one of my kids, for example, doesn't feel like she can write beautifully, write beautiful poetry or play instrument very well or what how do you. But she still has a unique perspective. I need to experience on life that I would love for her to be able to share with the world so other people connect with. Working with AI to get that out of her I think would be really beautiful. If AI can act in a truly symbiotic way to figure out what is the authentic truth of this person? I think that that's really powerful. I mean, I give you a anecdote that happened to me yesterday. I was working on my 20s, 26 goals and I was working on it in chat GPT and I was kind of iterating back and forth on it and then I asked chat GPT to give me really harsh feedback about myself. Based upon everything it knows about me on the internet, about what I've shared with it in chat. And actionable but like no holes barred, where am I falling short and how can I get better? And I would say I was reading the output to one of my best friends this morning at breakfast and 75% of what I read to him really hit me in the gut. Like it was real. It was a really insightful take on me. 25% was sort of like reading this trilogy thing where it's like that could be anybody but 75% was like no that's real that's me. Which is so powerful. We're only in the beginning of 2026. Like how insightful can this stuff get to understand us as people? And if you translate that back into the art format, what if I was able to partner with chat GPT to take these hard ugly parts of myself and translate that into art so that if you Dan are struggling with some similar darkness, some similar stuff that chat GPT has helped you on earth, my being able to translate that out into the world would resonate more with you and create a moment of meeting that otherwise wouldn't have been possible. There's this spectrum in terms of ethical approaches to AI on one end of this spectrum. Just free reign. Everyone should be free to use AI in whatever way they choose and let beauty's be able to be holder. Let us co-create with AI and let the public market so to speak decide if who we represent ourselves through AI or not is representative of someone or some being they want to associate with. On the other end, just radical levels of regulation, disclose, it's not allowed, it's cheating if you use it. Those are two extremes. But where's Jake's in terms of what's at a societal level? I don't mean in a classroom necessarily, but at a societal level, how heavy-handed should society have opinions about what's the appropriate use of AI? Yeah. I'll set aside the economic argument as well because we're not talking about should you pay the artist or what have you? In general, unlike the artist to get paid. But in terms of from a societal perspective, do I care if consumers know if art was generated or enhanced by AI? I would say yes when it's material to trust or the claim being made. I would say I'm generally pro-transparency, especially when there's a claim of authenticity or where trust is at stake, journalism, documentaries, educational material, commissioned work, etc. If it's purely entertainment, I'm less dogmatic. We've used tools forever, cameras, did a bit of Photoshop, auto-tune. The idea of being augmented by technology has existed in entertainment for a very long time. But in general, I think that default should be to move more towards disclosure as lightweight metadata. Where are people who care to know and for creators to be honest about the process? Because I think that the context matters in terms of how you experience art. We don't experience art in a vacuum. If I tell you a song was written in a hospital room versus written in a recording studio, you will experience that song differently, even if the waveform is identical. Humans are meaning-making machines. Providence changes meaning. I think over time, as AI becomes a normal tool, we will care less about what AI involved and more about what do the creator mean and did it move me? But in this transition period, I think disclosure changes interpretation, and I think that is important. But this almost comes back to self-awareness in some way. My hope for the future is that our daughters have more self-awareness around, did that genuinely move me? Not was this popular on social media or even did AI create or not created? It's like, did that vibe with me? I think we've used this term vibe coding and vibing a lot in the past 18 months. But what it used to mean is, did it resonate? I think this concept of resonance as a singer is very important to me. When I go into a space and I hear resonance, that's like, you know, singers love resonance. It makes a sound better. It's like we're better at singing. But like, I would love a future where our kids were more attuned to what content, what art is resonating with them. And I care a little bit less about who or what created that art. I frequently say more cynically than I choose to be, but we're entering a post-truth era. Where you have to question the authenticity of everything. Almost to the extreme that in the process of questioning everything, you assume that everything is not authentic where the definition is. not authentic is you don't believe who you were told created it actually created it and you're not sure if it matters You know and and and you just throw up your arms and say you know Everyone's cheating or everyone's using AI and that's not called cheating What's I was trying to take the positive on these things, but what's your perspective on what? What is what is truth in in kind of a ubiquitous AI age? Well, I think the secret to happiness in life is low expectations So if you go in thinking you know that everything is you know has a shade of You know not quite what it seems then you're not going to be disappointed I mean to me this is like also where art differs from like Journalism documentaries political Communication etc like I worry a lot about truth in those contexts like I still think it's It's created up to say this but in 2026, but I think it's very important that we as a society have a agreed upon definition of what is true from a factual perspective like did this event happen or not That feels critical to successfully functioning society So we have to do everything we can to maintain that and There's technological and political and cultural changes in the intake place and into ensure that When it comes to art and I mentioned before like I do think that disclosure is important Because I think it does change the experience of consumption But I do think like over time as it things get a little more ambiguous on kind of art and the cyborg nature of art creation I think again, people are gonna care more. I hope people care more on like did this move me like that Did I was I moved by this genuinely and not was someone else moved by this or should I do move by this but like did it like you Because you know in your gut when you listen to something is like whoa Or when you read something or when you see something and you're like god damn that was amazing I hope that we can tune into that voice more clearly in this era The other thing I would say we talked about the dinner. I believe that As AI becomes more ubiquitous in art The value of live performance is going to skyrocket I'm already a big fan of live performance not only because I do it myself But also like I love going to shows my wife and I love going to music theater whatever um I just think that's gonna become such so much more valuable great. There's no there's no question there when you're seeing the person on stage Performing there's Millie Vanille for those of you who remember Millie Vanille there have been you know There haven't left thinkers in the past but In general live performance is about it's happening in the moment Hard one makes it special is there is the uncertainty and like oh god. What could have anything it happened? This is so special. It's not pre-recorded. It's not art or it's not AI um And I think like that's gonna feel more and more magical um Which like unfortunately makes me more long live nation despite you know the The questionable nature of some of their business practices Apple tried to convince us otherwise right that you could put on 10 pounds of headgear and Experience the ludicrous concert right but there's something about the I will tell you and I because I've experienced it both as an artist and a and an audience member, right like There is a dynamic connection between those two things in live performance I mean like even in this conversation like if we both had spun up a eyes to have this conversation I'm responding to you shaking your head right now or nodding or whatever and you are responding to my wild gesticulations that the listeners can't see Um the same is true when I'm on stage like I am leaning into like seeing I In the metal show I did a few weeks ago my youngest daughter cried basically through the whole thing Because it was loud and it was late and I probably shouldn't a broader But like that impacted me and like at one point I'm kind of like talking to her and I'm moving and like that is changing the show and that's a silly example, but like live music By definition is dynamic. I mean even the eras tour Like even Taylor's tour which arguably is the most produced live performance in human history There were still moments with spontaneity There were still some moments you know of like what's gonna happen? What song is she gonna play? And I think in some ways like that made it even more magical and special for the audience In dad world did just give me chills because just think just think about you know in a decade Your daughter's getting married. She's you're sending off the cut whatever it is, you know and and she's Remembering that moment of connection that she had with you That's a human experience Right, I hope I hope it's not the uh the screaming and crying during the metal concert that she remembers, but uh They're they're half your moment so hopefully I promised you right memories get better over time Yeah, and you're sure the pictures, you know, you and your silly eye makeup or whatever, you know, you'll you'll you'll you'll laugh about it and When we talk about um, you know, just in our in our field both of our fields, you know, AI workers or you know Let's give them a Abortion certificate and uh, you know anthropomorphism and and pay a salary um In part I started this podcast because that just um That's just creepy um and and it doesn't achieve any any objective that either human have or you know any any ounce of my why Um, that violates um What's your perspective on what how should we bound kind of you know, how and when it's appropriate to not just anthropomorphize That's that's a little bit extreme, but like How much we think about integrating you know the physical embodiment of AI in society? I have like a somewhat optimistic take on this and my optimistic take is that I think The market will self-regulate here What I mean by that is like If enough people find it creepy that we're giving birth certificates to And naming these bots that are doing work for us Then that's not going to be the winning strategy Because like a buyer is going to reject it because the employees don't like it and it'll be someone else so like I'm a capitalist of hearts half my Fessional title like I think that that will correct and I think the same hopefully is true in in art like If we get to a place where Um, you know, there's a bunch of AI created artists that are putting out new music um, and you know people Dig it and they start to create you know whole backstories and there's a whole you know, you can imagine a there's a world that gets created around some You know, AI singer that gets created Um, there will be a demographic that's probably really into that and there will probably be a lot of people that are like No, no, no, gracias like I'd rather I don't want that um and that's okay too Because like people can opt in or opt out of it. So I don't know. I think that like They're actually I think a lot more about soul recently like the concept of soul and I think that people are drawn to others sole and I think that As there are more and more soulless entities that are in our work and personal lives They will play an important role And also the human desire or soul to soul connection Will increase I think we take it for granted for now because we don't really interact with that many robots yet But over time you're gonna miss the mistakes and idiosyncrasies and smells and weirdness of another human That's such a satisfying answer and just philosophically you and I agree the corner cases are Do we have to anticipate when that Capitalistic approach could harm humans do we have to wait for the market to tell to decide that it was bad after The robot surgeon kills a patient or after it crashes a plane or after it You know uses bias dated to deny alone and Mike yeah just capitalism would say like that's how market forces work is let it screw up and then we'll decide it was bad But I don't know if this is this is complicated My my default my reaction to that like my answer was really thinking less in the context of activities with Important like physical or economic consequences and more back to that kind of like Creativity vein like in creativity. I think we can afford a situation where we over rotate in the direction of a soulless entity and society as the self correct Um, and I think within companies I think as well like I was thinking in the context of like you don't want too many Um, you know named Entities with bursters biggest motor because it's creepy and weird Um, but should we have guard rails around like does you know, Ken agent autonomously perform task xyz without checking or the person Absolutely like we don't want planes falling out of the sky. We don't want um, you know massive fancial catastrophes etc etc And so I think like one of the big tensions likely in the next couple years of um In the tech industry as the agents get more and more capable is Bridling up against whatever guard rails are set either by companies or by legislation I don't think we're really in that area that because the agents aren't really that good yet Right like if you've tried the agents like they often break their brittle But theoretically over the next 1824 months they will get less and less brittle which will make people who are building them more and more likely to let them do bigger and bigger tasks, which is where guardels become more and more important. So my rough thinking is kind of like, it's sort of like a, put some guardrails on this, or some regulation, if it's something that could cause catastrophic harm, right? Someone dies, some massive economic issue. But if there isn't a high likelihood of catastrophic harm from the agents, then let the market figure it out. - Y'all in the Koon who recently left meta has a somewhat satisfying answer to this, which gives me hope if we as humans are compliant, but I don't know if we are, and that's that if we're brilliant enough to create the AI, we're brilliant enough to create it in a way that's safe. - So Jeffrey Hinton, who obviously has lots of disagreements with Y'all on this topic, I was listening to a talk he gave recently, and he is less optimistic than Y'all, and obviously he's become quite outspoken on the risks of AGI and ASI and such. Something he said, a framing he said that I thought was really interesting, was we should be building, basically, he takes a regard that the AI is going to be smarter than us very soon, and basically we'll be able to control us if it wants to, because once it crosses that threshold, it can build itself, it can take it exponentially, more capable. So he's like, "What do we do if we know that's coming?" We're obviously not gonna stop developments, 'cause that's a problem we can't coordinate, right? If one America stops, China's not gonna stop, so he's like, the best we can hope for is that we build this technology like a mother. In other words, we try to imbue it with maternal instinct, and have it view us as its children. So views us the same way a mother does, which is as a lesser being, a weaker being, that needs to be protected and cared for and nurtured, and supported, which I thought was both comforting and also scary, because the interviewer on my friend, Naima asks Jeff, "So how do we do that?" And Jeff, he was like, I don't know. So it was both the satisfying and deeply unsatisfying answer. All humans listening to that conversation, I know that great interview. I think their eyebrows raised to hear the godfather of AI kind of paused when asked for the answer to the riddle, right? Yeah, I appreciate his transparency in the, exactly he doesn't know. But it is just such a massively fundamental question. I mean, in some ways, this is why I go back into art, because when I feel kind of existential anxiety from some of these questions, I find it comforting to go back into the creation and consumption of art. And that kind of comes full circle. That's part of the human condition is we ask questions, and it's okay. It's a time when we're all vulnerable, and it's okay to be vulnerable together. And the only thing we shouldn't do is claim to have the answers before we do. Innovation is not the enemy. I think the enemy is not being willing to have the bold conversation. Yep, I agree with that. I think that this is the art is amongst the best venues to be vulnerable. And this is a moment where we need to leave. Jake, this one's been fantastic as expected. Whether you like it or not, you just became a regular guest in the podcast. I hope you find it. (laughing) I'm honored to hear that. Thank you, Dan. I don't know if that was your objective, but we're just getting started. This is really such an important topic. And I'm not letting you off the hot seat, but without answering one last important question for me. That's cool. You're an artist, you're accomplished in many domains. But what I wanna know is, when Jake's hit forgets the call to star on Broadway, what is the show? What's the role? What's the dream role for Jake? Oh man, that's a really hard question. Okay, so it's a long, I have a lot of answers here. So I really, I like play Ms. and I think playing Jean-Glaire would be amazing. I also like Jesus Christ Superstar and I think playing either Jesus or Judas would be an amazing life experience. You know, I'm messing around, man. This is a, these are all roles I have not played. I did play Harold Hill, the music man and the music man years ago. And that was such a fun role. It's just so silly and fun and dynamic and crazy. So I would play that again for sure. God, there's so many. So I mean, I say to see you as Javier, I'm Broadway. Oh my. Can you imagine that? We have not watched Les Mis, yeah, they're a little too young, but we will. I think they hear me singing it a lot. So actually there's one of my favorite songs and one of my favorite songs is called Take Him Home by, it's from Les Mis. And I love singing that song. Now on my head. Yeah, exactly. I would love to perform that song. That range. I can fake it. I have a strong falsetto. And so I can fake the high notes. Well, Jake ordinarily I would ask our guests where they can learn more about you professionally, but where can we see you perform? I have a banner as well tonight actually. And the question is like, did this band actually create, like how a gig? The last time we played this band, we played on my 40th birthday years ago, which was so fun. We played it a place called the Verity Club in the mission, which is awesome. And we're working on a new concept band right now. We got to figure out if it's good enough to actually schedule performance. It's somewhat ad hoc. I have this like day job and kids and such. So like I probably do a show every four or five, six months at this point. Well, you know what, me and my kids are big live music at Fist Sinata. So I promise you that at your next show, you've got at least three tickets sold. Thank you. Thank you. You've got to support our. All right, Jake, this has been so much fun. Thanks for hanging out. Of course, thank you, Dan. All right. Well, gosh, that's more than all the time we have for this week on AI and the future of work. As always, I'm your host Dan, church in from people rain. And of course, we're back next week with another fascinating guest. Yo, it's the end of the show time to power it down. Hey, I'm the sixth for humans with a crown. Team human in the house got the vibe. So clever. Dan, church in from people rain, making us better. We learned about AI for humans. We're ready to conquer the future that's all for now. See you next time.