Latest / SPEAK LIKE A CEO / How to Lead when AI is Smarter than You with Gerrit McGowan (Video Edit)
Transcript
- 0:00Gary, welcome back.
- 0:04Oliver, good to be back, man.
- 0:05It's been a while.
- 0:06Things the fourth time you're in the podcast.
- 0:08Always a pleasure to talk shop with one of my best mates in the world and talk about
- 0:12cool stuff.
- 0:13Absolutely.
- 0:14I had the pleasure of watching you on stage a couple of weeks ago at a longevity conference
- 0:17where we talked about flow and our perception of time.
- 0:19The awesome thing is the day before you told me that you weren't sure if you could make
- 0:24it because you were feeling well, so you created an avatar to give the speech in your
- 0:28whole set.
- 0:29In case you couldn't make it.
- 0:30Yeah.
- 0:31That was fun.
- 0:32One of my friends, who's a longevity doctor, asked me to speak at this conference about
- 0:36a year ago, and I said, of course, it would be my pleasure, although I don't particularly
- 0:40feel like a longevity expert.
- 0:42She thought I had something to contribute, but then the worst case scenario could possibly
- 0:46happen.
- 0:47The hand foot and mouth went around my daughter's quita.
- 0:51Apparently I've never been exposed to hand foot and mouth because I got it, which is
- 0:56another story within itself.
- 0:58But in early stages, before I got these ugly blisters on my hands, I got very horse and
- 1:05started coughing and was losing my voice.
- 1:08This happened 24 hours before this keynote speech.
- 1:13I started to get a bit concerned that I wasn't going to be able to do it, 800 people there,
- 1:18people counting on me, and I thought, well, what the heck am I going to do?
- 1:23And I figured I will go to my roots and what I know best and see how I can use technology
- 1:28to solve the problem.
- 1:29So fortunately, because I've done a lot of podcasts, I had a bunch of content of my face
- 1:34and my voice.
- 1:35So I went to 11 labs and I uploaded a whole bunch of my podcast recordings and created a
- 1:41twin of my voice, which was quite amazing.
- 1:44I could change the inflection and the tone and the energy and the error rates with that.
- 1:49And then I uploaded a whole bunch of video and I created a video avatar of myself.
- 1:52I took my speech, I wrote it out verbatim, and then I uploaded it into 11 labs, created
- 2:00the voice recording, and then uploaded that into Hagen and lo and behold, there is my face
- 2:07and my mouth moving to this talk that I'd never actually given.
- 2:12On the plus side, I woke up the next morning and I wasn't coughing so much.
- 2:16And as you know, because you were there and I was asking for your guidance, what the heck
- 2:19do I do?
- 2:21So I was like, I guess I'm going to give this thing.
- 2:23And the most surreal thing of all was, you know, I'm at the gym in the morning, I'm walking
- 2:28around town and I have my headphones on and I'm listening to myself give a speech that
- 2:35I've never given before to train myself to be able to give that speech for the first
- 2:40time.
- 2:41I cannot tell you how surreal of an experience it is to spend three, four hours listening
- 2:45to yourself knowing that you actually never said those words.
- 2:49I was so crazy.
- 2:50How long did it take you to put it all together?
- 2:54Yeah.
- 2:55Maybe 90 minutes.
- 2:56Wow.
- 2:57Wow.
- 2:58It was shockingly fast.
- 2:59I think I spent a total of about 30 euros to put it all together and there it was.
- 3:04So I had the backup plan.
- 3:05I'm glad I didn't have to use it because I would have felt like a mime basically playing
- 3:10record on my voice and standing there.
- 3:14But we think about deep fakes and all of these kind of scary things that people think
- 3:20about when they think about AI and it was the first time I realized it's that easy.
- 3:25And if you have any kind of public persona, if there is content of yours out there, your
- 3:31voice, your face, it takes very, very little to be able to do it.
- 3:35And you don't even need 95 video podcast recordings.
- 3:40I've been working on this other project and we're creating three dimensional content
- 3:45from 2D images, right?
- 3:47So essentially being able to upload a photo and turn that into a perfect avatar of an
- 3:53object or of yourself.
- 3:55So a technology is incredible and can solve problems like the one that I faced.
- 4:00But I think there's some use as well, absolutely.
- 4:04I did the same thing actually.
- 4:06That was what probably says communications exercise.
- 4:08So what I did is to use the same tools you used, ripped some videos of YouTube off that
- 4:14head of that organization.
- 4:15It's just a big important government organization.
- 4:18We put it together.
- 4:19Elevileps the voice, you know, ripping off some video content, rating a hot mic moment
- 4:23as if this head of this organization would have said something really, really terrible
- 4:27about the crisis we were simulating.
- 4:30And of course, confronting the people at the room with it and they're like, oh, shit,
- 4:34this is super real.
- 4:36And how do you react to that?
- 4:38And of course, the first step is to find out if it's actually real.
- 4:41So the real life situations you would have to find out with the person did you actually
- 4:44say this?
- 4:45And if it's a no, then ensure the world understands this never happened.
- 4:49And of course, there's no perfect solution to this because once the cat's out of the
- 4:53back, there's something called Buntolini's law, which means that it's very easy to put
- 4:58bullshit out in the world, easier than ever now.
- 5:01And it's very hard and it takes an exponential amount of energy and time to essentially put
- 5:06the record straight.
- 5:07Right?
- 5:08Buntolini's law is also called the bullshit asymmetry principle, conceptualized 10 years
- 5:13ago by an Italian engineer, but now it is essentially even truer and even more prevalent
- 5:20than it was ever before because it is so easy to create bullshit that's super realistic.
- 5:25Yeah.
- 5:26I mean, you know, I always think about what my mother used to say to me when I was young
- 5:30and like, look nice and how first impressions matter.
- 5:33And later discovered that there's plenty of neuroscience behind why that is the case.
- 5:38And here you are.
- 5:39It's something that you never said or did actually creates that first impression for someone.
- 5:46And there's a trust, a relationship of trust or a potential relationship of trust that's
- 5:52been broken at that point.
- 5:54And especially in the world of echo chambers that we live in, once you're in one camp,
- 5:59it takes tremendous cognitive effort to move from one camp to the other.
- 6:03So if you believe it to be true, the effort to get someone to recognize that it's not true
- 6:09is it can be an insurmountable hurdle, right?
- 6:13So I'm curious how you handle that crisis, right?
- 6:17You just try to mitigate more people getting that first impression because can you actually
- 6:23proselytize enough to convert someone back to reality once they've gotten very hot?
- 6:28I think what you want to do in an ideal case, you reach recutational escape velocity,
- 6:34meaning that you have essentially paid into your reputation, the bank account.
- 6:38So many good things, good deals, you're doing good work and you're saying the right things
- 6:41and you're out there and you're being present.
- 6:43So it's not the first first impression people have of you.
- 6:47They've had a hundred other impressions of you.
- 6:49So when they see something that feels out of the picture, kind of wacky, it's probably
- 6:55not real.
- 6:56And I think there's a lot to shifting information paradigm of we used to believe what we saw
- 7:02unless there was a reason not to.
- 7:04I think what we're having is we don't believe what we see unless there's the reason we
- 7:08do.
- 7:09So in my view, it's no longer that the medium is the message, the message is the message
- 7:14why?
- 7:15Because as he said, the relationship of trust is broken, but I probably have a few relationships
- 7:22of, let's say, podcasters, influences, thought leaders and business, who I trust, I've made
- 7:28the decision once to trust them.
- 7:30So I have this relationship of trust and it's hard to break that, but it's very, very
- 7:35selective.
- 7:36And I think that's what we're having.
- 7:39I hope so.
- 7:40You know, like I have this conversation with my son quite frequently, which is this
- 7:45idea of like retroactive cancel culture, you know, we have like artists and, you know,
- 7:53famous personas that have long been deceased where something comes out about them.
- 8:00And all of a sudden their reputation and their life's work is kind of diminished to something,
- 8:05to something else.
- 8:06I remember having this conversation about Salvador Dali, one of my favorite artists.
- 8:10Of course, he did have a dark side, which was arguably not forgivable over time spans,
- 8:17but it's very, very interesting how a personality can be seen in a positive light for generations.
- 8:24And then one piece of information can kind of, people can very easily erase that part.
- 8:31And I think because of the world we live in and because of communication technology, yes,
- 8:36we have these echo chambers, but it's creating more of a black and white narrative that exists
- 8:41there.
- 8:42And once you're black or once you're white, it's hard to switch those, switch those.
- 8:45That's right.
- 8:46That's right.
- 8:47It's more about believing, rather than knowing, and that's, that's a real issue.
- 8:50And echo chambers are real, of course.
- 8:52Well, you mentioned something interesting about the 3D internet.
- 8:56And when you first mentioned it, I pictured myself with VR glasses, but it's not, does
- 9:01all we're talking about.
- 9:03Yeah, the world of spatial technology is changing in an incredible way, right?
- 9:11So, you know, we can, you can think of, the 3D technology has been around for a long time.
- 9:16I mean, architects have been using AutoCAD and, you know, Ikea Kitchens for quite some time.
- 9:22Ikea Kitchens, yeah, or augmented reality, whether you could use the, the West Wing or
- 9:27Home 24, even now Amazon as these things when you can project an object into your space.
- 9:32If they all rely on models of these objects, so if you want to buy a sofa or a chair or
- 9:38something, the manufacturer creates these 3D objects that then you can render.
- 9:44What's happened with new AI, and it's specifically in the past 2 or 3 years, there was a, there
- 9:51was a technology called neural radiance fields or nerfs that was the first thing to do
- 9:55this that was able to basically take limited information and extrapolate what the rest of
- 10:00it would look like.
- 10:02Now, just in the past 2 years, there's an even more sophisticated algorithmic technology
- 10:07called Gaussian splatting.
- 10:09So what you can do is you can take a 2-dimensional image of something or even just descriptive
- 10:13text and you can extrapolate what the rest of that object looks like.
- 10:19So you imagine you're on an e-commerce site and you're looking at a sofa and you're trying
- 10:24to visualize how that sofa looks in your space and you can't, you can see one side of it.
- 10:28But with this technology and enough training into these models, it can, with a pretty good
- 10:35degree of accuracy, determine what the backside, the underbelly and the sides of that sofa
- 10:39looks like and create a 3-dimensional model of it with that limited information.
- 10:45And I think that is starting to, you know, we're already seeing it in the world of gaming,
- 10:50creating interactive 3D worlds, we're starting to see it in infrastructure and cultural
- 10:55heritage and now actually physical product.
- 10:58So it's just a, it's an emerging field that is very, very young, but some incredible
- 11:05stuff happening.
- 11:06And it just so happens that Germany is one of the world leaders in research and these
- 11:11types of spatial technologies.
- 11:13So it's been a fun sandbox to play.
- 11:16I imagine, with AI, AI reconstructs an image of let's say a car or sofa.
- 11:24So it takes all sorts of images of cars and reconstruct it into, you know, mold it into
- 11:29one car image.
- 11:31Whereas Gaussian splattering, the 3D imagery is a specific car, specific model or even a
- 11:37specific individual car, let's say, a collector's model, right?
- 11:40Is that fair?
- 11:42Yes, it can be, right?
- 11:46The more training that it has on information, the more it can essentially extrapolate and
- 11:51I am certainly not an engineer or a scientist, but the difference between, my understanding
- 11:56is Gaussian's are these unique shapes, right?
- 11:59They're kind of these all-blown shapes and it can kind of predict a little bit like
- 12:04an LLM, it can kind of predict what comes next on the boundaries of those shapes, right?
- 12:09And of course, the more information that it has, the farther that prediction can essentially
- 12:14go.
- 12:15So the leap is not in the technology, the leap is in the mathematics, right?
- 12:21And of course, we're able to do that because we have more computational power and, you
- 12:25know, we needed the emergence of AI and faster compute to be able to start unlocking these
- 12:33like deeper layers of technology and that's why I find this space so interesting is literally
- 12:4024 months ago, this is not possible, you know, five years ago, it's just being discussed
- 12:48in like, you know, research labs and all of a sudden we're seeing commercial products
- 12:53rolling out.
- 12:54I mean, you can think of some of the early like synesthesia and some of the early visual
- 13:00models that we were seeing in terms of like, you know, creating images from text like,
- 13:07what it looked like a year ago or even six months ago to what it looks like now in terms
- 13:11of the reduced hallucinations, the higher the quality is astronomical.
- 13:16So what is that going to look like in six, 12, 18 months?
- 13:20I don't think even people working in the field have a clear understanding because there
- 13:24are these massive exponential leaps that happen, throw in some quantum in a couple years
- 13:29and my goodness.
- 13:31So I think the immediate thing we will see is probably on the e-commerce side that let's
- 13:36say on Instagram, we will see 3D advertising, right?
- 13:40That's certainly something that that we will see more so handbags, cars, watches, whatever
- 13:44a protein powder.
- 13:46It may actually come in 3D because the world is 3D but the internet is 2D and the way
- 13:51it solves that problem is much more conducive to how humans perceive with our eyes.
- 13:57Yeah, I mean, that's already happening, right?
- 13:59There's a lot of companies that are creating, you know, 3D renderings of different products.
- 14:04So if you want to buy an IDES shoe, you can go online and you can move the shoe around
- 14:08and look at its underside and look inside of it.
- 14:12It's a really interesting new company that's kind of in stealth coming out in New York
- 14:16where you can take a couple photos of yourself, create a three-dimensional avatar and then
- 14:21find any article of clothing in the world and try it on in this virtual space, you know.
- 14:27So, I mean, but if you just go through your Instagram feed now, I mean, maybe it knows
- 14:33me to the algorithm, knows me too well, but I mean, a pretty significant portion is already
- 14:38AI-generated imagery.
- 14:40I actually saw post the other day and it was, this is actually a really cool account called
- 14:46the Wiz of AI and it's always kind of pumping out the newest technology in there.
- 14:51And it was one of those slides like 10 different images and there were two images next to each
- 14:55other that looked almost identical and it's a pick out which one is the AI.
- 15:00And I would say I got, I'm pretty good at it so I got more than half right but I definitely
- 15:05got some wrong where it was and actually the better the image, the one that has nicer
- 15:11lighting, deeper, richer colors is almost always the AI, right?
- 15:15Because it's lacking the human error component, right?
- 15:20And sure, you can train it to be more error prone.
- 15:23But I think that's the piece that's so interesting to me and actually it's one of the reasons
- 15:28why I'm really excited to talk to you today and it's something that I've been thinking
- 15:32about a lot which is the role of humanity in this world of AI.
- 15:39You know, I don't know how long ago it was now but when you and I had a little mind-melves
- 15:44actually.
- 15:45Yeah, in preparation for your TED talk.
- 15:50Exactly.
- 15:51And is that coming out pretty soon?
- 15:54It's out now actually.
- 15:55Yeah, it's just come out.
- 15:57You were supposed to send that to me, man.
- 15:58That's supposed to be the first one to see that thing.
- 16:00I can see some YouTube clicks on that for sure and I will put it in the show notes as
- 16:03well for sure.
- 16:04Yeah, thanks, buddy.
- 16:05I mean, you really helped me shape these ideas and what we looked at was AI and leadership.
- 16:12What does it mean for human leadership?
- 16:14And I started out with the idea or the prediction by futurists like Ray Kurzweil who say that
- 16:20artificial general intelligence is right the corner and by the end of the decade AI would
- 16:24do everything better than any human anywhere.
- 16:27So not just some things for some humans but literally everything.
- 16:31And I thought everything clearly includes leadership, right?
- 16:34So does that mean that we will now have AI leaders where we just take orders from algorithms
- 16:39and first of all, I don't want that to happen.
- 16:41So maybe I'm biased here.
- 16:42I don't find the idea particularly appealing.
- 16:45I'm camped human first, AI augmented.
- 16:47So I'm very bullish about AI and very optimistic.
- 16:50I love love the stuff that's happening that can make our lives better easier, more productive,
- 16:54give us more free time, give us better healthcare and so on.
- 16:58But there's this question.
- 16:59I just speak, you know, Gorilla in the boardroom that says, hey, I can take over from you.
- 17:04I'm actually better at deciding, predicting, looking around the corner, you know, creating
- 17:10visions and so on.
- 17:11And so to me, this is a fascinating question which I try to answer.
- 17:16Yeah, I mean, we'll see where it goes, right?
- 17:21Because AI is only limited by the amount of information that it has.
- 17:27And, you know, there's, I think a lot of experts in this space are, I listened to someone
- 17:32the other day that was talking about how, you know, companies like OpenAI and, you know,
- 17:37Google Gemini and Perplexity and Clawed and all these company Amphoropic are actually
- 17:43going to go to zero because right now, right now, you know, data is the asset.
- 17:51And eventually, everyone's going to have all of the data.
- 17:54So these companies will no longer have the asset and, and the technology and the compute
- 17:58power is going to be sitting on our device, anyways, right?
- 18:02Which some people say is going to, going to rule, rule that obsolete.
- 18:06So what happens in the future when we all have access to all of the data and all of the
- 18:10compute power remains to be seen.
- 18:12But I think one thing that compute can't solve is emotion.
- 18:18Right?
- 18:19It's, you know, the jury's out on creativity because right now, we're talking about patterns,
- 18:26pattern recognition and building, you know, LLMs are building off the, the what next.
- 18:31And maybe if you extrapolate that out far enough, you can get some pretty creative outcomes.
- 18:37But we're still seeing, you know, I don't know if you've ever tried to get Chatchy PD
- 18:41to write jokes for you.
- 18:42Yeah, the table.
- 18:43Yeah.
- 18:44Yeah, I mean, it's really good at placing the joke, like this is where the punch line
- 18:49should go, but it's really bad at creating something, you know, that actually connects
- 18:54with people because that requires, you know, something, something that, you know, shared
- 19:01experience, you know, something very humane, something very emotional, and that's what
- 19:08it's kind of lacking at right now.
- 19:10And when you think about leadership, I think it's kind of the same thing, right?
- 19:12Like, what's, why do we need humans in those, in those leadership roles?
- 19:18A thing is a good question.
- 19:19It's not easy to answer, actually, because when you look at the, so what leaders do at
- 19:24the moment, they're, they understand what's going on in the business in the world, right?
- 19:28And then they take decisions on that basis, who to hire, what to order, who to sell to,
- 19:32what's the visual, what's the strategy, and clearly AI is very good at understanding
- 19:36and deciding if we want to, because it has access to all the data and conquential data
- 19:40in no time.
- 19:41And the third element of leadership, to me, is communication, where I actually have
- 19:45to communicate this to people in the organization, but also potentially to investors to my board,
- 19:51to customers, of course, and those are human beings.
- 19:54And AI certainly has a role to play, but to me, I hope, and I believe that there is a
- 20:00role to play for humans, because yes, AI can simulate empathy, it can create empathetic
- 20:07words, but it can't feel empathy.
- 20:09And to me, that human connection is so central to our existence, the human desire to connect
- 20:15with other people to break bread and looking in the eyes and shake hands and maybe do business
- 20:20to you.
- 20:21To me, that's very central.
- 20:22What's also clear to me is that human leadership has to evolve to stay relevant.
- 20:26I don't think we can just stick to the leadership model of the past.
- 20:30I think we, we're moving away from the industrial model of leadership anyway, but I think this
- 20:34is the death now for the industrial model of leadership of command and control that's
- 20:38not going to work anymore in the future.
- 20:40I would recommend to everyone a quite too skilled and upskill in two areas.
- 20:45One is AI as a leader and the other one is communications that's human element of your
- 20:50leadership.
- 20:51If you go to those two, you will absolutely be relevant in the foreseeable future.
- 20:55Right.
- 20:56Right.
- 20:57Yeah, it's interesting what you said about empathy, right, because, you know, one way empathy
- 21:03in a vacuum isn't really, the magic of empathy doesn't happen, right?
- 21:10It's between two humans and the oxytocin that's triggered and the, you know, the feel
- 21:14good neurochemicals that flow through our bodies as a result result of that.
- 21:18So, you know, to me, yeah, AI can't trigger empathy if people know it's AI, right?
- 21:26So yeah, okay.
- 21:27You could pretend that the AI is human and, you know, tell rich, meaningful stories and
- 21:33use the right language, but I think that's very difficult to do, right?
- 21:37That's actually, yeah, and it comes.
- 21:39That's actually research now out, which was just the other week reported in the Harvard
- 21:43Business Review.
- 21:44And what they showed is that if your team believes you used AI to create messages, you sent
- 21:49to the team, for instance, trust goes down.
- 21:52And interestingly, the trust will remain lower even if your follow-up messages are written
- 21:58by yourself because the team thinks, oh, he or she's just using AI, right?
- 22:02So while it may be more efficient and time-wise that you can find the right words and be empathetic
- 22:06with AI, if your people suspect you're using AI to communicate with them without being
- 22:11transparent about it, trust will go down.
- 22:15And trust in a way is the killer app of the, you know, of the, of leadership in the age
- 22:20of AI.
- 22:21And what we have observed is that why do companies struggle introducing AI?
- 22:27Not start-ups, right?
- 22:28But larger organization, governments, big companies and so on.
- 22:32It's not because they can't deal with the technologies because people resist the introduction
- 22:37of AI often.
- 22:39And why do they do this?
- 22:40Because they fear the loss of agency and the loss of status and the loss of security.
- 22:45In other words, AI may take my job.
- 22:48And of course, there is some truth at that, right?
- 22:49We can't say others, non-sense people just worry that worry is probably justified in
- 22:53many cases.
- 22:54So if that's not the idea, if you want to use AI to augment your people, which I hope
- 22:59is the majority out there, what you need to do is build trust and help people bridge that
- 23:04gap between the people and the technology and take them on that journey.
- 23:08And we will only follow leaders who build trust with us on that journey.
- 23:13And only those companies can be successful in implementing AI and making the most out
- 23:16of this opportunity.
- 23:18Yeah.
- 23:19I mean, listen, I'm a big fan of the idea behind the, you know, solopreneur, you know, unicorn
- 23:27or whatever that might be.
- 23:28I find that I think that is inevitable in very, very unique specific cases where that
- 23:37could potentially work because there are, you know, scale economies that come out of
- 23:40this technology.
- 23:41But I'm really interested in, because I use it almost everywhere, like it's my course
- 23:46starting point right now.
- 23:48And, you know, I was someone who was saying the other day, like, my daughter is not going
- 23:53who's one and a half is not going to understand the term Google something, you know, like that's
- 23:59going to be out of the equation.
- 24:00Everything's going to be running through this.
- 24:01But the point I wanted to make is, you know, as someone that, you know, leads and runs
- 24:07businesses and I use this all the time, like, let's talk about operationally what a business
- 24:13leader or a CEO would do, like, where do they use it and where do they throttle back from
- 24:19using it?
- 24:20Because for me, one of the big benefits is to augment writing, like, because I find,
- 24:26I spend a lot of time writing and it saves me a lot of time.
- 24:30And I've done a lot of trial and error to figure out, do I start with it?
- 24:34Do I finish with it?
- 24:35Do I train my voice?
- 24:37Like, what's your take?
- 24:40Like, if you could create a playbook for leaders, how would you say the good places to
- 24:44use it and maybe where do you call it?
- 24:46Yeah, exactly.
- 24:47Two steps.
- 24:48One is to customize a GPT.
- 24:49So it rides like here.
- 24:51And for me, this was the big unlock when it started really doing that last year.
- 24:55It's actually not hard.
- 24:56Anyone can do it.
- 24:57Just watch a couple of YouTube videos and you can see on, you know, created gem on Google
- 25:01or created custom GPT.
- 25:03And in my case, I've used existing material.
- 25:06And of course, I've learned just because I've written so much.
- 25:08But let's say I've used podcast descriptions, right?
- 25:10I've fed in 50 podcast descriptions and said, hey, this is what it should sound like.
- 25:16This is the length.
- 25:17This is how we introduce the guest, et cetera.
- 25:19What we then do after the recordings is that we use the transcript, which the, you know,
- 25:23this platform gives us anyway, dump it into this custom GPT and it creates the, you
- 25:28know, thousands of words.
- 25:30It creates a couple of hundred that really introduces the episode well in my view.
- 25:35If I dumped it to the generic chat GPT say, I would get the bad average of the internet.
- 25:40I would get to get waffle, wordy, terrible sentence structure, M dashes without space,
- 25:45which I hate with a passion, the M dash, I don't even know where the keyboard shortcut
- 25:51is.
- 25:52It's so bad.
- 25:53You take some, take some LinkedIn posts you like and use that as training material.
- 25:56So you create your custom GPT and then I use what are called 10, 80, 10, 10 percent
- 26:02for me at the beginning.
- 26:03That is the, the idea, the hook.
- 26:05I always start with that headline.
- 26:07If I don't have a headline, I don't want to spend time creating a post on use letters,
- 26:10et cetera.
- 26:11So for instance, if, um, if I cut with a headline like, um, you know, what is it, um,
- 26:17how not to small talk, let's say for argument, I can use that as a starting point.
- 26:22I can say, okay, I want to create a post on how not to do small talk here, a couple
- 26:26of examples just from the top of my head.
- 26:28I don't type anything.
- 26:29I just, I just use voice to text function because I, I, I slowly type, I can speak relatively
- 26:34fast.
- 26:35Um, so I use that, dump it in there and it will give me 80 percent in the middle, right?
- 26:39So I do the idea at the beginning, 80 percent heavy lifting, there's stuff I could write
- 26:44easily because I have a lot of practice, but it just text on and then 10 percent by
- 26:48me at the very end to edit to maybe inject a personal story and to ensure it really is
- 26:53my voice, my values and also to filter out the stuff that I think isn't really what
- 26:57I would want to say on this.
- 27:01You mentioned one piece there for, I mean, I think it's a, those are really great pieces.
- 27:06The one I have struggled with is, you know, I was writing, I was just working on it yesterday,
- 27:12like my next sub stack post and, you know, I was kind of helping to create the outline
- 27:18of it. I was dumping, I had a lot of content from, from workshops and webinars into this
- 27:23thing.
- 27:24And what I, what I found missing was my stories, right?
- 27:30The experiential knowledge, the wisdom, the, I mean, I think the core to any good communicator,
- 27:36they're a good storyteller, right?
- 27:38Because they're, they're personalizing it.
- 27:40And the LLM doesn't know my life history, you know?
- 27:44So how to integrate it, you know, if I start with the LLM and it creates this model
- 27:49or it creates a draft for me, I find it quite difficult and demanding to now bring my
- 27:56wisdom in.
- 27:57I feel like I have to start with, like, here's the hug, here are the core stories that
- 28:01I want to tell.
- 28:02And now you can help me structure it and where to place them.
- 28:04I totally agree.
- 28:05And I use, again, it's, it's the voice to text because we can usually easily tell the
- 28:10story, right?
- 28:11Here's the hook and I want to use this story or this example, please integrate this.
- 28:17So this helps.
- 28:18And sometimes we spent more time than 20% on it, but I still think it does a lot of the
- 28:22heavy lifting.
- 28:23What is also great is the deep research function, right?
- 28:26So for instance, I want to write something on evocative and emotive language.
- 28:29I was like, okay, what's the research on this?
- 28:32And of course, it, you know, would have taken me hours to understand the research.
- 28:35If not days, I mean, people spend their academic lives doing this, right?
- 28:38So it, you know, I just type in a few words and after, you know, returning to it half
- 28:43an hour later, I have this beautiful research node, which I can then use as a basis for
- 28:48a post.
- 28:49And I couldn't have done this before because I could justify spending half a day just
- 28:52looking into studies and stuff.
- 28:54But if it says, hey, if you as a leader use evocative and emotive language, you know, people
- 28:59perceive you more positive view of 60%, you know, you're a bit here.
- 29:03Leadership rating goes up 60% according to this study.
- 29:06So this is cool stuff because it shows that we can essentially combine that sort of
- 29:11raw, fast intelligence with our personal wisdom to say, I think there is something there
- 29:17that people have overlooked so far.
- 29:19And once you combine that, that wisdom and the judgment, I think judgment is so important
- 29:24now, human judgment, with the raw intelligence and speed and compute of AI, often the result
- 29:30is really cool.
- 29:34It's interesting how you said judgment.
- 29:35I've never thought about it in those terms, but, you know, last year, I remember writing
- 29:40this article on the red sneaker effect, yes, familiar with itself, right?
- 29:45And it's like the kind of casual appearance of the leader gave them actual status, right?
- 29:52Maybe I've doubled out too far on that, but you don't get that in an automated process,
- 29:59right?
- 30:00You know, how do, how do people's unique personas and personalities shine through what is
- 30:08becoming?
- 30:09Yeah, I mean, we can personalize our voice with AI and make it better and better and
- 30:13better, but it is, my fear is the more that's out there, the more it's all going to mesh
- 30:19together.
- 30:20What, what does the future look like, right?
- 30:21Are the great leaders, the ones that really are the, the black sheep, the ones that stand
- 30:27out as being so distinctly different?
- 30:29Do we have to double down on our uniqueness and are the, are the leaders that are kind
- 30:35of maybe not as, because there's a lot of tremendous business leaders that are not particularly
- 30:41charismatic, right?
- 30:43Or, or, you know, fit into a large company really well and work their way up because they're
- 30:49best in class and what they do, but they're not the, the emboldened, you know, casual, loud
- 30:55entrepreneur or whatever it might be.
- 30:57Like, what's going to happen, what do you think's going to happen in the future?
- 31:00Is it, is it, are you going to have to be such an outlier to stand out?
- 31:04Have you had a dead internet theory?
- 31:07So that is the idea that's floating around that because the argument is that in the future
- 31:13or that content online would just be AI generated, right?
- 31:17So it's all just mediocre stuff.
- 31:19It's just flooding the internet and everything else is drowned out because it's so optimized
- 31:24to appeal to us.
- 31:26I think there's a risk that that may happen.
- 31:28However, I do think the internet is too appealing to import in and too useful for a lot of
- 31:32people to just abandon it and, and, you know, I don't think we all will turn into zombies
- 31:36just watching AI generated content.
- 31:39So I think what will happen is a continuation of the trend we've seen that we become extremely
- 31:44selective in who we trust and spend time with online, which means you know, I have our
- 31:48favorite podcasts and some of which we share and talk about.
- 31:52Those are the voices we look towards for information for filtering for curation, right?
- 31:58And if we do that, then the internet isn't dead.
- 32:01We just become extremely selective and going to ignore everything else because that information
- 32:05paradigm is shifting.
- 32:07What does that mean for a leader?
- 32:08I think the, there's an opportunity here for leaders to double down on their own unique
- 32:13voice, which I would find beautiful, which I recommend to everyone work with, say,
- 32:17hey, in the age of AI, your own unique voice matters even more than it has before.
- 32:22So for those who are charismatic and feel comfortable and want to go out and talk about ideas
- 32:27and want to shift me to move the company forward, I think that's a great opportunity.
- 32:32But your question is very justified.
- 32:34What happens to those who are not charismatic?
- 32:36You want to just, you know, knuckle down, maybe have a tech focus and work on the product.
- 32:41I think it will be harder for them to cut through.
- 32:44And maybe if you're in that situation, you're looking at partnerships.
- 32:47And again, that's something we see already that you have product heavy companies reaching
- 32:52out to people who have an audience who have attention, but to may not have technical
- 32:56knowledge for a product and say, why do we do this together?
- 32:59We give you some equity or, you know, we come aboard as a co-founder, etc.
- 33:03So I think the future's collaborations because it's very hard to be really good at both
- 33:08things.
- 33:09As I said, be good at AI, be good at communications.
- 33:11But in reality, some will probably be lean to one side or the other.
- 33:15Yeah, I mean, my hypothesis is that technology is increasingly democratized and accessible
- 33:26to everyone.
- 33:27You know, now that we can create apps with just a few words, that's not going to be the
- 33:33differentiator eventually.
- 33:35You know, there's always the, I think I may have said it on this podcast, you and I
- 33:38have talked about it a lot.
- 33:39There's two roles in a company, you know, either build shit or you sell shit, right?
- 33:42And all of a sudden, building shit is getting pretty damn easy, right?
- 33:47So now you look at the sell side and it seems to me like distribution is the tricky part,
- 33:56right?
- 33:57Now we can build easier, but getting noticed and getting recognized and building trust
- 34:01with the audience is getting increasingly harder.
- 34:03And we're seeing this fascinating rise of like the influencer investor.
- 34:09You know, you got Harry Stebbings in the UK and you got Stephen Bartlett, right?
- 34:16And really promising companies are going to them, not just because they have big audiences,
- 34:21but they have audiences that trust them, right?
- 34:24And I think this is a massive sea change that hasn't happened yet, but is getting really,
- 34:32really close.
- 34:33You know, who are going to be the leaders of the future?
- 34:35Like how do you, if you're a business leader, you're an entrepreneurial leader and investor
- 34:40or something, how do you prepare your position for the next 20 years?
- 34:44Like, is it, you know, having this public persona, having trust, having distribution?
- 34:51Or is it going to be the person that, you know, really integrates and focuses on the technology?
- 34:56My feeling is it's, it's the former, right?
- 34:59It's the people that, you know, have a persona, have this kind of broad relationship of trust
- 35:06out there that is going to have the competitive advantage over everything else.
- 35:09Because attention is so hard to come by and it will be harder to get in the sea.
- 35:14Hopefully not the debt internet, but the, you know, internet with lots of zombies running around
- 35:19and set code and set up this podcast, you know, mediocre content for mediocre people.
- 35:24And he pointed out the race to the bottom and said, you know, we have to get out of this mindset
- 35:28of using our eyes and race to the bottom because, you know, if you make sure you win the
- 35:33race and then you're stuck at the bottom, or worse, you come second, that's, that's
- 35:39that all you want to be.
- 35:40So we can use AI to race to the top and another thought that occurred to me in this, sort
- 35:45of in this context of of leadership and AI is, you know, we talked about extra, right?
- 35:51Build an audience, gain trust of investors of customers and so on.
- 35:55And internally, we all probably in the future going to lead not just human team members,
- 36:01but AI agents and our team members also have a team of agents, you know, I don't know
- 36:06when it would be 10,000, I have no idea.
- 36:10But this is a very sort of hybrid model of leadership, hybrid no longer isn't, I come
- 36:15to the office from time to time, but hybrid in the sense that they're humans and they
- 36:19are AI agents who are working towards a joint goal.
- 36:24Yeah, that's this idea of the leader being the orchestrator and chief, right?
- 36:28Like the, the orchestration agent that comes into play.
- 36:32And yeah, I think that's going to take a special personality to be able to lead the machines
- 36:39and lead the humans at the, at the same time.
- 36:42Essentially, I mean, look, maybe you're in a big enough organization that that's not
- 36:46the case.
- 36:47But when you look at smaller organizations, you look at startups, for example, and this
- 36:51is something, you know, we've been working on this, this venture for a few months and
- 36:54we're really trying to take an AI native approach, right?
- 36:57Like in the early days of, of any venture, like, you know, cost and speed or everything,
- 37:03right?
- 37:04You're trying to extend runway and, and move as fast as possible.
- 37:07And, you know, I can tell you, being in this position for the first time in a few years,
- 37:13the trajectory and the speed is astronomical.
- 37:17You know, we, in 60 coding hours with one person, we did what would have taken five or
- 37:23six, probably four or five months, just three or four years ago, right?
- 37:27So I love this idea of the more technology, the better in that regard, but just being
- 37:35able to make the right decision of, because there's, if you're paying attention, you're
- 37:42being bombarded every single day with a new tool and a new opportunity.
- 37:47And how do the good leaders know where to focus?
- 37:51Like where do I say, all right, I'm all in on the tech, I'm all in on the AI.
- 37:55And no, no, this is something I want people to, to lead and people to drive.
- 38:00And I don't know if I have that answer, I, I struggle with it with each of these decisions.
- 38:05Like, do you have any thoughts about that?
- 38:06I don't think I have a good answer because it's, it's a moving target.
- 38:10I think about it more as how to prepare for that.
- 38:13And I think that's probably also what, what anyone is listening, things like, what does
- 38:17it mean for me?
- 38:18What can I do?
- 38:19I do think it comes back to focusing the human side of leadership, communications and
- 38:24focus on understanding AI.
- 38:26And it's, I think it's totally fine.
- 38:27You're lean to one side more than the other and team up with other people because of
- 38:31the that way we are best prepared for whatever happens.
- 38:36And I called this in the title, I said, super computers need super communicators because
- 38:41I think if we only have super computers, but humans become essentially, you know, cocks
- 38:46in the machine that we have a real problem because we're making ourselves obsolete essentially.
- 38:50And that's not a good place to be for just societal stability, our well-being and so on.
- 38:55So definitely something to be avoided.
- 38:57So what I said is, you know, don't compete with the machine, use the machine to compete.
- 39:00And that's exactly what you're doing with your starter.
- 39:04Yeah, you know, I think we've been kind of, you know, praying at the foot of quantitative
- 39:12for a long time now since we've been able to do that.
- 39:15I mean, you look at marketing, for example, and how it just skewed so heavily towards
- 39:21performance, you know, and it's just quantitative analysis and just the more data you get, the
- 39:27higher resolution you get.
- 39:30And the less qualitative it was and the less human interaction, you know, it's human interaction
- 39:34by proxy.
- 39:35And I was just thinking about it, like technology is incredible for automating marketing spend.
- 39:42For example, it is absolutely transformative for writing code.
- 39:48Do I want AI leading brand?
- 39:51No.
- 39:52I don't think so, you know, like do I want it for, you know, like real customer success,
- 39:58especially in like a B2B context, support, yes, success, not so much, right?
- 40:04Do I want my AI trying to understand if my clients are happy with my product or trying
- 40:10to understand their needs deeper?
- 40:12I think there's places where humans belong and places where, you know, the AI will likely
- 40:18take over.
- 40:19And I think the really great leaders are the ones that are going to understand that.
- 40:23And, you know, what is that, what is that skill set?
- 40:26It is obviously the way we communicate and the way we interact, but I have a bias here.
- 40:32But I think the people that understand human emotions, the human brain, how we function
- 40:38biologically are going to have a big insight.
- 40:41I remember it's been five or six years, but Mark Cuban, the American entrepreneur who
- 40:46used to own the Dallas Mavericks, he said that the the era of AI is not going to be led
- 40:52by engineers.
- 40:53So it's going to be led by philosophers.
- 40:55And I always thought, wow, that's a bold statement, interesting philosophers.
- 40:59It might be neuroscientists, it might be something else, but I think the skill is going
- 41:04to be understanding the human side of things, right?
- 41:09And judgment, which of course rests on that understanding.
- 41:12So I look at this to keep it very simple, you know, is it 0% AI, is it 100% AI or anything
- 41:18in between?
- 41:19And essentially, it's my judgment.
- 41:20It's not that technology that holds me back, 100% AI, it may be okay.
- 41:25The call centers, right?
- 41:27Customers are quite happy.
- 41:28If you're transparent and say, hey, this is an AI, you know, assistant here.
- 41:32If you're transparent, they're quite happy because they get balance out of 20 seconds,
- 41:35rather than spending 20 minutes on the phone, waiting for a human operator.
- 41:39So they're happy and it can be 100% automated.
- 41:42Talk it to your spouse about, you know, after an argument, I would say 0% AI, LGBT may
- 41:48give you the perfect answer, but if you're reading from LGBT, you know, it's not going
- 41:52to go down too well.
- 41:53So I think the rule of thumb here is that the closer the relationship, you have with
- 41:58that person, the more important is to be yourself and show up as a human.
- 42:02Whereas if it is sort of mass outreach customers who are transparent with cool, that's totally
- 42:09legit.
- 42:10But be quite mindful.
- 42:11Is it 0%?
- 42:12Is it 20% AI?
- 42:13Is it 80% or 100% and that comes down to judgment?
- 42:17Yeah, I remember you sharing this kind of concept when we were talking about your
- 42:23TED talk, which is, you know, the leader that is anti AI, like 0%, or very low percent
- 42:30AI, they're going to get out competed.
- 42:32The ones that are 100% AI, they're going to make themselves obsolete, right?
- 42:37Because if the role of, like, what is the role of a leader of an organization, right?
- 42:42It's to connect with people, it's to inspire people, it's to build culture.
- 42:46These are very intangible, like, not quantitative concepts, right, that a machine can easily
- 42:53do.
- 42:54Sure.
- 42:55I think you could build a very successful company that could make a lot of money for at
- 42:59least some period of time that is just purely run by robots.
- 43:03There are use cases for that.
- 43:05But at that point, what do you need a leader for?
- 43:07Just turn on the switch and let it run.
- 43:10It's also a defendable.
- 43:11There is.
- 43:12That's right.
- 43:13That's right.
- 43:14Exactly.
- 43:15What is that?
- 43:16There is that balancing point that bell curve or whatever, where you sit in the middle
- 43:21where you're beautifully optimized.
- 43:23I think the challenge is that's a shifting target, right?
- 43:27As technology changes, do we have to zig when it's ag, or do we zig when it zigs?
- 43:34I think it's going to take a very special talent to be able to have a deep understanding
- 43:39of technology and an equally or greater depth of understanding of humanity in the way
- 43:45our emotions and our values fly, like tie into the equation.
- 43:48What would you recommend to people, as I said, for me, it's AI and communications
- 43:52focused on that.
- 43:53What's your takeaway to be prepared for this shifting target, or moving targets built
- 43:59on shifting sense?
- 44:00That's what it feels like.
- 44:02Well, I mean, I feel like I'm an old guy in this equation now, right?
- 44:07I'm in the second half, and I'm not sure I'm the best person to answer that question,
- 44:14but what I fall back on, so as you know, like I started my career as a development economist
- 44:19and I was working in communities where I was an outsider, and I came from a much different
- 44:27economic status than a lot of those folks.
- 44:29And the mantra that we always had was build meaningful relationships of love and trust,
- 44:36and that you do that before everything.
- 44:39And when you have that foundation, the opportunities become endless.
- 44:44And I think the more technology we have, the faster we can move, and so we're trying
- 44:50to skip steps all the time.
- 44:53And what I keep thinking about, whether it's starting a new company or trying to interact
- 44:57with a new market or new customers, is how do you build those foundational trusting relationships?
- 45:06And if you have that, then you can start really building off that and creating opportunities
- 45:13from it.
- 45:14So that might be old school, but I believe you start with that, and then you can look
- 45:20for ways to optimize.
- 45:21If you start optimizing from day one, you're going to miss where a lot of the magic happens.
- 45:27Maybe it's old school.
- 45:28Maybe it's just the wisdom of the second half, which I'm also, Garrett, it was so much
- 45:34fun.
- 45:35Thanks for your contribution.
- 45:36Thanks for all the insights, always, because it is such a moving target.
- 45:40It is so important to, I think, have this dialogue and also share it, I think, because
- 45:46if anyone has any additional ideas, please reach out, because I'm super open and so want
- 45:50to learn about this and get better understanding it, because there's not a single conversation
- 45:56I have with sea levels and founders, which I've talked to every day, without talking about
- 46:01AI.
- 46:02Because that's the questions everyone has.
- 46:03Follow it, tell you what's the impact what's going to mean for us.
- 46:06So I think it is one of these very fascinating, very rare occasions where the paradigms are
- 46:13shifting.
- 46:14And I just went through, which finally I want to share, because it just popped into my
- 46:17mind when you gave your last answer, which is there's probably some sort of a golden rule
- 46:22of AI.
- 46:24If you don't want people to communicate through AI with you on certain instances, then you
- 46:30probably shouldn't use it yourself.
- 46:32And something along those lines, right, think about do I want this to be a chat to be
- 46:37a team message, or do I want the person to communicate with me?
- 46:40So that could be a good way to sharp my judgment going forward when the lines are blurry.
- 46:46And I do recommend to founders that they should codify this internally.
- 46:51So when do we use AI, when do we not use AI?
- 46:56Meeting agenda's meeting points, cool, use it, customer communication makes a lot of
- 47:00sense, passing feedback sessions or interviewing new team members, that's probably better done
- 47:05by human.
- 47:06100%.
- 47:07I like that idea of having a policy, you know, this is where it's appropriate and this is
- 47:13where it is.
- 47:14And that might be different for different people in the organization, but setting that
- 47:18standard is a really smart idea.
- 47:22Yeah, always a great conversation, Oliver.
- 47:25I know that everybody listening, my man Oliver is such a humble guy.
- 47:31I got to hear some of his trial runs of his TED Talk, listen to it.
- 47:36It is amazing game changing will make you think of things in different ways.
- 47:40Oliver, I'm forcing you to put it in the show notes, but such a cool talk.
- 47:45I can't wait to see the final product and love that you've brought this sitegeist
- 47:50topic into, you know, the topic of communications, which is as old as humanity itself, and I think
- 47:58there may be no more important domain than communications in understanding where those
- 48:06machines fit in.
- 48:07So thank you for bringing that to the fore.
- 48:09Thanks, buddy.
- 48:10That means a lot.
- 48:11And yeah, we definitely put the link in the show notes.
- 48:12Awesome.
- 48:13Thank you.
- 48:14Till next time.
- 48:15Till next time.