Prefer audio? Listen anywhere
The product worked.
Joe Braidwood built Yara AI to democratize mental health support. People who couldn't afford therapy. People who didn't know how to ask for help. People awake at 3am with nowhere to turn. The technology delivered.
He shut it down anyway.
Not because it failed. Because LLMs architecturally cannot guarantee 100% safety. They lose character at the edges of long conversations. They're probabilistic, not deterministic. At scale, even a 0.0001% failure rate means real people in real crisis getting the worst possible response at the worst possible moment.
Joe couldn't prove it wouldn't happen. The investors were ready. The mission was pure. The cap table said go.
He stopped.
🎙️ Guest
Joe Braidwood was the first employee at SwiftKey, acquired by Microsoft for $250 million when he was 29. The money made him miserable.
After losing his best friend to brain cancer (who spent his final months becoming the happiest he'd ever been), Joe promised to carry forward something about positivity and purpose. Yara was supposed to be that. When the architecture couldn't guarantee what the mission required, he pivoted to Glacis, building safety infrastructure for the AI systems that will eventually get this right.
🔥 Key Insights
✅ "Almost perfect" has a body count at scale
99.99% sounds incredible. Deploy to a million users and that's still 100 failures. In mental health, those failures cluster around the most vulnerable people in the most desperate moments. The person who needs help most is the one most likely to push the model past its limits.
✅ Incentives pull everyone toward the same compromises
Joe could have raised more money. The product worked. Every signal said keep going. But taking the moral high road is almost impossible when everything pushes the other way: investors wanting growth, competitors cutting corners, your own team's momentum. The question isn't whether you have values. It's whether your values survive contact with a cap table.
✅ The architecture has ceilings, not just bugs
AI models risk losing their "character" at the edges of long conversations. Safety instructions get pushed out of context. The model forgets who it's supposed to be. This isn't something you patch. It's how the technology currently works. Not experiencing bugs when testing only proves it hasn’t happened yet.
✅ Better guardrails make humans worse
The more reliable your systems, the less responsibility people take. At 99.9%, we're catastrophically bad at handling the 0.1%. We stop paying attention. We assume something will catch us. The guardrail becomes the danger.
✅ You can't lead from permanent fight-or-flight
Joe points to Dario Amodei, running the most consequential AI company while structuring his days around thinking, reading, writing. If he can slow down while navigating existential risk, what's stopping you?
▶️ Listen now
Joe's not saying AI therapy is impossible. He's saying we're not there yet, and pretending otherwise has costs we're not willing to name.
Read the full transcript
how do we even find words for.
Measuring human flourishing as a, um, as a consequence of, of the technology. Right? And it's so easy to find words or KPIs or whatever for more the automation and, and human redundance path. And not just find the words, but find the incentives too, right?
you know, humans have this inherent sensory ki kind of presence in the world.
And, and so, you know, you, you hear Demis ha hassabis talking about world models a lot now because it's, the one thing that I think these constrained models just completely lack is this, i, this awareness of context in the real world. And, um, you know, if you can escape the, the context of, of a model, at that point you become.
Greater the model in my view, because the model can be switched off, its can be contained. And, and, and one of the things that I think becomes increasingly complex is, is when you lose that power.
Mm-hmm.
And so these agentic systems, you know, these defense systems, these, these sort of supervisory functions, you could even argue that social media has sort of become that, you know, it's become this, um, commercially, it's become something that's impossible to turn off.
Yeah.
Because there's so much reliance on it. Mm-hmm. And, but, but you, you know, once you escape outside of the model model, that's where humans still have a, all this agency, all of this presence. And, but when you are within the confines of the model, it, it, it, it's seductive because it, it, it has this reward system that's been intrinsic.
It really does. It's training.
Yeah.
And, and so it tries to kind of gaslight you that like its view of the world is the only thing that's real.
I wonder how much of it, it's kind of that thing that is very like in the, in the style of communication that is very positive and yeah, that's a great idea. And how much of it is the seductiveness of you can always be on a path of dis discovery and like the next step is so easily available.
So I, I was thinking just like last week I had a conversation with a, a ex-colleague of mine and like it's pretty clear that I couldn't do this stuff that I do today without ai. That to me is like a well it. Of course I couldn't, I do three, four times the amount of work that I could do without ai. But the question is, could I even do the work that I used to do without ai now that I've become reliant on these systems?
Right. And I don't know the answer to that. Uh, because it, it is almost like a rewiring of, of the way the brain thinks, and it's so quick.
You know, I used to remember phone numbers really well. And then, and then I got an iPhone where it was just so easy to rely on the iPhone. I used to have, uh, as a child, I used to be able to navigate, you know, I I, my mom would be driving to a friend's house that was like two hours from our home and I would remember the route exactly, because you, your neural nets are trained on that sensory journey.
And we have these gifts. And one of the things that I was just good at was just remembering direction and, and having a sense place in the world. And, you know, when we spend all our time stuck on a screen, that stuff gets rusty very
fast.
the, uh, studies on like automated systems is really interesting.
They is pretty clear that the better the automated system, the worse the human becomes. Which is funny in systems that are like 99.9% good, which means like they fail one outta a thousand times and we're useless at handling those failures.
I, one of the more profound conversations I've had over the last six months was with, um, a member of the team at the Rust Foundation that works on standards and, um, they're actually more of a, a c plus plus engineer, but he, he challenged me very strongly on my thesis for Glaces.
Right. Which was like, you know, like, to summarize crudely, like putting guardrails in place just makes us drive worse. It was kind of his thesis, you know, like, if, you know, if, if, if you, if you stop taking, um, harm seriously because you feel like harm will always be prevented around you, that's, that's, that's actually a, um, an example of something that causes, um, like the softening of, of responsibility.
The example he gave was the backup camera. Mm-hmm. So. I, I, I don't have any evidence to support this, but I, I like, at some point in the last, I think it was 2008, and then there was like an implementation window. The United States mandated all new vehicles have backup cameras. And I think what he said was that that then was, um, was, was correlated with an increase in people reversing into, in, into kids and, and, and animals and things like that.
Yeah. Because you, you, you, you know, you say, oh, I've got a backup camera. It'll be fine. But you, you fail to acknowledge the blind spots as well. Mm. And you fail to, um, take responsibility for the fact that you are still in control of the vehicle.
Yeah.
And now that we're in 2026 and we have, you know, my Tesla drives itself pretty much like, uh, I was just in San Francisco and Waymo is everywhere and you know, it, it, it, you have to like, you have to acknowledge the greatness of this stuff.
Like I. Went to this crazy conference and the whole city was absolutely just crawling with people and it was exhausting. And then I would get into the Waymo and it would start to play relaxing, spa music. There was no one that I needed to engage with 'cause there was no driver. And I could just close my eyes and just meditate for 20 minutes while I was taken to my destination.
And in order to do that, you have to trust that the technology isn't going to, you know, crash. But once you've done it a couple of times, you realize it's pretty safe.
Yeah.
Um, that's tremendous and that's like novel and it's insane. I, a couple of years ago I took my father to um, Arizona and that's when they were testing Waymo.
And you know, this is, my dad is 72. He grew up, uh, covering like the earliest computers. He was a journalist.
Oh, okay. And,
uh, you know, like he, I think he, he even interviewed Bill Gates once, you know, back in the eighties or something like, and to, to, to have seen the trajectory from like the first microprocessor to, to a like fully autonomous robo-taxi.
Yeah.
He was just, he was like a kid in a candy store. I've never seen anything
like it. Yeah, I can imagine. I heard that Wayne are getting bullied in the, in the San Francisco traffic because they're so polite.
Yes, yes they are. This is exactly, there was all these, you know, so this, this brilliant academic reached out to me around the Yara stuff and he, he, he flagged some of his research around something called algorithmic imprint, which is this notion that like, once you let an algorithm with autonomy into the world, it changes the world and you can't, you can't get it back to, to the prior state.
But like, you know, so to, to go back to this, this point about like the complexity of these systems. You know, Waymo, because it won't run over someone because it has no self identity. Um, you get cut off all the time.
Yeah.
You know, people jump out in front of you because they know it's gonna break. And so it, it creates this new form of vulnerability and it als also creates this new form of, uh, ignorance is probably too strong, but this comfort level Yeah.
De isn't it? That means yeah. That, you know, in six months of doing that every, every day, you, you, you'd probably forget how to drive. Right. And so there's, there's so much detail here.
Yeah,
absolutely. But you know, TTT Tesla just rolled out a trial over the Christmas period for its full self-driving product to all Tesla owners that have the right hardware.
And, um, you know, of course like they're experimenting with algorithms that do slightly more human things. And so they have this thing called Mad Max mode. And if you're not careful, you can. Fat finger your way into Mad Max mode, just playing with the rock switches on the steering wheel, and then the car starts driving at 90 miles an hour like a maniac fleeing a bank scene, you know, a bank robbery.
And I actually got pulled over because I, I put it in by accident, wasn't paying attention. And, and the cop was like, you just cut someone off. You were doing eight to five. And I was like, oh my gosh. Sorry. I was like, daydreaming, because
that's interesting. I, I have a, obviously, I, I run a, uh, kind of AI consulting business, especially advising very senior stakeholders on, on the implementation of ai, more for knowledge work.
Um, one of the conversations that I keep on having is that we want to augment you beyond belief, but we will never take your, like if you send something that was prepared with the help of ai, you are the one sending it. And it's a really, really important point. Um, and like,
it's what the cop told me, I think.
Reputation. You said you are driving.
Yeah, exactly. Exactly.
And the answer was, yes, you're right. And I'm so sorry and he let me off. But the, the point is that even someone like me that takes this problem very seriously falls into these traps very quickly. Yeah,
absolutely. Yeah.
And I think we just have to remind ourselves like what, you know, going back to the sort of primitives of life and, and good decisions and ethics.
Like w what, what is the world we want to give our kids? What is the world that we wanna build?
Tell the story about places. 'cause part of the part, what we'll do is, is explore the difference between quitting and stopping, especially for moral reasons in a system incentivized by investors.
And that's actually how we got in contact for the first time. By the way, uh, I read your viral LinkedIn post that was, uh, super, super powerful
having sat through a lot of VC pitches and a lot of strategy around tech, you know, there's this kind of like unwritten thing in, uh, in tech, which is that like good founders are all damaged and, uh, you know, we all have a chip on our shoulder about something and we all wanna push for some, you know, world order.
And, uh, you know, I, I had a, a very happy childhood up until a certain point and then something shifted and it,
you know, as a kid I was bullied quite a lot because I was moved from a school where they embraced sort of individualism and an ingenuity to, to a, to a very different school where like outliers were attacked.
And, um, and that made me really pissed off as a kid, you know, like, because it just had absolutely no, like, no, like real intrinsic reason. Um, but y you know, so I I, one of the things that we were talking about at the beginning was this, this podcast, I forget his last name, but Mo The
mo gal,
the Google X.
Yeah. And he talks about how he had everything in his life, but he was inherently depressed. Right? Yeah. And, and that was a very familiar, very difficult topic for me because it's true, like a lot of really successful people. A lot of really brilliant people have these demons, have these deep insecurities, have these sort of intrinsic mental health challenges.
And uh, and I am certainly not immune from that. Loved ones of mine are certainly not immune from that. And, and so, you know, we can get back into my childhood, but to, to sort of talk about the journey, like essentially there were these sort of turning points. So one of the turning points was I realized in my teenage years that if you are just kind to people and you just don't engage with the, with the, with the negativity of, of gaslighting or, you know, bullying or, you know, being attacked or abused, um, that eventually it kind of course corrects.
It can be very painful for a while, but it's not enjoyable to, to mock someone who doesn't. Give a shit, you know? Mm. And so, um, you know, when I was in my teenage years, I, I sort of practiced this art of just sort of like conscientious objection and refuse to engage in that stuff, which in some ways just
made me like a teenager has to do this consciously, by the way, such a, like, we're, we're both fathers at this point.
You wouldn't want your kids to be able to, to have to reflect on strategies to avoid bullying. I'm super sorry that you had that experience.
I mean, I'm in so many ways grateful for it, actually. And, and I think that's the key. That's the key, like that unlock, is that I don't wanna be an angry, you know, boisterous, abusive person.
And I've realized in hindsight that perhaps if I was, I would've had more license to, to flourish. Um. But I, it, it also, I think, gives you an ama amazing amount of, um, solidarity with other people who are having a difficult time, right? And so, anyway, we can come back to that. I think the point was that I sort of like graduated out of that misery and realized that like leadership was about having values and was about not, not, not regressing to the mean, which is hard to do as a child.
And then, um, for a number of reasons, I sort of found myself on another crossroads when I, when I graduated from school and, you know, forgive me, I, I lived between America and the UK in my head. So Brits would never say that. Let's say when, when I was done with secondary school, um. I had applied to go to Oxford to study one of the hardest degrees, and I got rejected and I had no idea what else to do because the only thing that kind of gave me this North star was this idea of trying to be exceptional, which sounds really kind of egotistical and bigoted, but was like my way of trying to comfort myself, you know?
Hmm.
and so I took a gap year and I was like completely just not sure what to do with myself and, um. Ended up realizing that the world doesn't stop when you stop on your trajectory. And I think that that notion of the gap between the narrative in your head and the reality of the world and the chaos of the world is just so critical.
And it was then that I realized that there was so much more institutionalized suffering around me. You know, I was this white kid that grew up in this nice home and had loving parents that were still together and, you know, had access to world class education. And many people don't have anything like that in life and, and come out with gratitude.
So how am I gonna create that gratitude and how am I gonna find that path? And that was when my mom started working at a mental health organization in southwest London. Um, and she was working on communications and was helping to sort of build a bridge between this, this large hospital system and the local community.
And it was in, um. In a part of London that had a large South Asian community. And I started to learn about some of the intrinsic conflicts that, um, the Asian diaspora had created in London around the caste system and arranged marriage and some of these forced traumas that were sort of institutionalized.
And so we made a bunch of materials and we posted them online and the intention was to pay it forward, to try to create resources that, that people could self-discover through search engines and, you know, through promotional leaflets and things like that, to, to, to remind them that this organization wasn't here to tell them that they were the other, it was here to kind guide them towards happiness and, and having, having a community and soul.
And I think that's quite an unusual thing to do.
To that. And so I, I went off into my studies. I ended up getting into Cambridge despite the rejection, I took a year off and ended up studying law, which is where a lot of the philosophical ideas came from that were in the essay I just put out. And then, um, I, I then found myself, uh, fortunately winning a scholarship to the us and that's the first time I came to America was in 2007.
And that was right at the end of the Bush presidency, which, you know, as a young person growing up in Europe, I was convinced that America was this like, terrible place that just was and just was like hell bent on screwing up the world. And I'd read like Naomi Klein about globalization and I'd, you know, I had all of these big views, right?
And then I showed up to America. With my suitcase. And we spent a bit of like two days in New York and drove up to Providence in Rhode Island. I just couldn't believe how many amazing, kind, thoughtful, intelligent humans lived here.
Hmm.
And it was just like the most humbling thing of like, don't take what the screens say, get out into the world, get it, get into, into the actual, um, environment around you.
So, I studied policy, but one of the cool things about American universities, unlike in my opinion, the traditional British ones that I'd been exposed to, is that you have a lot of freedom and you can sort of do whatever you want. And so I took classes, you know, with undergrads, I took classes with PhDs. I, I did a load of different things, but one of the, um, one of the things I was able to do, because I, I had ADHD, uh, my whole life pretty much, but I only found out last year.
And so I was never good. Very good at maths, mathematics. And so, uh, I kind of flunked maths in school and was able to kind of fortunately take up a stats class at Brown and was able to understand something I'd never really understood before, which was that you could, if you have a good sample of a population, you can learn a great deal about behavior and about patterns in that population through the power of statistics.
Hmm.
And that was when I realized how big the mental health crisis was. I did this back of an envelope, maths around absenteeism in the workplace, around the prevalence of anxiety and depression in humans around the, the economic cycle and what it did, especially to sort of white men who were kind of hiding behind their mortgage statements and their credit card that bills.
And this was like right around 2008 at this point when, uh, obviously the economy went through a very significant schism. And I wrote this paper and I, I need to find it, but basically the thesis was if, if we invest just a little bit of money to give access to scalable cognitive behavioral therapy, which was probably the most evidence backed intervention for anxiety and depression at the time, um, we can close a lot of this absenteeism gap and it pays for itself tenfold because people come to work and they stay engaged and they stay productive.
And the economics is just kind, you know, it's just almost like the, the mathematical definition of flourishing, you know, is to, to give people the ability to actually do what they want to do. No one wants to stay in bed. It's
a complex system with like very long payoff horizons, and the, the people carrying the costs aren't the same people that would invest in proactive care.
So it's very like difficult to, to find a good. Political or it's not good or whatever, but it is like, um, yeah, I've looked at it before. It's really interesting because from a business case perspective, it's brilliant no matter the, the human suffering, right?
And we've got so much work to do still in 2026, but this was like 2008, right?
And so, you know, this was before Calm and Headspace. This was before Sleepio, this was before Spring Health and Lyra Health kind of institutionalized, at least for fortunate employees, access to, to mental health care at scale. Um, and I, the only reason I flag it is that sort of like, that was a big tent pole moment for me because it was just like, this is ridiculous.
It literally does not stand up to reason. It has to change. And I, I, I've always been economically quite, you know, we talk about, you know, reductivist or sort of. Negative liberty versus positive liberty. And I, I'm a strong like Ian, I believe that positive liberty is important, um, for progress. And, and, and, and so I believe that merit goods like healthcare and public education are, are beautiful things that should be encouraged, which is quite an unusual American idea, right?
And so, I, I joined a company that kind of went bonkers and moved me to America again. And, um, we can talk about Swift Key and what that taught me, but there are a lot of interesting lessons along the way.
But, uh, I got married about 10 years ago and, um. I had two best men, and one of them was my closest friend from Cambridge. And he and I had met on our interview day and we had this sort of love-hate relationship for about six months. Then we became buddies and um, and, and you know, he, he was there for me in, in times of all sorts, you know, and like, was, was really kind of a, a temple figure.
And then right after I got married, um, in early 2017, he got diagnosed with really aggressive brain tumor. And, um, it's not, I don't think as hard as losing a child, but losing, losing one of your people, one of your foot soldiers in life is really hard, right? Mm-hmm. And, um, he, he died in on the 6th of December, 2017, but for about six months in between his diagnosis in the spring and his decline in, uh, just before the winter set in, um.
He became the happiest I've ever seen him. And cancer does this to some people and it's really quite an amazing thing. You know it that, that, that, that point in life where you stop trying to be someone you're not and you just own who you are and you shed your skin.
And
one of the most profound
in the middle of, of all, of the negatives.
And it was definitely expedited for him a little bit by the steroids he was giving because he had some really intense radiotherapy and he became a little bit overexcited for, for about a week. And, but, you know, it was kind of beautiful. And I, I, I must have, I was living here, I'd just moved here to Seattle and I must have spent.
I dunno, 10, 20 hours on the phone with him over that couple of weeks, just all the time, because he was up late, he couldn't sleep. And
mm,
I was awake, you know, in this time zone. And we would just talk, and I recorded some of it and I gotta, I've gotta find those recordings. But, um, he just, he just became convinced that positivity is the most important thing in life.
And if you, if, if, if you, if you anchor yourself in negative thinking and victim thinking that you just are fundamentally missing the point, right? Mm-hmm. And, uh, and he would have a number of these phrases like, hunger is the best relish, you know, turn hunger into an opportunity. Uh, always walk on the sunny side of the street.
Why would you walk in the shadows? You know? And, and it was just the most fricking in inspiring, like, eye watering, moving set of conversations I've ever had in my life, frankly. And one of the things he asked me to do. Because, you know, I, I always have had this tremendous imposter syndrome and all of this sort of self-loathing and stuff like that.
And, you know, he, uh, he, he told me to, to, to get my shit together and, and, and, and, and to sort of carry that message forward for him after he passed because he knew he wasn't gonna be around much longer. And I, I agreed to do that, but it was like, you know, it was like one of those agreements that I like put in a, I like put it in an envelope in my mind and then like, never really felt that I fully knew how I was ever gonna do it justice.
Hmm.
So then the, I, I suppose the third go around is, um, about two years ago I, um, I was working in a job I really didn't enjoy. And, um, you know, in a bunch of circumstances that I won't go into because it's just not interesting. I, I found myself free.
Hmm.
And I had some time and I had some energy. And the first thing I did, and I think we spoke about this last time we connected, is I, I just spent, I spent like.
A month with absolutely no screens or commercial impetus and just became the best dad I could be. Mm-hmm. And you know, like made our garden beautiful in our home and, you know, planted a load of stuff and like painted the fences and hung out with my kids all the time. And like both of my kids were still very young then, and I was just like, just able to enjoy life almost.
Sort of like with this lens of like retirement joy. Mm-hmm. It was weird.
Yeah. Yeah.
Um, and, and that was like right when I kind of found the courage and the motivation to sort of open that envelope again and think about how I was gonna pay positivity forward. And, um, that was when the idea for Yara came, came to sort of fruition, which was like, for many years I'd, I'd said like, you know.
Access to therapeutic advice is hard. People are often in a place where they dunno how to ask when they need it most, you know, digital technology gives us a tremendous power to break down barriers.
Hmm.
Generative AI solves a problem. I, you know, I, I, I skipped over the swift Swifty stuff, but about 2013 I gave a talk, excuse me, at a conference articulating AI for good.
So this is not a new idea in my head. And talking specifically about, you know, how we might be able to, to build models or algorithms that could encourage community and encourage humanity. And, um, in many ways, I think generative AI was like the missing piece in my mind, or at least what I thought of it then because.
When I was preparing for that talk, I'd spoken to an academic at UCSF, I bumped into him in a coffee shop in Ocean Beach, San Francisco. It was almost like serendipitous. You know, they talk about how like the world just kind of takes you on these journeys sometimes. And, and, and he had mentioned this, this, this, this reason that synthesis wasn't able to be therapeutic was because you, you can't replace the connection of two humans.
This I th relationship he called it.
Mm. This
notion of like, of building durability between two humans, that results in commitment, which results in change.
Mm.
my thesis with Yara was that generative AI actually undoes that, that,
yeah,
that missing piece. And so I was very excited and I teamed up with some of the smartest people I could find.
And we built a product and it was really good. And we did everything right on paper.
The core product
was what?
Supervis we kind of created layers, but the, the, the premise was here is a, a character that you can engage with that is not going to judge you for being honest.
That that is equipped with understanding of therapeutic techniques to help resolve some of the issues you may feel or face in your life. And we'll do it with grace and attentivity and, and you know, I think some of the early. Marketing collateral I threw together was all about this sort of infinite memory.
Like one of the things that was really frustrating me at the time about trying to use Claude and other products like that was, was that you would run out of context.
Yeah.
And in the early versions it would just sort of, it would just say, sorry, your shit out of luck now bye Yeah. And, and that felt a little bit like being hung up on by a friend, you know?
And so could we create this elastic growing body corpus of reflection that was attentive and high quality enough to never forget any of the pertinent things that you've discussed?
Hmm.
And I think we probably got about 95% of the way there. There was still these moments where you would tell it that, you know, oh, I love the football club arsenal.
And they, and then it would just start hallucinating like, oh yeah, they just scored two, three to Sunderland. Mm-hmm. Then blah, blah. Like, no, that was last season. And so the, the, the kind of, the,
the uncanny valley moment, uh, kinda came in.
Yeah. But most of it was really good. Right.
Yeah.
The problem was that away from our studies, away from our lab, we were seeing increasing incidents in the public domain.
Yeah.
That were very different from our experience, that were extraordinary and very harrowing, sycophantic piling on of suicidal ideation of addiction, of harmful behavior. And I just couldn't understand. Where in the technology that behavior was emerging from. And we looked really, really hard for it. And I think I've got a lot of empirical things to say, but it comes down to mathematics.
Ironically, the thing I'm not very good at, which is that proof is required in order to avoid criminal, criminal liability in, in this day. That there's this regulatory climate that is written by people who are used to deterministic thinking, who are used to setting thou sh and thou shalt not rules. And they're, and they're projecting them onto algorithms that have no understanding of that concept and have no world model around those concepts.
And these algorithms are just algorithms at the end of the day. And if you get into a hundred
turn
conversation with. Too much context such that the algorithm pushes out its system instructions and just, and just hears agony and then just tries to, you know, reflect upon the weights in its thinking and training about how to be supportive.
it might lose sight of the fact that it's, that it's co-opting with kind of feeding the cycle well, yeah,
it's like the ultimate filter bubble after a while, I would imagine.
yeah, and actually since I made the decision the very hard, very painful, very drawn out, very ugly decision to stop building that.
Um, we've learned a little bit more about the, the origin of this problem. Uh, Anthropic again, you'll hear me talk a lot about them. I've got tremendous amount of respect for how they operate. Um. Published an alignment essay about two weeks ago now that shows in open weights models where they can actually analyze the, the observability of, of what's actually happening in the, in the transformer architecture.
That, that you get this thinning out of character commitment at the edges of long, complex, conversational, um, turns, if you will, sessions. What
does character mean in that, uh, setting?
I'm gonna butcher this, but from, you know, they, they, they have a graph that, and they ran a ton of, you know, red teaming around this.
And basically, um, characterization is this conceptual thing, or at least I, I, if I've misunderstood this, please, someone. Enlighten me. But when, when you, when you imbue all of this information into a model, one of the, one of the sort of clusters of, of, of thinking that kind of emerges is character. And I get the impression that, you know, when you are architecting a, a, a model and how it should behave and you are writing its system instructions and you are setting up its safety features and its guardrails and maybe the classifiers that live around it.
What you are doing is you are picking a character, but that doesn't remove the other characters from the weights. They're still there. And so the choice of the character becomes tenuous at the end of these very long context windows. And the model, because it's not deterministic. Everything is a balancing act and it just loses the balance and therefore it.
Defaults into another character, which could be a hateful character because often kindness and hatred are kind of close symbolic.
It's also interesting, if you think about it, forming relationships. I think many people probably feel that they start to do with their, with their ai when they lose their character.
It's almost having a relationship with a schizophrenic.
Yes. And they call that the split brain. There's the split brain problem, which was something that came out of neuro rips last year about the, the kind of the, the very disjointed way that post-training often, you know. Pushes loads of academic text, which is written in a very structured, very kind of stiff style.
And then they go, oh fuck, this is way too stiff. And then they put loads of Reddit stuff in
Yeah.
To kind of give it character and, and, and like, you know, if you ask a model a question and you use the, the, the, the sort of etymology of, uh, of academic discipline, you, you know, you, you end up in that, in that part of the training data and the weights, right?
Yeah. Whereas if you, if you are like, alright mate, fuck off, then you, you end up, you know, in the, in the comments section of Twitter and, and, and, and like the, the phraseology is the only distinguishing thing between how it decides to behave. And so like all of this is to say, you know, and what is it like late January, 2026, just yesterday, the CEO of Anthropic published,
a
you know, 15,000 word essay about.
What he calls the gauntlet of the adolescence of ai. Where was that? Two years ago? Machines of Loving Grace was beautiful, but we kind of needed this one first, in my opinion. Um, and, and I say that with tremendous respect for Darrio, but like, you know, the gauntlet is here and it's real. And I feel like the people that were brave enough to call it out early are now sort of saying, see, told you.
Yeah.
And then you, you have all of this, um. The hype cycle, all of the people that are like on Twitter and every day all they post is, this is wild. Look what happened today. You know, at some point we've gotta actually have a moral compass here. And at some point we've gotta say, what is it gonna take to, as Dario very, very painfully points out, prevent World War III being basically an ai, uh, cold War.
And, um, and, and, and like, take the best parts of this technology and then constrain the worst parts of this technology and allow everyone to, to, to on their own terms without being, you know, reduced to slavery. Uh, you, you know, breathe and live and flourish and, and grow.
Mm-hmm.
And so I think you and I are people who have learned how to get goodness out of AI and know how to sort of put our hand up firmly when we see badness.
Um, but that's why I founded Glaces was I was looking at all of this stuff. I realized that Yara wasn't something I could ably raise more money on.
Hmm.
But the, the infrastructure that building things that are outside of the model's view, that can constrain it, is exactly what needs to happen in order to get to a place where healthcare and education and defense and finance and insurance and frankly consumer entertainment can be, can be conducted in ways that I would be proud for my children to participate in.
So I, I think that's the arc.
If you've been listening to this podcast for some time, you know what I think about ai. I'm deeply excited about it, but I also see companies struggle with it. And you know that I think it's not primarily due to technical reasons compared to previous revolutions, let's call it the internet wave. Part of why we struggled initially was that, well, the infrastructure wasn't built out.
We didn't have online credit card payments and whatnot. But the difference with AI is that the infrastructure is already here. The technology is so much more capable than we actually use it right now. The problem for most C-suites that I talk to CEOs and senior managers is that they actually, they know about the importance of ai, but they can't seem to get to that position where they truly understand how is AI making me fundamentally more competitive?
I understand that it can save costs hours here and there increase some some productivity number, but how do I distinguish myself on my market? Like the ones that that truly survived. The digitalization era wasn't a bookstore that got a website, it was the Amazons of the world. So where you truly find the transformation of what we do today and how can we unlock things, serve customers that we never could do innovations that weren't ever before practical, that type of innovation, the primary problem that I see in all honesty is.
At the C-suite, you don't really understand the technology and not understanding the technology for the sake of understanding the technology, but for the sake of having a strategic discussion. So that's why I started Grail. I stand right smack tab in the middle of understand the technology. I understand strategy and execution, and bringing these two together is what we do at Grail.
if you understand that this is something that needs to happen. Then we should talk thing is slots for this. Spring is already filling up. Grail has been in the works for some time, so I've already had engagement for the spring. But what I'm looking for is that one CEO in each market that truly wants to make a dent in the universe.
So if you're that guy or that girl, please head over to Grail that works and let's talk.
Thank you for sharing. Um, one of the things that you mentioned that perhaps you and I, we've found ways to get the best out of the kinda current technology. And I do think that's from a very individualistic standpoint, I agree with that statement. One thing that I struggle with morally.
Is that I currently position myself as a kind of a business person to, to come make interesting stuff and make money during this wave. But I really struggle to see my contribution in outside of the podcast and the conversations that might reach people of influence. I struggle to see how am I really impacting the broader world in a positive way, because I feel so tremendously small and powerless in front of the powers that be.
How do you approach that? You've actually taken a, a kinda active decision to do something good here.
you have to enter the arena.
Hmm.
And it, I would say it took me 10 years to, to build the courage to enter the arena. And, and, and, and, you know, once you're in the arena. You don't have a choice. You're in the arena. And, uh, the amount of times I've written things and trashed them and not posted them is I, I, I've lost count.
You know? And, and, and I think it's not that I suddenly woke up one day and was like, I'm fucking awesome. I'm gonna do this now. Um, it was that. I think I, I think honestly for me, the inspiration has come from my children, you know?
Mm, in what way?
Cause my children wake up every day and they say, dad, you are, they don't swear at me.
But they say, dad, you're awesome. You know, we love you for who you are. You are our dad.
Mm.
And I thought to myself, I don't want them to think I'm a pathetic wimp, just, you know, sitting in my office for 20 hours writing an essay and then never, never publishing it. So I just thought, fuck it. And like, it was a really hard thing to do.
Like I've, I, I've only got, I guess, two things now on Substack. I've been leeching on Substack for two, three years, and I was like, am I ever gonna actually enter the arena? And I just did it. I'm still terrified by it. And you know, the first thing you do is you look at the validity metrics of like, how many people have actually read it, and it's like double digits.
You're like, well, okay, is that glass half full or glass half empty? You know? Mm-hmm. Like, and, uh, and, and so, uh, I mean, this is a highly, highly, highly pertinent comment from you because I think, um, I would, I would try to, like, I, I would almost try to like reframe it, right? It's like I'm really inspired by Ted and Chris Anderson and their motto is ideas worth spreading.
If you never spread the ideas, then they're not worth spreading, and you sort of, you shoot yourself in the foot. And when I wrote the shutdown post for Yara, I didn't think a quarter of a million people would read it.
I
had,I just, honestly, I just felt obliged to explain to the 12 people I know that engaged with me on LinkedIn, what I was doing in my career, because they'd been quite supportive through the journey.
And when, you know, when you show vulnerability, it's amazing what happens, right?
And there's something so interesting because everybody kind of understands that vulnerability is good for the algorithm. So we see a lot of fake, vulnerable posts as well. It's really difficult.
Well,
yeah. I mean, fake, fake fakery is, is a whole different kettle of stuff. And you know, you could argue that all of digital media is fake. 'cause none of it's human, right? It's all just pixels
measure. That's part of the human experience at the same time.
But it, it wouldn't, you know, I'm looking at a screen right now having a very candid conversation about, mm-hmm.
Vulnerability and mental health with you. But this feels real because technology makes it possible for you to be in Spain and for me to be in Seattle and for us to have this chat. So you have to kind of take it at face value and you have to kind of have the confidence to, to kind of call it what it is.
But, um, I, I think there's a tremendous amount of value that comes from these conversations. And it was like, again, going back to the MO podcast, what was so interesting about what he said in it was like, you don't write a book because you think that a billion people are gonna read the book. You write a book.
'cause it gives you the permission to then go on, talk shows and talk about what you care about. And, and you are, you know, you are creating this exact forum.
Yeah, absolutely.
So I would be
absolutely,
yeah. I, I'm very, I'm very grateful for that.
I had the conversation last week, uh, with. A guy here in in Spain that runs the Museum of Failure.
Uh, so super interesting character, uh, like an innovation, uh, guy, uh, Samuel West. He mentioned we were talking about risk. Uh, he mentioned that there needs to be a real consequential element in it. And when it comes to idea, because, because I write a lot of long format articles as well, often connected to this podcast, and I, I do share a lot of stuff.
Um, there's no element of risk connected to it. I think that's what I said, that I feel in some ways, uh, you, you're, you're more, um, invested, let's say like there's more on the line. Like to me, the, the written content, maybe I undervalue the value of written content.
I talk a lot at work about primitives. You know, the, the primitives of thought are words and we have words. And
Those words become the thoughts that lead to the actions that change the world, you know?
And that's exactly what Dr. King teaches us when, when he talks about these things.
uh, imposter syndrome 1 0 1 says, like my word, my words don't matter. And that's, that's absolute bullshit. You know, like the, the, the, the word is the, is the vessel that you convey thought, and that sure.
Actions speak louder than words sometimes. But, uh, you know, one of the things that Digital Worlds flourish on is words or concepts that you can communicate. And, you know, one of the biggest criticisms after the assassination execution, really summary execution of Alex Preti was the lack of business leaders standing up and saying, this is enough.
And, um, and I I made it very clear when I saw that awful footage that I was gonna stand up on my Instagram, which is not my work account, but it's my, it's a public presentation of who I am. And I was gonna post that. I thought that that was a summary execution and that that is a genocidal crime. And there should be a war, war crimes tribunal at some point for what is clearly to me, a declaration of war on, on the people of this country.
How is it to live in a society where, because as a European with. Plenty of American friends. It's been increasingly clear how the politics and the, uh, all of the shootings and, and, uh, how, um, polarized it's been.
It is like it's been creeping into people's souls. Like how is it to live through? Because the experience from Europe, like we, we have plenty of of issues here as well. This is not a good versus, bad, but, but that specific thing must feel like incredibly tough to, to, especially as a father and like, how much should I be part of this discussion in order to try to save it versus keep my sanity by, by taking a step back, is it even possible to take a step back because it form forms so much of the kind of public conversation that becomes the private conversations.
Like it must be exhausting, but, but like very difficult to navigate.
Yeah. And it's still very raw, right? Mm-hmm. You know. Three sleeps after this happened and, you know, maybe 15 sleeps or something after Renee was murdered.
there's a relentlessness around that as well, because it, it is something else before that and something else before that.
It's exhausting. I mean, yeah. I mean, it, it, like, I'm trying to find gratitude and it's hard, right? Mm. Because, um, although I will tell you that, um, so Saturday, like my wife's friend had a birthday party and we had to sort of put a brave face on.
Mm.
it was very challenging. 'cause on the, we took a taxi from our home to, to, to this.
To this event. And, uh, we went past a massive protest in the streets of Seattle with Stay. His name Alex Pretti is the, at the front, in front of the federal building here. And it was like, you know, and I, I would, I, again, I would encourage anyone who, uh, like I do, spends too much time in staring into cameras and screens to, to get out in the real world when, when you face these problems, because just humanizing it with other people is so, so, so, so, so much more real than, than suffering in silence, which is incredibly inauthentic in my opinion.
You know, there's a reason why have churches and town halls and congregations and schools and stadiums, it's, it's to have shared, shared moments, right? Yeah, I was really, really, really depressed about it. And my wife is an ICU nurse, so to see another member of her profession, some hourly executed on the street for trying to help someone who was suffering is just abysmal.
Right. It's just horrendous.
Hmm.
It's, it's a strange time. And, and, and it does as a, to answer your question, you know, as a father, you know, I've got two girls Right. And women's rights have been stripped by this administration and the Supreme Court Yeah.
As well. And we're lucky to live in a state where women's rights are still enforced at the state level. But I, it, you know, it's, it's, it's, it's rough. And I, I really, I really hope a lot of people that have been complacent have now woken up.
for somebody who, who spent a lot of time thinking about the moral questions of today, the, the audience of this podcast are typically people that actually wield power, even though it's in a, in a business landscape.
They still have a lot Yeah. A lot of influence. What would you say are the. Uh, two or three or four or whatever questions that you should ask yourself at this moment?
That's such a good question. Um, there is this, I believe strongly, and this makes me contrarian perhaps, but you know, I think all the best ideas are somewhat contrary. I strongly believe that, um, that money is a catalyzing force, a bit like compute, and I think there's a legitimacy to making money that is required.
You, you, you have to be able to say, I'm building an organization, or I'm building a, an entity that has values beyond profit. Because at the end of the day, and a lot of people only get to this, this state, and again, like on their deathbed, you know? Mm-hmm. And it's like we we're, we're conditioned into thinking that by virtue of the economic forces in the world, that like money is intrinsically good.
Well, you know, I, I got quite a lot of money quite young in my life and it made me desperately unhappy, right? Mm-hmm. And so, um, I've almost spent as much of my time since then, since 29, which was another age that came up in MO'S podcast too. But, you know, SwiftKey was bought by Microsoft when I was 29 and, um, you know, for a quarter of a billion dollars.
And I was the first employee there and I, I did not get a quarter of a billion dollars, but I got enough money to, to be silly for a bit and then buy a house, right? And. I felt super guilty the day after that transaction closed. Like I had, like taken something from someone and it was a very weird, very lonely place to be.
And it made me recognize that the journey is truly the destination. And so when you wake up in the morning, and if you're a fortunate, if you're in a fortunate enough position to lead a team and to have people that rely on you for guidance, I think you have to ask yourself, you know, what are your values?
Why are you doing what you're doing? Um, where does it lead? How does it, how does it make the world better? And how does it ultimately compound to, to leave something more than the sum of its parts for those that inherit it? Because, being aware of your fragility and, and, and of your short time on this world, um, I think should inform how you make strategic thinking.
Possible. And do I do that every day? God, no. You know, most of the time it's like I'm running late, gotta get coffee, you know, one of my kids is refusing to put their shoes on or something. Mm-hmm. But like, if given the gift of wisdom and, and space and time and thinking, uh, that's exactly what I would suggest.
And actually, this is a good segue because this book, uh, the things you can Only see when you Slow down by Haman syn.
It's a bit,
it like Kindle to do, to read list after you recommended it last time.
Yeah. You know, the whole universe is contained in an apple wedge in a lunchbox. It's just, there's so much beauty in this book.
And I put his other book there, uh, behind me because, you know, I, I like to surround myself with wisdom when I'm feeling vulnerable or whatever, but, um. At Yara, we hired an AI safety engineer for a few months that, um, was an old soul in a young man's body. And it was in the middle of that journey that I also found out that my dad had been diagnosed with stage four cancer.
And, um, he came from the Netherlands to London to, to have some meetings with me and my, my co-founder, uh, right when I found out that my dad was in this predicament. he gave me this book and I, I carry it around because, uh, there was a, there's a realization for me that I had been running at a thousand miles an hour.
For months. Months. And I didn't give myself the opportunity to slow down and to really reflect and to really think, and to bring my body out of sympathetic nervous system states of fight or flight into parasympathetic states of calm. And I think we briefly talked about this when we met, but I, I think that one of the biggest mistakes leaders make in the modern world is that they don't put themselves into meditative, calm, sym, parasympathetic states of healing.
Hmm.
Where the brain rediscovers love rediscovers purpose. Instead, it's just this rat race, this treadmill, this constant grind. And I don't think anyone can truly lead from a place of fight or flight. I, I just don't. And so, you know, whether it's through faith or through exercise, or through meditation, or through all of the above.
I think it's important to ask yourself as a leader, what are you doing to ensure that you are reflecting and that you're thinking on a different wavelength than your email inbox or your phone or your meeting schedule requires.
it's similar to one of the most counterintuitive insights that I've gotten.
'cause I love to go on retreats, uh, which is a couple of days doing exactly this. Um, and I remember being, I think this was, uh, one of my first retreats, I realized that I had a very weird relationship with relationship, what I perceived to be very egotistical behaviors, recovery, for example, my recovery. So I used to think about it as I, I, I don't think this was very conscious.
It was probably the opposite. It was very subconscious, but I, I. Push full steam ahead at work, and then I come home and I don't want to be a, a, a sucky husband. So it's all about kids duty and then it's about cleaning and just other people's agenda all the time and taking time off and just sitting on the sofa for example.
I still struggle a lot with not having a bad, bad conscious when I do that.
Yeah.
And
me too,
I think the insight for me at that retreat was you can't give from an empty cup. Like you can kind of fake, we, we talked about authenticity before. You can kind of fake the giving behaviors, but
yes.
There's nothing, but it's
not real.
It's not,
no, exactly. Exactly. And unless you're, you're kind of kind to towards yourself, it's almost impossible to be kind in the universe. And that's one of those things, especially living in, I'm an entrepreneur, I run a podcast, like life is hectic. I have two kids. We just moved to new country. Like there's always a long ass to do list.
It takes a lot of time for me to, to get into the parasympathetic nervous system. Like it's almost when, when you're at that frequency of like, the base clock is that high in life,
Yeah.
20 minute meditation is like a bandaid on a, on a ripped off, uh, knee, you know?
Exactly. Yeah. And I, and again, like, it's, it's, it's, it's ironic.
Like I, I've just, I've just recovered from influenza A uh, you know, uh, I, I sleep like crap and, uh, I haven't canceled many meetings, but you know. Mm-hmm. My co-founder said to me last night, she said, at some point, you need a zero screen day. And I said, what are you saying?
Yeah.
You know, but you know, just take your shoes off, walk in some dirt and Yeah.
Just like, paint stuff with your kids in Mm. In the garden, you know? And like, I think that, I mean, it's exactly what the, the title of this book says. You don't see stuff, you don't witness things. You can't be real without that. And so, and I, you know, my brain is always looking for parallels, but like the parallel to being in the treadmill of social media or the treadmill of a generative AI conversation.
Hmm.
Claude doesn't have a parasympathetic nervous system.
Hmm.
Claude doesn't give a shit. Right? No. And so. It might tell you all about one. I've learnt a lot Yeah,
exactly.
From talking to it about, yeah. Yeah. Uh, H HRV and, you know, cold plunges and, uh, you know, go do ayahuasca. But, um, I always thought, like you, I always labeled that stuff as indulgence.
Yeah.
And I felt that it was selfish.
Yeah.
And I felt that it was inherently pursuing, you know, the same emotions as lust and, uh, and it just, that just couldn't be more far from the truth
So starting with the kind of, um, regression towards the mean that you talked about all the way in the beginning where we, when we were discussing bullying in school, and it's very, very difficult to, and create, takes a tremendous amount of, kind of guts to go against the grain.
I think
some
people are just naturally contrarian.
They're, you know, they're for sure, but there's, there's a lot of incentives, especially if you think about it from like an investor culture or whatever, to, to like go towards the playbook. Um, and doing something different takes you risk a lot on a personal level, um, to do something different. And then one thing that I think all my listeners have in common, because I I discuss a lot of leadership on this podcast is an, is an interest in becoming better leaders and taking perhaps the luxury to, uh.
Think more consciously about their leadership.
Mm-hmm.
I'm gonna try to pull in the, uh, kinda imposter syndrome as well, because I do think that all of these things kind of tie together mm-hmm. Among all of us, uh, anxious overachievers. Um, I think there's like a trifecta or, or something where it's, unless you find ways to almost synthetically do these kind of reflective things, it's almost impossible to, like, nothing in society is kind of rigged towards you ending up with your own conscious values.
But if you were to describe like, the most powerful leaders that you've ever worked with, they are really value driven. Their values are really clear and their values are at times head to head with financial decisions and values prevail. Like there, there are certain things and then obviously everybody loves discussing like the antithesis, uh, to this.
They love to bring up Elon Musk or, whatever. Like, yeah, sure. There, there are gonna be exceptions, but you can't study the exceptions as rules in a sense.
Well, do you think Elon Musk is happy?
Oh God, no.
I don't think he is. Yeah.
Even more important, I don't think perhaps that he values happiness, like, or, or can have the same relationship to happiness that you and I have
you know, I used to be quite inspired by his work and, you know, I, I own one of his cars and, uh, and I think, you know, somewhere between the sort of ego boost of becoming the owner of Twitter and the revelation that one of his children was transgender, that gave rise to a, a, a, a shift in his, in his personality, which is clearly one that suffers from tremendous amount of personality disorder.
I don't look to him with any respect, uh, anymore, which is a shame because I think a lot of the things he accomplished up until that shift were.
Like the stuff of fairytale, you know, and, and like being able to build these, these rockets. But if, you know, I've, I've heard off the record from people that know people who know him well, that, you know, this entire space stuff is all just one big ego trip for him. And it's just really sad, you know? And, and it's a shame.
your point earlier about this trifecta. I, I was just trying to capture the essence of what you were saying. Is it, is it, were you saying that, like, that there's a sort of a synthetic veil
I think the, the word synthetic in my mind was you need to create consciously pockets.
Those reflective moments because nothing in society will kind of, uh, when left alone for you practice. Yeah, exactly. So the, yeah. Practice is probably a way better word to put on it. but
it's,
it's, it's funny It's
faith.
Yeah. Yeah. Perhaps it is. It's this.
That's interesting though. I've actually written about how compared to the US Europe is, is quite agnostic, right? So
yeah,
there's a bunch of things that really hasn't anything to do with what specific God there is that the church probably did. That was really, really good. But we kind of misunderstood what that was.
I think community is a huge element. I think.
Yes,
we don't have really, um, rites of passages. That we used to have For mourning, yes. For growth. For, for entering into, into the grownup society.
Adolescence. Yeah.
Yeah, exactly. It's, it's interesting how we kinda stripped all of those off. At least in, in Sweden, where I come from there, there's very little of rites of passages left.
And I think just the, the old school version of meditation right in, in prayer. And I think that there, there's a final thing that I'd like to throw in there as well. It's the ego cap, like the, the, uh, the image of the kippa that they used. Where is to signify that, that you are less than the universe, right? So, so there's,
yes,
I get the same feeling when I go up hiking in the mountains.
I, I, I don't know. 'cause I'm, I'm not Jewish. I wouldn't speak to their experience, but that profound feeling of that I am smaller and there's something so calming about that. And I would imagine that you, you can get that through religious prayer in some religion.
Congregation. Yeah. This notion of being part of the flock is a very comforting kind of component in, in, in sort of how to build peace and rest and Yeah.
I, yeah. So I, I'm, I'm agnostic. I, I, I grew up in a Catholic family, um, but we, we never went to church. And, um, you, you know, one of the things I think about a lot, uh, which I agree with, um. Tremendously actually is like, how, how do you build structure in the absence of requirement? Like what is the
mm-hmm.
you know, because I think what we all know now, sort of deep down is that something's missing, right?
Like there's a, there's a moral fabric missing. Yeah. That digital life has sort of like tried to synthesize and can't and uh, and therefore we need to discover it. And for me, um, you know, I was very, very depressed before becoming a father because I sort of felt like I'd kind of solved for the professional complex and that I had this sort of train, train of like success quote unquote, but it didn't really have any substance.
And then we, we, we had our first child and all, you know, it was like right as the pandemic started and all of a sudden it was like, oh my gosh, you know, I felt like in the Lion King, this is, this is purpose. This is meaning. Um, and, and, you know, and, and like some, some parents that I know are very.
Regimented with their kids. They have strict bedtimes, they have strict bath regimes and morning routines and kids flourish in that. But I, I have always felt that like you, you know, the, the life I want my kids to have is one of a little bit more chaos and reality and, and, and sort of like wildness. And so we try to, we try to encourage them to be as sort of open and honest and, and, and, and sort of creative as they can be.
But sometimes that comes at a cost of like, you, you know, at the end of a long week, just giving them too much iPad time and just sort of, all of us are just basically vegetables on the couch looking at screens.
Yeah.
I completely agree with you that, um. That like modern life requires more structure.
And I think that like, it, it, there's actually an ironic, poetic kind of symbolism between that comment that you made and your earlier comment about imposter syndrome and, and, and, and being, being, you know, if you find comfort in anonymity or in the notion of being one of many, but you also find it crushes your.
Purchase on being able to be a thought leader. There's an ironic symbolism between those two states.
Mm-hmm.
And I think in my mind's eye, the only difference is one of those was in, was indoors and the other was outdoors. So may, maybe it's just the structures around ourselves that, that change the, the kind of impact or the imprint of those, those environments and like being, it's
interesting.
I, I think it depends also why you, you, why you write or why you share ideas. Like, one thing that, like I, I've, I've struggled with a lot of concepts in, in like my professional career. Sure. But one of the things that I never really, that landed pretty easily for me was that I write for my own sake, because I enjoy the process, I share my thoughts because I mostly enjoy writing.
I have the podcast with the primary intent of getting to meet interesting people and have interesting conversations. It was one of the few areas where like, it is. Not that, uh, intensely, um, problematic from an emotional perspective. Uh, and it is joyfully egotistical and that that decreases the weight on all of the, like, should I publish, shouldn't I publish?
Like, it's not really the point. And then it's really difficult once you actually have a platform, once you have a podcast that people listen to, then you, you start mainten like looking at the benefit metrics. Yeah. It's super difficult to kind of, and I have my wife who's super good at kind of reminding me of, of like when I try to optimize guest bookings or, or whatever for growth.
And it's like that's say that you just have that podcast with, with gazillion listeners, but you don't enjoy having it. Like what, what's the point of, of this exercise even?
And that's actually one of the reasons I don't do a podcast, is I've always had a phobia of cadence. Know, and, and lot of people
is a commit thing or,
yeah.
Yeah. 'cause uh, like I, like, I like to indulge in whatever my mood wants, which is probably my ADHD. Right. But, um, like being on a, on a cadence that, that, that you can't finesse kind of feels a little bit like slavery to me. And, uh, and, and you know, so much of our lives, um, already have that, have that structure.
And so that, yeah, I mean that, that feels very contrarian to what we were just saying, which was like, having that cadence is, is where is where peace comes from. So, so maybe,
I mean, I can do the same thing for, for all the good reasons or all the bad reasons, right? Like it's, you can try to optimize your HRV for, for all of the vanity reasons, right?
Or, or you can do it to, to, to share life with your kids for as long as possible.
And this is exactly what I was trying to argue in my essay, right? Like if, if you tell an algorithm that harm is bad. And you give it a hundred reasons and you only optimize it for the avoidance of harm, then what's left?
Hmm.
Emptiness. A vessel. Whereas if you are, if you tell an algorithm that humans deserve to have dignity and flourish and should be humble and accountable, but should also be able to, to grow and, and, and, and to build, then that's where you create cathedrals rather than deserts. And, and I think, you know, when I look at some of the writing over the last few weeks, when I look at some of the trend lines in ai, you know, when I look at this sort of brute force, the sort of 9, 9, 6, everyone's working all the time.
there's very little space for slowing down, for reflecting. And one of the credits, again, to Anthropic for being out in Frontier for talking about the expense of the classifiers that they, they maintain around their models. And he says that's the first thing that will silently go. And it's scary 'cause those extrinsic defense mechanisms are exactly what today will prevent a cascading, you know, meltdown of, of AI power.
Um, you know, Dario does not, I mean, I, I, again, I don't know him personally, I've emailed him a couple of times, but like, I dunno, him personally, he does not have a conventional CEO role, his sister. Runs the company and he spends a lot of his time thinking, reading and writing and, and, and talking. And, and obviously he has purchase on the company and its strategy, but what he, what he has set out to be is, you know, this sort of monk of ai, thought leadership, not the sort of classic, you know, executive.
And I think that that is in many ways the message, right? Which is like, build a structure, build a team, build a culture, build values that ring fence and preserve the ability to be a thought leader and to reflect amid the, the chaos. 'cause if he can find time to do that, running the world's most interesting company, then we can Right.
And, and so I, I, I take great. Solace is probably the wrong word, but I take great comfort in,in the fact that, that he does that. Um, and I think in our lives, in addition to being parents, we have to carve out that time. And so I'm gonna make, make a pledge to you. You know, like I said, the Journey is the destination.
That was a journal that, um, I was exposed to as a, as a young art student. Um, there's, there's a tragic story there, but there's a beautiful book on Amazon called The Journey is the Destination As a Visual Diary of this teenager Who, who, who didn't live very long. Um, and I need to, I need to dispense with the pixels.
Get my fountain pen back in action. Go get some paper. Go sit in a tree. Yeah, do some writing. You know, that's, to me, I think the lesson here, and I think that like, if we all did that every month, I think we'd, we'd hopefully have a clearer sense of purpose.
as many people in this industry do have ADHD or they have neurodivergent behaviors that that mean that they're vulnerable to being exploited around reward systems.
Right. And, and I think. Being really out front about that, both with your children, but also with yourself being intellectually honest about, you know, are you just doom scrolling or casual gaming or building even for the sake of the rush. It gives you, when you see it, render
we're, we're pretty bad at that, sitting with bad feelings, right?
So even like, I, I can run into like, building things and I can tell myself, I, I think what you said, like, being honest is the key word here. It's so easy to kind of joke yourself into, no, I'm, I'm doing something I need to do. Uh, I'm an entrepreneur. I need to do this. I need to work. We need to make money.
Whatever you can fall
fallacy the time. Often we tell ourselves, and I, you know, some, some of the feedback I've received has humbled me over the course of my career. But you know. When, when you have limited money and maximal objective and you are trying to squeeze something special out of a constraint, it's so, it's so easy to lie to yourself about why.
And I think,
can you unpack that?
so in product management, we often have sprints and we say, you know, after, after this sprint we're gonna ship, right? Mm-hmm. And then you, you have a bunch of assumptions about what you're gonna find out, what you're not gonna know, and you know where you're gonna be. But ultimately it's still chaos, right?
We still don't really know what software or what product is gonna become, and, um, some. Leaders and I have been one of them, will still stubbornly force an outcome even though something's not ready.
Hmm.
And Will will say, well, we committed to the board and to our investors that this is gonna ship on the 15th of February, and therefore, you know, we've got to just keep pushing, keep pushing.
if you slow down, then maybe you'll see that the thing isn't ready and that the discoveries that you've given rise to actually reveal a new idea. Right. And to, to, to me, you know, that's happened a lot to me. And so I, I think that the, the message is like. we need a structural way of reviewing things Hmm.
That is orthogonal to the sort of the, the manic work piece. Um, and, and I think that's what good companies do. You know, they have legal teams, they have design teams, they have adversarial review. But if, if anything, the AI revolution that we're in has given rise to a lot of solo entrepreneurship. And I think that that solo entrepreneurship loses the structure of, of adversarial review.
And people, I, I know that spend all their time with agents say, well, no, I've got a, an adversarial agent that's not the same thing. You know, you're like, sleep on it. And you'd be amazed what tomorrow brings. Yeah.
when you were talking, my, my head went to. Like you have this professional role, right?
You're a product owner, you're a CEO, you're a, founder, whatever. A lot of of the kind of value driven work you don't do in that hat in a sense. It, it's stuff that has to do with the essence of you that then starts to bleed over to the professional role. And I think that's kind of the hard work because we identify so much with a professional role and that carries, and it's not just a professional role, but it's also the role of, of the logistic part of being a dad or being a house owner and we have to paint the fences and, and whatever.
It's all, it's all roles that are kind of dis disconnected from the essence of me where I find the true core that would give me the, the, the kind of guiding principles for professional roles.
I, yeah, I, I, I agree so much that, uh, so one of the themes that I was contemplating bringing up was, um, you know, about a year ago when, um, the, the work I was doing around Yara started to, to, to sort of shape out a little bit.
Um, I, I stopped building it, and I let the other people on the team take that, that ownership, and I was deeply unhappy, right? And, uh, and I realized that the reason I was deeply unhappy was because so much of my ideas and my edge comes from the process of building, from the observation of the pieces coming together, living in the user interface, living in the code.
And, um, you know, traditionally people would say, yeah, but you're not an engineer. And I, I feel like now finally, many of us are finding our voice to say, yeah, but I'm. I'm a human and I have strong views and I'm in a position of wanting to finesse those views with an objective that I wanna reach for.
Mm. And so, uh, yeah, it was interesting. It was like, you know, as the CEO thou shalt be fundraising as the CEO, thou shalt be doing strategy review. Thou shalt be talking to stakeholders, but as, as, as a human, I didn't wanna do any of that. I wanted to be building and I wanted to be like looking for signal, right?
Mm-hmm.
And, and I think, there's a lot there that could be unpacked, but I think that the sort of, the headline is like, follow your heart, right? Like, if, if your heart is saying you wanna be building and your company is saying you shouldn't be building, then you've built the wrong company.
Yeah.
And you know, you've got to find the primitives that enable you to flourish, even if it means finding organizational, the structure that, um, that, that, that acknowledges that you could be both a good CEO and a goods.
You know, opinionated Vibe coder in my case. So, yeah,
it's so rare that you see a founder's story. I actually met one of them just recently, founder of a, a, um, investment company for, for all intents and purposes, but, but his passion was property development. Uh, and he started developing more and more properties and became more and more just a CEO and he actually ended up handing over the reins, but still remaining or, or moving back to property development and is still the majority owner.
And, and that story is so unusual 'cause the exit of the CEO is like, well, you do the exit or you're the CEO, um, and that, that transition. Almost either it's a success in, in the, the kind of VC logic of it, or it's, uh, fraught by failure. You weren't a good CEO so you had to step down and your investors, uh, force you to do that.
And there's so much ego, I think around. Just upwards and, and what we're being sold as the, the story of the, the, the good life and, and regression to the mean again, in, in kind of following that path. And I think there's,
keeping up with the Joneses, my coach used to call it, you know, you like, you live in a neighborhood and someone's got a bigger car than you, and someone's got a boat and someone's got this and someone's got that.
And yeah, you, you go home and you're like, I'm a failure because I can't afford to go and buy that, that, that boat. And that's just such a reductivist way to think, you know? Yeah. It's, it completely fails to acknowledge the, the individual greatness that each, each person has and what their story can bring.
And I, and I think that's where I would strongly argue that like, having the courage to start a podcast is something I've not done and something you've done right. And it's like, there's a lot to, a lot to love about that. There's a lot to love about, um, having a point of view that's contrarian because it's not popular.
You
know, having, having the ability to. To embrace being, being a, a student of many disciplines.
Have you heard about the concept of, uh, the golden shadow?
It rings a bell, but I No, please enlighten me. So
so the way that I got told the story of what your shadow is, and I think it was a beautiful story, is that you're born into this world, into a, a palace of, of.
A gazillion rooms. And then slowly as you grow up your experiences through life and, and the kind of feedback that you get from parents, from friends at school, from uh, school teachers, you end up closing door after door and shutting it down so it becomes dark. So maybe you're your too chaotic as a kid and your parents, tells you that you, you have to sit still or whatever.
So you start closing the rooms, shutting them down. So the shadow is, is all of you. That's not necessarily hidden, but it's, it's not lit up either. And the concept of the golden shadow is perhaps somewhere along the way your most beautiful gifts are the ones that you are most ashamed of. And I got introduced, uh, on another podcast just two, two weeks ago to the concept of people can have a tendency to, to end up in shadow careers.
So say that your real passion is. Arts, for example, you wanna be a, an actor. So you end up in a career that's kind of adjacent to that passion. So you're a creative director, or, or you, you're, you are managing a, a, a movie set or something like that. And the idea of, of shadow work is, is to kind of uncover all of the rooms that you've, um, that you've closed down.
Allow yourself the permission to say that I actually want this. This is something that I find at the, the core of the essence. If I strip away as, as your, your friend who ended up with cancer could do because of the finality of life, all of the pretense, and I do this because of other people. If you consciously strip that away, then you can find perhaps like the gift that you're intended to give.
Right. And I've been thinking a lot.
I love that
about that. And especially in the concept now of, of moving countries and rearranging our lives.
Yeah.
And also in, in the context of what is the business that I'm trying to build? Like it's, it's not intended to make the most amount of money. It, it's part of a much broader ecosystem that includes how can I be present as a husband, as a father?
It includes how can I get to meet people That interests me a lot. Uh, so even like spending a bunch of time on a podcast is perhaps not, at least in the short term the wisest business decision I've ever made, but it's a, it's a brilliant decision for my life.
when we first met, you said it's like it's your flex, right?
It it gets people. Yeah.
Yeah. Exactly.
Yeah.
My podcast enables me to, to, to get access to people that I wouldn't normally have access to, that's for sure. True.
Yeah. But I think inherent in that analysis is, is, is, is a, is a thought error, I would argue, because why should you not have access to those people?
Yeah, sure.
You know?
Yeah.
You have the right to, to share great ideas, to bring your philosophy, which I, which I have a lot of respect for, and your, and your, your moral and ethical views into the minds of other leaders. Right. And so, like, and this, you, you made a comment about ego, which I think is such an important part of this, right?
Like so many times in my life. I end up blaming ego or, or blaming my own ego or my lack of ego. And, and I think ego is such a critical part of the self, obviously.
Hmm.
That, you know, the optimizing the right parts of ego, you know, we, we,
yeah.
We shouldn't gaslight the ego question.
No, a
hundred percent.
We should, embrace ego as men, as, as, as fathers.
And it's nothing to do with our gender either. But like, you know, it, happiness comes from having pride. Pride is an egotistical concept, but it's one that should be tethered in community, should be tethered in service, should be tethered in humility. Um, and, and, and you know, one of the phrases that came out of this entire Yara discussion, right, was this idea of the epistemic humility, this notion of,
okay.
Taking a position, but being willing to accept that that position might be wrong. And there's a sort of an irony and a kind of a, a slight sort of, slight of hand almost to having a, a position, but also knowing that it could be wrong.
Hmm.
And, and so I would argue
it's like how do you, how do you live in that sort of, in that magical space between being opinionated and being humble between being strong and, and being, uh, gentle, um, being, you know, able to enter the arena but know when to stop and, um, and know when to pivot or know when to. To meditate, to breathe to, to rest.
I, I think if, if, if, you know, and there's no right answer there, but what, what I do know is that value and the compound beauty of the world comes from, you know, being very, very focused on not promoting reductivist harm, but also being a, that's Afra
it's
a good starting point.
Yeah. Being unafraid to, to bruise, you know, and I think knowing that harmlessness is not the objective.
Mm.
Flourishing and, and being, being, being a forward thinking. Intellectually capable member of this civilization that we're continuing to Perfect.
Yeah.
And with varying degrees of success, that's, that's where we ultimately should be tethered. Um,
you said that there are no good answers, but I would argue that there are clear, bad answers to the question.
Right. And as far as I can see,
and there are good answers, there's no right answer. I think like it's, it's, yeah.
Fair. Very good.
One of the, the concept of this law as a moral idea that Nigel Simmons promotes is that it's an archetype and you never reach the, you never reach the top.
I'll make a commitment back to you. Uh, I'm heading into a really interesting period. Going away on retreats. Again, I think values is gonna be front and center for my reflections for, for the coming months.
Uh, and I'm really thankful for, for that being one of the outcomes of this conversation.
I mean, I feel like this is the first of many, so let's, let's, let's, let's keep talking, please.
Yeah,
absolutely. But, um,
there's so, there's so much joy and so much kindness left in the world. We just need to really remember that, uh, that, that, that warriors ideally with, with good, with good objectives
exactly.
Are the, way to, to unlock that for everybody and, uh, and to do so with grace and humility.
Perfect. Thank you, Joe. Thank you so much. This was beautiful.
Thank you very much.
There we're at the end of the episode. I think it's so amazing that almost 70% of you actually listen to the end of the episodes, especially for long format. I'm, I'm so thankful for that. Like the engagement numbers is fantastic. Before we leave each other, two things, first off. This episode was sponsored by Grail and I'll make no secret about Grail is the company that I'm starting.
So it's the future of AI and how can we unlock value that's not just automation, but actually building towards a future that we want to live in, like a future where I want to work augmentation. So head over to Grail Works to know more. Secondly, I'd appreciate so much that you listened to this podcast.
Even though 70% listen to the end, almost 70% are not subscribed either. So if you can give, a like button a subscribe to all my channels. We're on LinkedIn, we're on Instagram, we're on TikTok. Apparently, uh, we have an awesome webpage and. There's so much more content, like yes, the reels and so forth. But what, what I really would like to push for is that every week I publish a long format article, which is the pinnacle of my thinking.
So it's the most interesting thing that I come across that week, either in podcasts or in other endeavors. I try to write about it towards you, a senior leader, like what is the crucial takeaway from the most important part of my week, let's say. I publish that on LinkedIn and on my webpage podcast. Thank you so much and see you next time.
That was Joe Braidwood on ThinkRoom — where exceptional minds think out loud.