The Inner Life of a Responsible Tech Practitioner with Benjamin Olsen

There's this quote I love the German phrase I stick on top of LinkedIn, which is, whoever sees more is more right, this old proverb. And I think to make it concrete, just in the daily life of the response, one of the major response, I practitioner fundamentally, I cannot do it alone. And so my daily life looks like I work with dozens and dozens, sometimes hundreds of people on problems. So it's at the core coordination problem, even with the fundamental group of folks building the technology, right?

So a feature comes through like the most innocuous kind of things on your operating system, right? How can you make that experience in that process super responsible the sense of we are going to bother beyond to make sure that this is a safe experience and what that means in practice are things like abuse, monitoring or content moderation and so on. And so what looks on the surface to be an innocuous piece of technology can be used in all sorts of terrible awful ways that one does not want to imagine, but practitioners are responsible. Tech these days are asked to imagine those things.

Any piece of tech, any piece of something we've invented can be used in ways that we can't foresee sometimes, but again, whoever sees more is more right. Welcome to the second episode of humans on the loop. I'm Michael Garfield and you're with me on my exploration into wisdom and technology, agency and automation. And what happens when we ask not whether we can do more with our tools, but what is actually worth doing if we can do anything?

When technology gets Byzantine, when the heady early years of cybernetic mysticism give way and our software engineers become the new priests of the Catholic institutions of big tech, maybe we can learn a thing or two from a Byzantine Catholic who's made responsible technology their life. This week's guest is just that person. Benjamin Olson is the head of Windows responsible AI and data compliance at Microsoft, where he also pioneered their first AI and ethics education programs. He's also an advisor for AI and faith and has worked as co-chair of the World Economic Forum's responsible learning and education program and member of their responsible development and deployment of technology steering committee, the former responsible innovation lead at meta and a part of the IEE's working group on responsible AI.

His online courses in analytics, data science and responsible tech have been taken by millions of students in more than 120 countries, but it's his writing at the intersection of religion, spirituality, technology and human flourishing that caught my eye. I met Ben through Andrew Dunn of the School of Wise Innovation, where I was on faculty for a course on embodied ethics in the age of AI with Josh Shri last spring, and was immediately taken by the clarity and heart he brings to places I have always guessed were frankly soulless. Speaking with him, give me hope that maybe all this hype is actually the evidence of earnest and concerted effort in some corners anyway, to do the future right and not just big. I hope that you enjoy our conversation.

Okay, awesome. Benjamin Olson, you are on the loop, I guess. How are you feeling feeling on feeling ready? Yeah.

Good. Why don't we start by talking a bit about your history? I like situating this in the human narrative. So knowing your choice of background story and specifically how that brought you into the work that you're doing today, like how you came to care about the things that you care about with respect to wisdom and technology is in my book, the right place to start.

Yeah, excellent. Well, I'll start with, I never intended to get into technology. So I was a philosopher by a training and my undergrad, yeah, I barely made it through with that degree. So I had studied a little bit of everything.

Like I was at this tiny liberal arts college, say history, classics, philosophy, theology on and on and just re-inveraciously and with no intent of monetizing that skill set and found myself out of college with a degree to be five years because I booked a one-way ticket to Rome. That's kind of another story, but we can go there if you want. But I became clear I needed to also make money after college. And I had a girlfriend and thought marriage was a good idea.

Still is. So not spoiler alert, but I found my way into a little tech company, a consulting company. I sometimes say through nepotism. So my father did have a little consulting firm and it was clear at the time that some philosophy skills were transferable even 10, 15 years ago, right?

Like logic and set theory and all that. It was easily translatable into programming into at the time I was doing data analysis and SQL and so on, right? And found my way through that with mentors, with kind of the school of hard knocks of Microsoft with really tough people to work with. But at the same time finding it interesting enough to keep going.

But through that time, I ended up teaching at university a little bit. So I really loved data visualization and that side of the work and the arts and science and then started to dribble in things like the ethics of data visualization and like how you can line with statistics and make charts, do all sorts of funky stuff. And so that led me to teach even more. So I built a couple of online courses for and this is after I worked at T-Mobile, Costco, a bunch of other places inside of technology and technology adjacent companies.

And when I took those courses on, I thought, there's a different way to do this. So there's a different way both to teach and to tell stories and bring data and technology together with philosophy, with art, with culture. And so that was the winding road to building MOOCs for a time. And so I built Microsoft's first AI and ethics course.

I built many data science and AI courses and quote unquote, top millions of people. And that was a great time. And it was also where I learned that scale sometimes doesn't really matter. So like, yes, I taught a million millions of people, but what's fulfilling?

What's why am I here? And those things started to coalesce into bringing that AI ethics work internally to Microsoft. And found myself at the nexus of really the first wave of responsible AI at the company. And that was a trip.

So I could go deep into that. I could fast forward, but really that was where we started to call ethics, responsible AI, responsible innovation. Those terms weren't widely used when I first started in this field. And so that was even just six years ago.

So that does you how fast things have moved and how it's kind of the lingua franca of today, right? It's everywhere. And people care, people are intrigued and scared and all sorts of things now with AI becoming what it is. So I was just having a conversation yesterday with Kevin Clark of content evolution.

Does advisory work in leadership spaces. And we're talking about a kind of simplified mass loss hierarchy in which we have survival success and significance. And I like that. I also want to deconstruct that a little bit with you because it strikes me that the more intertangled we have the powers of artificial intelligence.

And let's for the sake of this conversation, I want to take a more general definition of AI to consider algorithmic optimization processes, which would include the revenue models and organization structures of the businesses themselves. So like thinking of human institutions as a kind of prototypical AI, I think helps us get into a framework in which we can generalize how to think through the kind of constraints that you think through in your daily work in a really meaningful way. But yeah, your point about responsible AI being something that people weren't spending nearly as much time thinking about a few years ago seems to be due to to be kind of cynical about it. The existential threat to companies, to their survival and their success of these news stories that show up like the infamous story of Google photos confusing black people with gorillas.

Suddenly they have a bottom line problem. And that's maybe the only way that these kinds of ethical concerns are actually legible to the organization itself in like a James C. Scott seeing like a state way. They only have so many different kinds of inputs.

They only count so many different kinds of things. So yeah, I'm curious about at the level of the organization and then also at the level of the people that come through these conversations with you that work alongside you. I know that there are probably people for whom ethical concerns seem just a bother trivial or a distraction from building truly excellent products. And so I'm curious just perennially about this question of how do we make the significance level apparent to the survival and the success level or another way of putting it that you put in a lot of the work that you've written on and shared on LinkedIn.

Also a perennial concern of mine, which is how do we render outcomes in much longer time horizons hundreds or thousands of years? How do we make those outcomes visible to processes that historically we're only concerned about the next financial quarter? So that's all related and I'd love to hear you riff on that. Yeah.

Well, first I'd say organizations are going through this learning, growing, changing process where the language of ethics, the language of like you said, significance and meaning is not something that's come naturally because to your point it used to be so simple. Like it's like profit and loss. If you fabricate something, you sell something. You have to imagine folks working in the automotive industry or what have you back in the early 20th century.

We weren't really thinking about downstream effects of the consumption of fossil fuels, even the tail of like it could have been otherwise we could have had electric to begin with, but it went a weird way. And so I think of this lack of an ability to see those downstream effects is, yes, on one level in immaturity, a lack of desire, but it's also just like infancy. And so I think of it like there's this babyhood that corporations were going through and then infancy and now they're transitioning to toddlerhood or something, right? We're looking at these many dimensions of what it means to be here, doing what we're doing.

When it comes to business, when it comes to working together, it's materializing in so many interesting ways. I have in my mind, it was a funny quote from like a priest for Breeze from Comedy Bang. Nick Carl saying something like, on the directional, we have to be omnidirectional in our work. And that's crazy making, right?

Like we have to think even in the microed terms of like team interactions, that would be inclusive and build everyone's voice into the process. And that means our stakeholders, that means like downstream effects, quasi-cultures, so the many micro communities and macro communities, we affect through to like the 50 or 100 different conceptual lenses you can filter things through. And it starts to become just insane, right? And so I think what in organizations, and this is true of mine, this is true of Google, and I spent time at Meta, I can tell you a little bit about how we went through that at Meta for responsible evasion of bringing in these different prisms and lenses of what to look at.

And we would call them dimensions, right? So omnidimensional, multidirectional, looking at through like, yes, trust, safety, privacy, security, responsibility, but then there's planet, right? There's communities, there's freedom of speech. Some of those are what came through at Meta.

And something I'd care to be about is freedom of religion and freedom to practice your faith. And those things, like all those did come to the fore there at Meta to begin with, but that team only lasted about two years. So like the team went away and they got dispersed and there's a whole story there, but just to say it was difficult to translate that into product at various times. And now in Microsoft, we have six, right?

We have six response way, I have principles, those principles, I can rattle off and use all the time. It's six as easier, remember than 12 or 20, but it's still hard, right? Like you still have to think about the relationship between inclusion and privacy and safety. And sometimes those can, but heads in interesting ways and there's tensions there.

You know, even before we started the recording here, we talked about like open source, right? Like how do you think about open source and all that in relation to the ecosystems you work in? And when you really push on it, some of those frameworks and those philosophies like don't reconcile at the end of the day. So like open sourcing personal data is a no go, right?

You can't just open source a bunch of personal data online and expect folks to be happy with it. Those are called breaches. Like those are called, today we call this breaches in part because they're behind a wall of safety, right? And privacy and those are expected, those are fundamental human rights as we talk about.

So what ends up being elevated to that level will open source ever be that. I think you have to pull on the thread of how the concept is related to all the other concepts. And then if you really want to get crazy, you start to play with those things like you mentioned, like time horizons and you push it out to say, how can we know beyond the next six months to a year, where things are going? Ultimately, what should we do with these omnidirectional, omnidimensional considerations against X timeline, right?

And I think what I do there is like starting, I start with things like bodily practices with like being the most centered, most careful, prayerful person I can be in relation to making those decisions in the context of my role, which is very small and the large scheme of things now. I used to feel like I was a little bit of a conductor, right? Like I was videoing the world's top superstars. I was pulling the strings behind the scenes, making sure that things were coordinated and going the way that it should for responsibility for the world.

That hasn't really set Microsoft, which was, so we're adding a lot of those things. But yeah, where it is now is almost a journey of humility of like, OK, I'm just this one person and what to do. And I can unpack that in terms of my role. What should I do?

What do I think others should do if I can be so bold? And it relates back to that humility, that fundamental sense of who I am and why I'm here. And only then can you start to be omnidirectional, multidimensional, omnidimensional. So I want to think through the humility component with you, you mentioned Yolanda Gill, a computer scientist and her notion of super responsible AI.

You talk about an example she gave at a conference in Hawaii years ago that we want cars that drive better than humans drive cars. We're not actually going for human level drivers. And Kevin Kelly has made a similar point about what is the point of human level AI. Why are we trying to reproduce a human thinker when we have the capacity to develop all of these other modes of cognition that could supplement or complement our own, the calculator is already super human level in its own respect.

But this question of how do we even strive for systems that are more responsible than we are? I'll just point to Doug Rushkoff. And a point he's made on his show, Team Human over and over again, is that he gets called into these big tech companies as a consultant and the people inside of these companies are like you. Like they are aware of the externalities that are produced by these institutions, want to make a positive impact in the world and they are inside the machine.

They are, when I was at the Santa Fe Institute, the last episode of Complexity Podcast that I hosted, 106, was all about the tensions between us and the thing in which we are operating and how a lot of these are kind of irresolvable tensions. Because for instance, if you have a learning network and you want that network to learn to be adaptable, then the individual nodes of that network have to have shorter memories. They can't be stubborn and recalcitrant. And so there's a certain amount of, you can think of it as employee turnover, right?

You hear this again and again in stories of large organizations that flush out people that are the storykeepers of those organizations because those are the people that are preventing whatever kind of change wants to be driven from the top. And so I want to ask you about what it's like in daily practice on the ground in your work, but I want to do so through the aperture of this fabulous quote you pulled from Martin Luther King Jr.'s Nobel Prize lecture in 1964 where he says, we have allowed the means by which we lived out distance, the ends for which we live. So much of modern life can be summarized in that arresting dictum of the poet's throw. Improved means to an unimproved end.

And this is where I'm angling toward your writing on responsible innovation and the inner life of a responsible innovator. I want to start by asking this question of like, why are we doing this? And then also, how do we even know? Because it strikes me that there are people in Web3 that are thinking this way about quadratic voting and like, well, actually we can design systems that aim for properly capturing and referencing a true diversity of perspectives.

And that's how we might come upon something that we would collectively agree is super responsible. But yeah. Yeah. Well, goodness, there's two ways to go there.

I think of it like what you mentioned about the Web3, where I'm going to see this startup called Holis. Have you heard of this? It's one of these nonprofits out there or I don't know if their structure has changed, but the concept is beautiful in that way that they're like, we have to solve this democracy problem with technology, married with the latest in algorithmic developments and getting people involved and active in the democratic process. And so we'll pull people, we'll create knowledge graphs, we'll integrate it with like actual countries and systems.

And it's a beautiful vision. And I bring that up just to say, if you look at any specific combination of considerations right now, it feels like there's a startup or there's a nonprofit, there's a company trying to do something about it. And I think that's one of the things that gives me hope in this moment about who we are and what we're doing is it's like, we're nothing but opportunity seekers at the end of the day, carving out a path of as much as we can see. And so I think there's this quote I love in the German phrase I've taken on the top of the LinkedIn, which is, whoever sees more is more right, this old proverb.

And I think to make it concrete, just in a daily life of the responsible innovator, responsible AI practitioner, a leader is fundamentally, I cannot do it alone. And so my daily life looks like I get on calls, I get in rooms, I work with dozens and dozens, sometimes hundreds of people on problems. And so it's at the core coordination problem, even with the fundamental group of folks building the technology, right? So a feature comes through like the most innocuous kind of things on your operating system, right?

How can you make that experience in that process super responsible, the sense of we are going to bother beyond to make sure that this is a safe experience at large. And what that means in practice is are things like abuse monitoring or content moderation and so on. And so what looks on the surface to be an innocuous piece of technology can be used in all sorts of terrible, awful ways that one does not want to imagine. But practitioners and responsible tech these days are asked to imagine those things.

Red teaming has a, it's a pretty benign name, but sometimes can have the feeling of looking at the worst of what humanity has to offer. Any piece of tech, any piece of something we've invented can be used in ways that we can't foresee sometimes, but again, whoever sees more is more right. Like we, you can foresee that there's a problem when you put something into a process and you think about it, like, okay, how many ways can it go wrong? And then you work with a team to iterate through that.

And then through that process, you figure out what are potential mitigations? What's the space of possibility? What's the tech opportunity to solve this holistically? And for those things, like I mentioned, content moderation or abuse monitoring, those aren't always perfectly preventative measures.

Those are sometimes post hoc. Those are ways to ensure that we can trace things or make sure that things are being used responsibly in terms of use. But when you put something out, let's even just say like file storage, you have it on your phone, every computer, every digital device will have some sort of storage and compute and so on, right? You can put all sorts of terrible things in there.

And yes, there are protections. But there's space for people to do terrible things when there's possibility. And I just go, okay, like with the tech available with what we have in terms of our processes, with experts, we have in safety and privacy and fairness and inclusion and reliability and child safety and regulatory compliance. That's another thing we can talk about for days if you want to, but it can be pretty boring.

But the thing is, all that comes together to really shift and shape that experience to make it so that we have what's in place today to do our best. Now, is that what I categorize as super responsible? I think it's certainly very responsible, but my dreams, it would be something like, let me just share a vision of what this could look like if we're really different. If you had a group of technologists with ample resources, with a proper platform, who could meditatively prayerfully consider what's being done and with great skill, bring like many stakeholders along, maybe even bring some sort of democratic process onto the table, you could bring people from all sorts of communities to give you back through out slowly and as tech progresses, as these guardrails progress, you can add them in.

And there's like a 50-year rollout for a future. As we analyze the downstream effects and understand the tertiary effects and the butterfly effects of all of it for a single feature or maybe many interacting, what are like momentastic, excuse me, corporate structures of tomorrow that make that possible. And I just have no idea. But to me, that's where I wish it would go.

Not that feels very possible because of what you mentioned of pressure as time horizons, how the work is done. I certainly know that by virtue of being where I am, I'm complicit a bit with things that have happened, with harms that have been done, with me supporting or being participating in it, there's blood on my hands. And what to do with that? And I mentioned this maybe another post where part of my job is as a sin eater, right, metabolizing the worst to make it the best.

And that's not always put in the job description. Like it can be such a tough aspect without the proper foundations, not using the word proper, like a deep well to draw from to give you life and sustenance that's apart from the work. And that's what I meant by the inner life of a responsible tech practitioner. Yeah, I want to edge a little bit more fully into that article.

But on the way there, I want to embroider a couple of things here that are starting to shine through in this question of super responsibility and its relationship to the humility that you're talking about in ethical tech practice. And it occurs to me that the reason that perspective diversity matters in product design is, well, in part, I spoke about in the introduction of this series on Substack. William Gibson has that great phrase, the street finds its own uses for things. And like you were just saying a moment ago, red teaming involves a small group of people trying their best to imagine what billions of people are going to come up with.

So in a way, part of this is that the humility comes from inventors realizing that they don't actually know what in the Stafford Beer Sense that a system is what it does. They don't actually know what they're inventing. It's like in 20 or 50 years time, it may turn out that this thing that you thought was helping you do your office work is like critical Mars infrastructure or it's like it was World War Three or whatever. The other is that super responsible.

I love that quote, who is right is the one that sees more. That suggests that super responsible means we don't even know what that looks like. This is a common thing in Eastern spirituality, this notion of the crazy wisdom or that Yoda, like this character that is inscrutable to Luke Skywalker. And it was like, well, it's because this is a wiser being than you are.

And so you get into these weird, a question like the question of how would we even recognize artificial super intelligence, right? Or another one that I love, which is why did it take us so long to even consider the possibility that corporations were actually intelligent entities in their own right? Does a rabbit even know that a wolf can think in the way that these kinds of things? And then I want to lean on this idea that another part of the humility in this whole process comes from the fact that for these reasons, you cannot decide in advance what super responsibility is, right?

Like as you put it, the creation of AI systems should never have as its goal to be perfect, rather a system and an individual or group creating or using that system should play citing James Beacars, an infinite game of evolution and growth. So again, it becomes much more rigorously a scientific thing rather than we set out to prescribe the conditions whereby this thing that we are aspiring to can be defined, that it looks much more provisional and you're not doing ballistic calculations anymore. You're doing complex system science and you're running simulations and you're constantly going back to adjust the parameters and look for new variables. Like you said, that's where this question of content moderation about, as you put it the cilla of censorship over blocking, over prescribing and the carib this of under blocking or allowing people giving people even more leverage to come up with harmful or exploitative ways to use the tool.

It's interesting because, yeah, like with a child, for instance, a lot of my thinking on this is related to Hans Moravic's mind children and like thinking about technologies as children with their own destinies. And I know it like you said, you're a parent and it's just, yeah, like tell your kid who to be. There is a certain amount of, for lack of a better word, faith that comes through in this process. And so to me, an honest accounting with these considerations does lead us into a corporate practice in the years to come that does look, however secular, isomorphic to a monastic lifestyle.

And yeah, I'd love to just, here you go on that. Yeah, a couple different frameworks are swirling around my head and some examples. I'll tie them together loosely, real fan off the super responsibility concept because I do think in certain cases we can articulate what that super responsibility is, right? Like we have the car example is easy.

It's almost like no death. Like how about no death and super fast transportation? That'd be nice. And also no emissions that'd be even better.

And I really want it super comfy and like extremely private. I want my transport to be private and comfy and certainly espresso and can be a little vehicle I can do intimate things and or look, I can, on and on like maybe there's going to be that in the future and that could be pretty cool. But I think of it like for some of these other ones that just aren't so clear cut actually for most of all the other ones. Do we know that we need them or do we know that they're good?

Like everybody's asking that anytime any big tech company comes out with something now and they should be and they should be and we should be asking the same. But so while it can be hard to anticipate or imagine what that super responsibility would look like, I think it's holding that possibility space open while you settle. You have to and I think of it like this content moderation of use monitoring thing. You want no harm to come from the use of any feature and yet from the dawn of time we've accepted risk as part of our life and no risk, no reward.

Where sometimes failure ends up in harm, sometimes it ends up in losing a limb or losing a life. What we care about as a society somewhat funneled through this concept of heroism. We want ourselves to be a protagonist, we want our children to be heroes, we want our children to be presidents, me I hope so. But the wish materializing as the highest levels of possibility, potentiality for individuals, for communities, for societies, that can look as real as we want it to.

If we forecast without everything we have, like I think of it like your musician, your artist, your calling forth, the spirit as you feel it, went to the present and in a dynamism that moves people to experience new things and feel new things. I think of it like technology as dance as spiritual practice is music. We don't think of it that way, the royal we, but some of us are starting to and maybe one additional frame to think about is I've been thinking somewhat about this like opening eye stages of AGI. So it's like, okay, we have what we have now and we build onto agents and we have agentic architectures where AI is here already.

It's not like that's far in the future. In fact, I'll riff on this in a sec because the next stages are AI acting as team members or teams and then the last stage is before AGI or whatever term of art open AI wants to use because it's like the organizations are fully AI'd. You can have a company or a corporation as an AI in all this constituent parts and then you think it gets back to your point about a small group of humans trying to predict what a billion humans will do with a feature is crazy, let alone and you throw agents on top of it and AI is interacting with it. What's gonna happen?

Who can imagine that is combinatorics of that situation? Obviously no one. Okay, so, but that whole framework is so thin, right? Our goal for AI is like people, teams, organizations, isn't that a great feature?

Like I want more corporations in the world. That's a sad dream. What's a better framework? What's a better framing?

But what? Why else are we here? We're not here to just feed the mega structures that are, right? Yeah.

Now seems like the time to soften into something a bit more imminent. You've touched on this at various points in this conversation. Reading the piece you wrote on the inner life of responsible innovation on LinkedIn, I'm frankly just kind of shocked that people like you exist that are sitting there, holding their awareness on their breath and this attentiveness to the present moment while in corporate design sessions. And I would love to hear more about your inner life.

We've talked a lot about ideas on this, but I want to know more about as you put it in this piece, how far might we go together if we invest in equal, if not greater quality and quantity of attention to the flourishing of our inner life as technologists? Just this entire show, this project emerged out of a concern that I experienced while working in an AI startup. It was like, I think this product is amazing. I can see what the founders of this organization see, how they imagine that this is going to help fulfill the original vision of the founders of PCs and the internet, how it's going to help us communicate better with each other.

But again, we spend so much time improving the means and we're not reflecting or as you put it elsewhere, we can do better than reflecting, we can illuminate the ends of all of this work. And so there's like, what is it to be the one who is slowing down and cultivating one's attention amidst all of this engineering? Yeah, holding that together as I can. I think just, right, start is, there's always a talk about like, you think a lot more about yourself than other people think about you.

And there's a framing of like, it's so easy to over blow that into something that it's not. And in the daily life that I'm trying to live, it's just not about me. You have to do the stupid LinkedIn self-promotions or you have to still have a career and all those things. And maybe this goes back to my point about being complicit in what we build and the harms they're in and all that stuff.

There's the baggage we carry there, but I think the potential in the work is this inter transformation, no matter the work. And I'll bring an example of somebody I've been reading about recently. So it's this, this is this pre-spec in the 20th century using this water-water chisic. And the water-water chisic had this dream, this vision of being super responsible and going to Russia and helping them out the time of 20th century madness, right?

I want to help the Russian people out and minister them in the words of how Catholics framed that stuff. And I think of the, and served them. And what happened was, yes, he found his way to get there eventually through Ukraine and snuck into Russia and became a priest over there, but was quickly captured and put in solitary confinement and tortured for five plus years and really broken. And yet during that time, he had his practices and he had a formation in the Jesuit tradition, which involved like a daily prayer rules.

It's called and these processes that he went through. And through that time, yes, eventually got out of solitary confinement and then was put into labor camps for like 20 years. So quarries and he became a mechanic and like, you know, so how do these weird jobs? And obviously just an amazing person that could even survive such a thing, let alone them thrive and find his way through.

All through that time, he had this inner life, this kind of secret life force, this power to draw from that wasn't just technique, right? And I think this is something where, and I'll bring a full circle in a minute, but I think like right now in the circles that I've been in, there's a lot of emphasis on technique and doing the right breath practice and taking the right psychedelic and having the right guru and the right set of practices or beliefs or even no beliefs and that's the best way to go and there's all those sorts of things. But for me, I think through my winding way of like, I've left with all of them for a while, became a Buddhist teacher and meditator and because that's what American Buddhism does is you can instantly become a teacher. I fell into that trap, but I think doing hard practices, working on or calling the Brahman Bajaras, so these infinite without measure kind of practices where you expand the heart and open the heart.

And for me, through that process, really deepening that practice at the other side of it, I kind of came out along with some other things. I actually think that there's more that exists than meets the eye. In fact, I'm not here because I've willed it. I'm here because it's been willed that I am here and I've been given this life to serve, I've been given this life to create and have conversations and all that for a greater purpose for I think in sharing this with Father Walter Chisak for the glory of God.

And what that means is I hear the stories and I'm like, you know, I'm like, anything I go through at work has nothing to do with the direct suffering of a Soviet solitary confinement torture session. And yet, like, why am I not building that foundational set of practices and finding joy in the process? I could be a mechanic and I could do the same thing. I could be a responsible AI team leader, but a big megacorp and do the same thing.

It doesn't matter where I am. It comes back to this concept of another recent saying it was Catherine Doherty of this concept of the duty of the moment. What is the duty of the moment? What are you being asked to do right now instead of pretending like you have all the answers, this goes back to like the only way to be omnidirectional or omnidimensional is to inhabit the life of the one who is omnidimensional and omnidirectional?

Oh boy. OK, so let's see if I can pull this off. My friend Larry Mulstein, who works at a somewhat similar layer to your responsible AI at DeepMind. He and I have had a lot of conversations around his ideas about technological love.

And as you say in your inner life article, responsible innovations, real name is love. And if I think about it's funny, my kids are obsessed with this YouTube channel, Danny Go. And Danny Go has this episode of he's going around asking all of his friends what love is. And we're Danny Go to ask me, I would probably say something like what you just said, which is attentiveness to the needs of the subject of that love and the kind of more esoteric Christian, I thou sense that I was on the fence about whether to make my recent interview with Timothy Morton on their book, Hell in Search of a Christian Ecology, an episode for Humans on the Loop or for Future Fossils.

And ultimately went with Future Fossils just because it got extremely weird, which I love. But more directly on these particular investigations. And one of the things that Morton says I've heard alluded to several times in what you've discussed here is that evil seems to always come from people trying to do the right thing from some idea of the good. Like you think about the decimation of the indigenous peoples of North America as the fulfillment of manifest destiny, right?

Is people that were on some sort of spiritual quest, the Crusades. Like everywhere we look, it's an ideology that has a sense of identity associated with a particular telos, like it's going in this direction. And in going in that direction, you end up with, you know, what Carl Jung called an anti-adromia, which is you end up creating the opposite of that which you have intended, right? And so there's something about the difference, and I guess this is what I was saying about kids earlier, it's like you may have an idea of what is good for your kids.

And that's a very different thing than simply attending to your kids. You may have an idea about what is good for the world. That's different from actually listening to, and again, this is where the openness and the sort of ongoing infinite game and stuff comes in. It's different from actually listening to your customers and then also the people that turn out you are interacting with, you're serving in some way that you didn't see as your customers.

It's about finding the kinds of value that you didn't realize you were creating in the world. And so when Larry and I talk about technological love, it's about, again, who sees more is right? It's the ability to expand one's sphere of moral concern and to realize I'm serving the entire cosmos through this interaction I'm having with my spouse or something like that. And like how that shows up in a responsible AI is especially fascinating because there's a sense in which this sort of mass capture fishing approach, like AI is already trying to do this.

To some extent, it's just botching it. Every day there's a new instance of some situation where people that are trying to serve the world with better AI products are buying up data sets like the Reddit deal, right? It's like, whoops, those people 20 years ago, when they signed up for a Reddit profile or whatever, they didn't think that it was going to be in service of the feeding and caring of some machine dragon that they don't approve of. And so Morton talks a lot about how permission, like you can't have love without consent.

And so that's where I would love to hear you expound is in how this kind of present moment being with necessarily involves an intersubjective agreement between the organization doing that is trying to love well in the creation of their products and in the actual concerns and consent of the people that they, or beyond people, the planet that they claim to be serving. From the modern conceptions of consent, like either you have GDPR, right, that defines very specifically like what consent is, according to general data protection regulation in the European Union, right? And it has criteria like it needs to be specific, like it needs to say one thing. It's like folks need to know what they're consenting to, right?

And there's this kind of transparency, there's this, you know, directness about that definition. And then the handshake is made, then the data can flow, the spice will flow, right? The experience will flow, right? And that's kind of the concept there.

And having listened to your conversation with Timothy there was wonderful. Just hearing about expanding that to religious frameworks and territory I haven't thought about much before, but once I started playing with it after the fact I was thinking, it's good. Consent is good, obviously consent forms a foundation of a lot of social contracts and things that we care about. And again, why are we here?

What should we be building? What do we want to be doing? And there's a consent in the beginning of a relationship where like, yes, I will date you, yes, I'll get married. And then I think you were talking about how do you serve the university interaction with your spouse, right?

And I think it's inherent to the definition of a good spouse to, on day 7,201, our 14, you have an opportunity to support her after something scary happened with one of your kids. And then you just you just shut it down. You were terrible. And at that specific moment, things crumble and like, hey, that never happened to you.

It never happened to me. But the technology follows a similar pattern of like trust is gained and drops in lost in buckets. And this, I think, consent without ongoing trust and ongoing safety without the beauty of the container being this place that not only is safe, but also the place to play and to explore and expand and use for, like you're talking about before, the better things of life than teams and organizations and corporate mega structures, like dance and art and living together community and those better things. The frame of consent only gets you so far.

And I know means do I think that because I'm in a consensual relationship that's enduring, aka I'm married and I have that fundamental commitment, like that there's a lack of mystery or surprise or even a possibility for me to be terrible or for me to rise to the occasion of being like who I'm meant to be in this relationship. The corporations need to think that way too. And I think that's why there's a, I'll tell you like a new thing I care about a lot. So there's this idea of multicultural, multilingual effectiveness of AI features, right?

And I think of it like lots of companies now are funnelling different languages through GPT or different foundational models and thinking like that's good. I have English now instantly have Italian and Chinese and Japanese and Korean and that's it. Maybe I'll slap a consent in front of it. But use this was a trained English, but we did this and use it for that.

It's in your language. Good luck to me to tie back even to like the super responsibility concept is like there's an infinite game of how good you can get with a feature in a different culture and different language. And that can look as extreme as the 50 year, 100 year rollout model. I imagine as like a monastic tech practitioner, sure, talking every single person in the country and see if they want it and whatever that would actually look like.

But I think there's just always more you can do. And I think it's important that we at least try to make any feature that's global for AI to be globally responsible has to not just filter things through GPT, but have like sociologists and linguists and people who get compensated appropriately for giving feedback on things like block lists or content moderation or keeping features safe and making sense when you search for a cat. You don't get something weird with because of strange embeddings and different languages. And not only that, this goes back to the principle of not being perfect, but still aiming for perfection.

There's that paradox and it's just part of this practice to, to, I think to your point, even just to have that unimaginable open ended perfection that's possible while not being scrupulous. And the spice of scrupulosity kind of bumping heads with a practical virtue of it's good enough. How do you navigate that? And that's the human magic space.

Just if you happen to be a corporation and you're listening, one thing I would love is to get paid for surveys. That would feel more like a marriage, right? But it's not like you're just trying to figure out how to extract. You're not just trying to figure out how to get me in bed.

You're trying to figure out what makes me happy in bed and that's paying me for my data, which by the way, I don't know if you listened back to, I'm going to create a list of precedent setting future fossils episodes for people that are following this series. And one of them is the conversation I had with Amber Case and Michael Zargum, in which they talked about terms we serve with, which is this project that they're working with organizations to normalize retractable permissions. So that if people, if users of a platform ever do feel like their trust has been violated, they can withdraw their data. They can withdraw that consent.

And I think that that's huge. That's very much in line with the metaphor that we're playing with here about what it means to give consent at the altar versus how do you make sure that this is still a thriving love relationship 20 years in or whatever. But I only have two more questions for you today. And one of them is just because this is a show about agency in the age of automation.

And again, you've made comments about this and about your own will in the world, your own choices versus accepting this is where you're going to be. The world has put you, but I would love to hear specifically about how you exercise agency in a world that is so enormously constrained. The quote unquote social contract that we have now as modern beings is so multi layered and often opaque. And so just knowing how you work to exercise agency in your own life and how you might encourage others to exercise agency in there specifically, if you have ideas for this, that people may not have considered ways of thinking about how to bring more deliberate and intentional action into the way that they interact with these bizarre unknowable super structure.

Like if you don't like nobody reads the user agreement, come on. So how do you exercise agency in that? Yeah, all the do my favorite thing, which is pairing the megamistical with the super practical. So I just think of like on the one hand holding this understanding that on the level of decision making an agency, I feel like there is a side that's like very practical and transparent and we can kind of study us like we study monkeys.

And there's an approach there of I think of the decision science work in Stanford for like Jip and Dan Heath comes to mind. I studied decision science a little bit at Stanford for a certificate program because it sounded like fun when I was doing my data stuff and really influenced by them. And I love this little framework. That's just something that feels so right for the kind of work that I do and that you're doing.

And I think of it like the acronym called wrap. And so it's like folks, before you make a decision before you want to do something to maximize your ability to have just that mode come a little bit more agency in the context of the thing you're working on as to widen your own life. And then you're going to do something to make sure that you're in your options to reality test your assumptions, attain distance before deciding and prepare to be wrong. So those four things.

And you do that. You're going to do pretty well in whatever you're facing because I think of it like it's a pithy way to wrap your decisions and wrap the experiences. And to me, I just love that because it gets to what we're talking about. It's omnidirectional, omnidimensional side to when you're options.

Reality goes deep as you want with that statistical way down to embodiment and a world soul and whatever else you want to say. And then you know, attain distance before deciding is providing a space and I'll pause here for a second because I actually think this relates to my more mystical comment, which is like I think agency. Sometimes my presuppose that we know who we are and we know what this is. And your mystery and yourself and who you are as being unfolded in time.

Who you are is only known by the timeless and the thing in my tradition. It's sort of like you have a secret name that no one else knows written on a stone at the end of time. And so this is cool quote of truth. It's like snow in the hands of one who melts like snow in the hands of truth by the tsunami, what you are as a mystery.

And so who you are as a mystery and explore that while you also wrap stuff, wrap those presents. I can say other things, but that's my second. That's wonderful. So just in closing, because I don't feel like I know the optimal path with this series, I'd love to know who you recommend to me and to whoever is listening, the business leaders, the engineers, the ritualists.

Who do you recommend as thought leaders in this space, however you choose to confine the scope of investigation here? Who would you suggest to someone but say building AI or running an AI company to absorb as deeply as possible? Or who would you like to see interviewed? Yeah, yeah.

Okay, I'm going to cheat. So first Shannon Valor is somebody I always look back to, the recent author of the AI Mirror and Technology in the Virtues. It's been foundational in my thinking and Shannon's always just finding a way to articulate a lot of things that we've touched on in very clear, exciting ways and also just really cares about other things too. And I think she's a great guide for this age.

So I'd say that, but then I'm going to just give you two other names. She's talking about Dan Zigman, who's Zen Monk and head of Special Projects at Apple and two of the great things there. And then another person that no one loves is Zbika Krieger. So Zbika was my boss back at Responsal Innovation in Meta now is doing some really interesting stuff as a spiritual leader, but also a consultant or and also there's no buts here.

It's the both hands. So I think that's what's coming through your show is this both and finding a way to embrace these paradoxes and not just do good in the world. But why are we really doing this at all? And what should we do together?

Beautiful. Thanks a lot for sharing your software with us today. Thank you, Michael. Yeah, so good to be here.

Thank you.