Kathryn Hume: Full Interview

Hey Siri, play Machine Yearning.

Here's Forgive and Forget, Isabella Machine Summer's remix by the Kooks.

No, no, Machine Yearning podcast.

Hmm, I'm not finding anything for that.

Not again. It's Machine Learning from Assist. It's a podcast where we think and dream about the future of AI, the talking internet, and how we're reshaping our culture. Catherine Hume came to us via Robin Sloan, a great friend of the pod. One of the things we loved about talking with Catherine is that she gets the poetry of language. She's also really smart about how that poses some major challenges with AI. Katherine is also someone who can say, "I think math is cool," and you can tell she means it. Her blog, Quam Proximae, is a must-read. Katherine brings a wonderfully diverse view. She works at Integrate AI now, and has done product, marketing, investment, and mentored startups. Katherine gets that nobody really knows where this thing is going, and that's okay. But also, we can gather some clues about where we're going by looking at how humans don't change. We loved our time with her, and we think you will too. Enjoy.

Thank you for coming.

Thanks for having me.

So how have you seen the conversation change in the last 12, 18, 24 months?

Ooh, great question. I mean, there's certain things that stay the same and certain things that have changed. I'm a big fan, and when I'm thinking about even the future, and I'll come back to the question about like what's changed, of when you think, all right, so what's gonna be inevitable as opposed to being like, this technology is gonna get to this point. Which are normally just, it's a fool's errand for us to try to predict where the tech's gonna go. It's better to stay like, think like, what is gonna stay the same? So Jeff Bezos from Amazon, like he says, can I envision a world where customers are going to wanna pay more as opposed to less for toilet paper? He's like, nope, that's not changing. So therefore, if I make decisions on my tech stack that make it so that prices will stay cheap or decrease, that will be good for business. You know, it's like that's an inevitability. It's an assumption I can work with. So I think about that with regards to like the future of AI where it's super hard for us to predict exactly what's going to become possible and what's not. But there's a lot of aspects of human nature that are going to stay the same that we can use as like a grounding as we think through things. As we go for the past and like what's changed over the past 2 years—

What are the specifics of that, though?

Of like what's going to stay the same?

Of the human— You said there's some things that are gonna— there's some things that aren't gonna change.

Yeah, so I think, I mean, maybe this is just my own sort of beliefs about humanity. I don't think we're ever going to want a world where we are isolated and alone and we optimize our technology so that each of us can live in our own little bubble. As much as it seems great to have like the eye of everything and like everybody gets to be in their own little island, 'cause I feel like a lot of tech sort of is focused on the individual and like the power of the individual. I think we never want that, right? Like, yeah, and a lot of studies have shown that, you know, when you think about like longevity, right? What should we be doing? What should we be eating? How often should we exercise? What are all the little things that we can control in our lives so that we like live an extra couple of years when we're in our 80s and 90s, presuming that we're lucky enough to get there in the first place? And I think there's been a lot of studies recently that show that like the best guarantor of longevity is that you spend time with other human beings. You feel loved, you feel connected, you feel like part of it is you have a purpose when you're younger. When you're older, it's just like you're connected. Like someone's there. We're mammals, we're animals, we have limbic systems, and we like our hearts to be next to other hearts 'cause they can regulate together. And like, I think that will always be there. And I think that, you know, a lot of those, you sort of read big picture narratives around AI and we worry about the machines becoming smarter than us. That is focused on some assumption that the most human aspect of our humanity is this little part of our brain that's the prefrontal cortex that does like rational calculations. And like, God, that would be a horrible world if that's what we get reduced down to. Like, who wants that? So I think it's pretty inevitable that we are gonna really stay animals and like love the aspect of us, of our animalness, and that we can think about developing technology to support that as opposed to alienate ourselves.

That's so interesting. Do you think AI actually saves us from the machines? Because the last 10 years, we've really been programmed to just be addicted to the machine.

Yeah, I mean, I think if we make design choices where we want that to be the case, for sure. So I know in the latest Android release, they've now added a couple of features where you can sort of program— it's sort of like a dimmed color mode so that it enables you to sleep a little bit better. Or even you can decide. That you no longer want notifications after a certain time, or however you'd like to configure it so that you feel a little bit more autonomous as opposed to tethered to this machine that's guiding your every dopamine hit. So I feel like those preferences, it might be hard for us to determine exactly what we want from first pass and sort of wake up on a Monday and be like, this is how I want to live this week. I certainly know that I try to— I always want to shift some habit and fail miserably by Monday afternoon, you know? Like Monday morning I have high hopes and by like 4 o'clock I'm like, "Oh God." Don't we all? Yeah, with the inevitability of the 5 o'clock shadow, it's like done, you know?

What do you find the most or least surprising about the current state of AI?

Ooh, great question. Let me think a little bit. I mean, I think I find the least surprising but the most boring is our imagination of AI as robots that are supposed to be more intelligent than we are and sort of have an antagonistic relationship with us. Least surprising because science fiction has had a big impact on our cultural imagination. So like, and the science fiction writers, by the way, were drawing upon the Judeo-Christian tradition. So, and like the medieval and ancient epics. So there's like 2,000 years of intellectual history shaping what we consider to be the future, but it's kind of old and stale. So it's not surprising at all that that's what we're worried about, but it's also so restricting as opposed to our sort of opening up our minds to have more creative sources of inspiration where, I don't know, animals are super intelligent. Dogs— imagine being a dog and being able to smell the world they do. Like they do, right? And imagine like if we were to use that as a source of inspiration for thinking about some cool product, what would that look like?

How do you have such optimism for AI and how do you give that to more people? I believe, or how do you have so much?

I mean, I think math is cool and a lot of people aren't interested in like, you know, understanding the, all this is is linear algebra that ends up having these like miraculous Turing test like experiences where the Turing test is like, you see it, you know, a computer seems intelligent, A computer is intelligent if we mistake it for an intelligent agent, when on the back end all it is is math. So somehow modifying— we have to modify our educational system to make math more fun and not just this thing we have to go through to sort of develop intuitions. That said, I also think that there's a lot of concerns and ethical risks and things we need to be thinking about with AI. They're often different than the ones we worry about. It has to do more with some of the existing inequalities in society that are being— exacerbated and reflected as through a convex mirror with these tools. But I'm even optimistic about that because it's not the, like the linear regression is not evil, right? A mathematical model, it can't be, it doesn't even make sense when you say something like that. What it is doing is, as this guy Jonathan Zunger wrote on the internet, AI does not permit polite fictions. So we walk around with our like, nice little polite lies that we tell one another related to how well we get along with one another in society and like how much we've progressed since the 1960s. And these systems come along and they're like, nope, there's still a lot of systemic racism. There's still violence, there's still bias, like that's in the data. The data, it's showing us loud and clear. So we have to be very mindful as practitioners developing these systems of like, where things can go wrong. And I think that there's a lot more work to do to sort of, you know, get like sort of up the ethical awareness and consciousness of a lot of not just technologists, but basically like dialogue between philosophers, like policymakers, ethicists, technologists, et cetera. Can't just like lay all the burden on, you know, the engineers at Google. That's not the solve. What short-term feels like this, like, unbearable burden of trying to develop systems that are fair and not have all these mistakes take place is actually a wonderful long-term opportunity to try to become more aware of inequalities and do something about it. But again, it takes that shift in framework. If we view AI as a tool that we are empowered to design and build and shift and not this source and force we can't control. Machine learning.

We told you Catherine gets It's the balance of pragmatism and poetry.

Machine learning.

Machine learning is all about the ways we navigate a new world where identity and technology and commerce and security have to live together in ways we can't even fully imagine yet. We have some great guests lined up, but we want to hear from you. Who would you like to hear on machine learning? Catherine came to us through a friend, and we bet you have good ideas too. Let us know. DM us @assist on Twitter. Coming up, how math gives us an approximation of meaning and other things that blow our mind. But Katherine explains it so well.

What is the difference between AI thinking and natural thinking? Are they fundamentally different?

I think there's certain aspects of natural thinking and AI thinking that are the same and others that are really different. So one of the big capabilities and breakthroughs in AI over the past like 5 years has been in the field of what they call computer vision. So this is basically you show a picture, excuse me, you show a computer a picture of a cat or a dog or a glass of wine or like a baby, and basically it could come out and it can give you a label and say, "This is a cat." So then you say, "All right, so is that like the same way that we as humans recognize a cat?" Yes and no. So from the yes perspective, basically these systems are trained where they make a mathematical model. So when I say mathematical model, I literally mean like, If you remember back to when you learned algebra in high school and you had like y f, like a function. So some of those functions can be pretty simple, like y 3x. And unfortunately, that's not complex enough to do something like recognize a cat in a picture. But if you take it and you say y equals like— imagine projecting this out into 50 dimensions, which is a cool imaginative exercise, right? Like we can imagine 3. 'cause we live in them. We can maybe add on like time as a dimension or maybe color and get us ourselves up to like 6. But imagine now having to go up to like 50 million, you know, and it's like my mind gets blown. There's no way for us to even have an intuition. Like we can't like imagine that in our mind, but that's how many dimensions these functions are working with when they are presented with some super complex data and then they are asked to do something like, say what's in that data. So then we say, all right, so is that like or dislike human thinking? It is because our brains are super complex computational machines that can take in all sorts of information and do stuff with it. What we can do that computers can't though is if we're shown 3 pictures of cats, we kind of start to see the patterns and similarities between them. And if we see a 4th picture, we can get a sense of what that is. If you're a machine today, you need to see like 50,000 in order to start to get a sense of what catness is. So they can't construct a schematic representation of catness as easy as we can. So we come hardwired with some pretty cool processing tools.

And if so, what changed in the world that there's just enough machine-readable data to start training? Now, and all of our language is online. And so we're putting all this machine-readable text all day into an area where it can be read. Is that what's changed in the world of like— or is it the models have gotten better or the data's gotten better? Like what's changed?

It's a combination of the three, but I think it's fundamentally the first two that you mentioned. So computers are a lot faster. And also the shape of the underlying hardware has shifted. The, if you think about like Silicon Valley and what they call central processing units, basically chips that were like, you know, sort of running the technological revolutions of the '80s, '90s, early 2000s, they took electrons and they processed them in a line. And the math that's powering AI does things in sort of crisscrosses. Like, it's like, so it's beneficial to have these lattice structures where electrons can cross paths and move in and out of one another. And those chips actually came from video games. So they're graphical processing units and they were developed for purpose A. And then somebody was like, hey, what if we try this with this AI stuff? And they were like, oh shit, that just went like, we went from having to take weeks to train an algorithm to being able to do it in like hours and if not minutes.

Once it's trained, is it trained or does it always keep learning?

Depends on the system. So for some of them, like, you know, you sort of do this, huge training process and then let it loose in the wild and it's good enough for jazz. Most of the time, the phenomenon that you're trying to predict or model changes. So it's actually really great to start off with a model that makes a lot of mistakes, but then learn and get real-time feedback from the world in time.

Can we focus back to language? Because you said what changed about photos breakthrough in the last 5 years. But are there breakthroughs in language?

So there was this big shift over the past 5 years in terms of how statistical systems, machine learning systems are programmed that are enabling them to capture more of the nuance in language. So old school natural language processing, which is the subfield in like the overall umbrella term of AI that focuses on language, basically was built on the— so the task at hand is language is a mess. It has so much nuance. There's so many, like, nitty-gritty details that we process when we extract meaning from some utterance and, like, sort of sequence of sounds. And in the past, what a natural language processing practitioner would do would be to take a messy sentence and sort of break it down into parts. So it would be like, this is the subject, this is the verb, this is, like, the object, right? So things that you learn when you learn grammar when you're in grammar school. And then they would use those parts and sort of put them into a database and start to find regularities on them. So there was a huge step that was the human mind coming in and taking something complex and making it more simple so that it could become tractable using computers.

Do they call that entities?

Yep. There's entity resolution and extraction and stuff like that.

Got it.

And, um, the shift in the new deep learning paradigm is we say, let's go bottoms up as opposed to top down. So, let's stop trying to impose meaning onto things and sort of tease it out using structure and just feed in millions of examples of real live sentences and allow like the complexity to stay and not reduce it down into these simplifying assumptions and just allow the computers to sort of pick up those patterns and be able to execute tasks like translation or sentiment analysis, whatever that may be. And therefore we're enabling, like we're keeping some of the complexity in the language in the model. So it's able to pick up a lot more interesting and nuanced stuff. The trade-off there is that sometimes it's harder for us to understand exactly what the model's picking up on, why it's found nuance. So we sort of lose control over the system because we have to in order, you know, we're not using our mind to try to sort of bake simplicity into the process. We're sort of allowing there to be more complexity.

So how does this work? I take millions of sentences of any language or one language?

You normally want to stick with one as opposed to just sort of random.

Does a certain language work better than another?

Depends on how many training examples you have. So what we're doing is you take a sentence and then basically the first pass is to represent it as a series of numbers in what we call vector, and the numbers are sort of projected into space. So you turn it into like a geometrical arrow, if you will, okay? And then this is what I found. This is why I got into the field in the first place. It was so fascinating to me that when you turn language, sort of process it sequentially and turn it into this number, you can then do mathematical operations on it. Like you can add, subtract, multiply, all the things you do when you've got like, you know, a line in space. And then you, so you do all that processing in the world of math and then you output again, sort of a language, like a sentence representation. And it's as if it's been doing sort of language the way we think about it, right? It's as if it's been finding grammatical relationships or semantic relationships, but all that it's doing is it's using math as a proxy for meaning. And miraculously it works. And it doesn't work perfectly and it makes all sorts of mistakes, but like we can use math as this approximation space for meaning. Which is, like, I find mind-blowing. I mean, really so exciting.

What does that mean?

It means that there's something going on in how our minds— if we were able to, like, cut open our brains and look in and sort of observe the processes of how neurotransmitters are flitting around from place to place and all these little connections that are made between synapses to make meaning like we're doing right now, having this conversation, we'd see something that was a much more complex, fancy version than these artificial intelligent neural networks are doing, right? So I don't want to give the impression that it's like there's a direct analog between how brains work and how these machines work, because they're not. It's more like a metaphor than it is like an exact replica. But it does mean that there's some sort of mathematical foundation for the way in which we communicate with one another. You know, if we, again, if we could peer into our brains with some special magnifying glass, we'd be able to see.

Wow. And I would wonder like, what is meaning then?

Yeah, for sure. Well, does it matter? Does it matter as much as we think? I mean, in terms of linguistic meaning.

Does it or not?

I mean, I take that from a perspective of it certainly has transactional value. So, So, you know, we're here today because we were able to exchange an email and agree upon a time and a place. And that time and place had a referent. So indeed, you know, 65 Langton Street is a place that I can look up on a map and I can come there. And when you asked to do a podcast, that had meaning for me. Like, it wasn't, you know, and it was different than coming over for pizza with somebody I've never met before. So, you know, that's something different.

Meaning has so many variables.

It has so many variables. And so it has a— it's a token of exchange for us to be able to— like, communication is just a form of being able to do stuff together, as opposed to sort of just existing in our own little world where we're unable to actually find a space for commonality.

Does Miss Sun— oh, go ahead.

But I was going to say, but that's a really different sense of meaning than the emotional meaning of falling in love. Of watching a movie that makes you cry, and then you say like, that's, you know what I mean? So something that is meaningful versus like something that has meaning, right? And has sort of pragmatic transactional value. Machine learning.

We told you she had a lot to offer. Catherine has her own podcast called In Context. We hope you'll check it out.

Machine learning.

Coming up in this final section of today's Machine Yearning, Shane and Catherine dig into feedback loops and how they help brands finally build real genuine relationships with a community of customers. This might be the most vital part of a conversation that gave us a ton to consider and apply. So let's get to it.

Take capitalism and business and how it impacts communicating with your customer. Like we're communicating as two people who have found meaning. How do you find meaning with your customers?

It's a great question. I actually do believe that sort of the big shift that's rendered possible with AI is that customers now give feedback and can give feedback and can be listened to even by their actions and transactions in a way that really wasn't as focused on in technology systems that just sort of push stuff out. What's neat about new products, like I love Spotify, is that they have this very simple interface where we can say plus, thumbs up, thumbs down, and shape the set of choices that will be presented, like the experience that we have in the future. So that's fascinating.

That's a great example of AI being used very well for good.

Oh, for sure.

And my weekly playlist on Spotify is amazing.

Somebody asked me recently in terms of how AI is going to impact sort of the human experience. They were like, is it going to take away self-determination? I was like, oh, actually they also said A, self-determination and B, serendipity. So like, you know, the experience, sort of a nostalgic experience where you walk into a bookshop and you think you're going in to buy Oscar Wilde's Portrait of Dorian Gray and you sort of browse around and find something that you never would have discovered otherwise. And that's great. Like, I'm all for that. But I think Spotify does that. Because of the Discover Weekly, where we don't have a lot of time to go on and search for new music unless you're really a diehard and you're reading Pitchfork every Monday or every Tuesday.

So if I'm a brand in the world, how should I think about language? Or how should I change my language? Or what do I need to do?

I mean, I think it's about that feedback loop. So you need to open up the ability to listen to customers as individuals. And here it's funny I say that, but my mind immediately goes, this doesn't mean that it's become creepy and collect every single solitary piece of data about somebody in every aspect of their life so that you know them better than themselves and can do creepy stuff with data. But rather understand that you might be underserving certain populations if you go in with your expectation that everybody works this way or these ossified, approximative stereotypes on what people might be interested in. If you can get some vehicle, it has to be easy because no one wants to fill out a survey. Nobody wants to do extra work. They just want to be delighted. And also what I love about design thinking and applying this with AI is that when you think about offering a product, don't view it as— this is the product, but rather, what does it mean to experience this product? So Warby Parker is actually really great at this, where they don't sell glasses. They sell the end-to-end experience of what it means to fix your vision and include in the customer experience everywhere from being able to try on a couple of pairs to figure out what works best to— I think they have some— I actually don't— I'm not a Warby Parker customer myself because my eyes are very close together and standard eyeglasses look horrible on me.

Interesting.

So I have to go to kids. Glass at Shops, which like sucks, but it is what it is. But they just contextualize in the full experience. So I think that there's really this magical opportunity with AI. We tend to think of it as like it's going to— there's fear of job automation and everything becoming the realm of the robots. But if we reconceptualize the tech as what I consider to be the connective tissue between humans who want to love one another as individuals or what it means for a company to really love a customer. And loving them means empathizing with their experience and different people's experiences. I think it's a wonderful way to view these tools.

And brands are driven by attention.

For sure.

Right?

As capitalism is. So when they can deploy AI and if attention is the game, and it's focused on that, is it going to get bad?

I mean, for me, this goes to long-term versus short-term. So if the brand is optimized to get short-term revenue wins and attention and hacking attention is the hook there, then they're going to optimize for that. If, on the other hand, there starts to be sort of a— consumers don't want that anymore and the brands really need to focus on their long-term like loyalty and relationships with people, then for me, somehow it has to make sense. And of course, I don't want to be overly idealistic. There needs to be the right balance between P&L holders and businesses that have near quarterly revenue targets and fiduciary duty back to shareholders. And this is how the system is all structured. But we're starting to see things fray a little bit at the seams.

What are some examples?

So gosh, even in the company that I'm working with right now, we— There's a notion in sort of thinking about working with data and consumer enterprises of lifetime value. So like, you want— and you want to maximize the lifetime value from a customer. So it might be that somebody comes in and they spend like $700 on day 1. And you're like, ooh, interesting. Like, let's focus on this person. But then you notice that they spend once and then they're just gone. So like, their lifetime value is like— is exhausted from that first transaction. And others where it starts off kind of slow, but it builds over time. And you got that like— It's almost like a one-night stand versus a long-term enduring relationship. Amazon. Yeah, for sure. Amazon Prime actually is— it's a brilliant mechanism to ensure customer loyalty. And what I've heard— and this is a little bit of just— it might be apocryphal. But all of the features, like Amazon Video, that you get as a Prime subscriber, those are tricks to get people on Prime. It's not a service in itself. It's an additional benefit. Because what they really care about is having that subscription, because that leads to— the correlation between that being a leading indicator of stickiness with the customer is so tight that they're just like—

The best business is subscription.

For sure. And once you've already invested that sunk cost and you get all of the benefits of it's going to arrive in 2 days, the amenities that go along with being a Prime subscriber, of course. Now, does that mean that it's really in the customer's best interest? Good question. It certainly is the case that as you think about why Amazon is the business that it is, that they invest in long term. Like they're thinking about when they may take an action today, it's not oriented towards short-term profitability and horizons. And this is a lot of the sort of companies that, like large Silicon Valley companies where if you look at their balance sheet, it's so much startling that the valuations are so high, but they're sort of playing a different game than most other companies are governed by.

If I'm a brand, and even where we in our past got taught is to have the answer and like raise our hand and say, and get an A and not really be cool with saying, I don't know. In this space, how is a brand who's probably not cool with a bunch of airs cool with airs?

They struggle with it.

That's why people struggle with how to even understand to implement AI or bots or whatever.

From my take, it's the biggest cultural challenge is getting comfortable with making mistakes, learning, having the output of a system be a probabilistic guess. That may or may not be right and being OK with that.

Especially because software actually was built on the internet in a way where it worked or it didn't.

Yeah, for sure. People are used to also even in agile product development being able to know how long it takes to build x or y widget. When you move into AI and you're like, we're going to do some experiments. We start with this hypothesis. We get some data. It's like, nope, it didn't work. There's basically re— it's almost being more like an investor where you sort of place your— you spread your bets across multiple companies knowing that 9 out of 10 are going to fail, but there could be the opportunity for one to be really great.

It's focusing on long-term opportunity, knowing that errors and failure is a big part of the new world.

It's part of it, and it's the only way to learn. Also, even as humans, you learn a lot more from the mistakes you make. And it's helpful when the end of the story is positive, right? If you're just sort of always failing, it's not really the best recipe for well-being, mental health and well-being. But if you don't make mistakes, you learn the most when you try something and it's not as great. You can refine over time.

What are the 2 or 3 words that you think, if I was a company, long-term I should focus on to get me started today to walk into this space?

So I think a lot of it is going to depend upon data. So imagining what could be possible and then ensuring that you're designing interfaces to collect really great and interesting data to help solve that problem.

What do you see as the best interface?

So I just did a podcast interview on my podcast with my friend Eric Colson, who is the chief algorithms officer, which is such an interesting title, for Stitch Fix. And so Stitch Fix is this— are you familiar with it? It's like this—

Explain it, though.

Yeah, it's an outsourced personal shopper company where it's an e-commerce play. And you can come in as a user, and you just sort of input information into 60 questions. And then that's it. They send you clothing, and you can get 5 items per month, 5 items per week, however many you'd like. And what's really fascinating about what they've done is they've actually— they get a lot more data about people. So they've designed this interface where it's not just your transactions, like what you bought, but they've got measurements, size, taste, right? So it's a much richer collection that does require that upfront effort from a person to come in and fill out the initial information. But that effort is then offset by the fact that going forward, they don't have to shop anymore. You spend 15 minutes to get rid of the shopping entirely. And in the process, they're getting a totally different rich set of information about people and their preferences and what they're interested in that a typical e-commerce play doesn't have.

Fascinating. So because of the relationship channel that's one-to-one, it's not a bad idea to make me work up front. As long as you deliver on the promise.

It's an interesting trade-off. Because anytime there's friction and there's a little bit of work up front, it's like, you're going to lose some customers. Because they're going to get to that interface and be like, screw this, I'm not doing it.

But then your value is in the relationship because that's the only thing that keeps people coming back.

Yeah, for sure. Yeah.

Interesting. It's even thinking about a hotel. I hate when I have to rebook again and re-sign it. Everything you have to do again. It's almost you got to get me to never have to do anything.

Oh, for sure. Once you've like— I mean, they know you well enough to know a little bit about your preferences. And if you sort of set this thing and each time you stay, it should become easier to stay the next time.

That's a good motto. Every time the customer interacts with you, it should be easier and more delightful.

That's right. Yeah. I think that's a good way to think about it. And then that flip side of getting there is knowing that there's going to be some error. It's not going to be perfect every time, but in the aggregate, that overall, every time the customer interacts with you, it's easier and more delightful.

How do you interact with the customer if it is an error? Do I just tell them, hey, is it more like what we're trying to do as humans, like more direct, honest feedback? We're like, we're really sorry, we just don't know that answer yet, but we'll let you know when we do.

I think that's the way. And then you say, well, how do we get there as a culture, right? Where, what do we have to do in terms of educating everybody about how this technology works so that we can have more honest, conversations via AI interfaces to be like, we made a mistake. We're really sorry. Here's why. What can we do differently next time?

We need to reprogram ourselves to be cool with making mistakes.

And that goes not only at the company level, but even at a— think about when— I mean, I'm in sales meetings all the time. And I've always found that it's a great way to gain the prospect or the customer's trust by saying, I don't know the answer to that. I'm gonna go home and do my homework. And sort of, it almost legitimizes then everything else you've said, 'cause then it's like, everything you've heard here to date, like, it's right. I wasn't, I'm not faking. Going back to the beginning of our conversation, like, you know, how do you know if this is sarcastic or real? It's like, one way is to express vulnerability and be upfront with areas where, you know, you can't, you don't have the right answer.

It's the key to trust.

Yeah, and I think trust is a huge, I mean, it's a huge thing right now between consumers and brands.

Thank you very much.

Thanks for having me. It's fun. M-A-C-H-I-A-E-Y-E-S-I-X-Y-A-G. All right.

Thanks for listening. Get in touch on Twitter @Assist. DMs are open. We're super interested to hear who you think should appear on the podcast. Please subscribe so you never miss an episode and share this with someone who cares about how we make sense of these changing times.

Machine learning.

Machine Learning by Assist is made by Paul Ciuffo and Michael Elsesser for Limina House. Have a great day.