Is It Time to Stop Optimizing? (Coco Krumme)

Listen now (39 mins) | "And the larger issue is that we start to see the world as something to be optimized. So even if we were able to optimize unquote for the right things..."

You can also listen to this episode on Spotify, or wherever you get your podcasts.

Coco Krumme is an applied mathematician and writer. She got her doctorate at MIT and has spent some time working on data modeling in Silicon Valley, as well as some time living a bit more off the grid, on a remote island in the Pacific Northwest. Through her work, Coco came to question how, and why, we’ve become so obsessed with optimizing nearly every system in our country, and every aspect of our lives. She wrote a book that explores the myriad ways we’ve optimized too many things; what we’ve lost by allowing optimization to become our denomination worldview; and what might be a more compelling and ultimately life-affirming perspective, or priority. That book is aptly titled, Optimal Illusions: The False Promise of Optimization.

Today, we talk about the kind of optimization we’re seeing play out in our culture now. We recorded this conversation in late March, so there was much to discuss on this front.

I also asked Coco about why she thinks the threat of surveillance capitalism is overblown. She shares her thoughts on AI, the hype and money bubble around it, and why she’s more optimistic about what humans can do than what these technologies can do.

MORE FROM COCO KRUMME:

Optimal Illusions: The False Promise of Optimization

Coco’s Website

EPISODE TRANSCRIPTION:

ELISE:

I don't remember who said this, but I think about it all the time that the universe or the cosmos can be explained through music, mythology and math. Let's talk about math and have you abandoned math or will you always be a data scientist at heart?

COCO:

Those two questions need to be answered separately. I don't think I've ever been a data scientist at heart. I think that I'm a really good faker because I like numbers. I'm good at numbers, but I think I've always been more of a creative type, which actually I think mathematics gets a reputation as a more scientific discipline when I think it's more of an art, especially when you're thinking about how to represent the world, which is what a lot of mathematical modeling is about. What's your framework for understanding the world? What's your metaphor? And that to me involves a huge amount of creativity. It's not cut and dried. I think when you tell people you're an applied mathematician, they often think, oh, you're an accountant or you're a scientist, right? That's a long way of saying I don't think I'm a data scientist at heart. I'm not good at details per se. My mind meanders, but to answer the other part of your question, I haven't given it up entirely. I still teach in a couple of university programs and data science. I still do some consulting work and really enjoy that, and I think the mathematical way of seeing has certainly as I described in the book, influenced a lot of how I interpret the world, and I think it represents how a lot of us interpret the world right now.

ELISE:

Yeah. Well, I think too, you write about this balance between parts and whole and it feels like math is, and I used to be a total math nerd until I just stopped when I got to college. It could have been my fate too cocoa. I'd been on this accelerated math course and then never took another math class, which is probably not a totally uncommon story. And what's so interesting about math is you are either building from the bottom up or you are revealing what's already there, explaining phenomenon in the world through this symbolic system or representative system, and we'll get into the contours of the book, but it feels like we really get into trouble when we start in this highly reductive atomized way or that we're the creators of the world through math rather than using math to describe the whole. Does that make sense?

COCO:

Yeah, yeah, for sure. I think there's a few things in there. One is in my mind remembering that it's, but one way of representing the world that's important and there are other ways of understanding, and then the other is this distinction between the kind of bottom up or top down and seeing mathematics as almost as this godlike force that we control versus a toolkit that we can use in limited ways for certain things, but it's like a children's play kitchen. We can cook up some fake cupcakes that we can't control the material world fully with this toolkit. At least that's my opinion.

ELISE:

Yeah. Well, you write about Newton too, who I think is obviously for most of us, he represents the end of one system. What he has described does a really good job of describing a lot of phenomena in our world, and then it breaks at some extent, right? At a quantum level, it doesn't work, but you describe him as sort of the last of the magicians or actually coming out of a more alchemical tradition, which I didn't know,

COCO:

And that's not my phrasing actually, the last magicians was, I think Keynes actually described Newton in a speech that way,

ELISE:

And I think it was Keynes too who talked about Newton as someone who had a sense or perception of something and then would go and find the proof rather than building a model of the world based on proof, which is a subtle but important distinction, I think.

COCO:

Yeah, yeah, and from what I understand of Newton's biography was frustrated by he was searching for these ideas of alchemy, and he was very much a scholar in that sense, driven by his intuition, but kind of unearthing and mining rather than proactively or in this very measured way going out and saying, I have a hypothesis, and now we're going to go through this process of collecting data and testing it in a rigorous way. I find that tidbit fascinating that he is come to be known as the father of modern science when he was actually operating in this older tradition.

ELISE:

Well, in so much of your book and this fever that we find ourselves in particularly now, even more so than when you wrote this book, the Fever of Efficiency, the fever of optimization comes from this bottoms up. How do I take what's present and then just sort of coerce it to a point, and I want to talk about all the fallibility that then creates for us, but I guess you could have seen this coming. We are on this collective H, there's this fever of efficiency, DOGE, and

COCO:

Yeah, no, for sure. I mean, one might also ask the question of how efficient is firing summarily firing?

ELISE:

Yeah, exactly.

COCO:

Large swamps of the federal government. That's maybe a separate question of whether this project is in fact efficient by its own definition of efficiency, but there's no doubt in my mind that it's operating with the idea that efficiency is the highest and best goal, right?

ELISE:

Well, and I think it sets it up as this problem that should we reach peak efficiency, whatever that means, that somehow everything will invariably work better and we'll have achieved some sort of newer standard of governance where there's no waste and all of these things that sound, sure, they sound compelling to everyone. Nobody is like, yes, let's pay for a bunch of things that are ineffective, and sure, I think we can all understand sort of the baseline reason, but then when you're, again, you're not holding the whole, as you're looking for the parts, you start to see how ridiculous the whole construct is. I mean, maybe it's not a good example actually of

COCO:

No, I mean, think it's interesting. I think there's certainly something absurdist too about this moment. The part of me that likes to laugh and see the absurdity in the world is certainly appreciating this moment. I don't know a huge amount about this, but I know that there have been other, this isn't the first movement by a president to go in and kind of clean up or make the federal government operations more efficient. There have been other presidents who have done that. I think what is maybe different about this time around beyond just the various absurdities of how it's being conducted is that rather than the measured surgeons approach where we're going to go in and surgically remove those parts that aren't working or consolidate this, maybe surgeon is the wrong word, but a craftsman's approach of we're going to make things more efficient by going in and shaping and molding carefully right now it's this megalomaniacal.

We're going to make it more efficient by knocking things down and then rebuilding it in the model of Silicon Valley or these sort of abstracted notions of efficiency, which as we all know that that world, even if Silicon Valley did operate according to these very utopian abstracted models of efficiency, which it doesn't, that model is very different from how a bureaucracy like a federal government works with redundancies and so on and so forth. So I think it's completely ignoring or rendering itself oblivious to the fact that in a lot of human governance structures, you need those redundancies. Efficiency is not, or total efficiency is not the right operating principle,

ELISE:

Right? Yeah. It's not irrelevant, but it feels somewhat not necessarily the point when you're talking about the impact of some of these agencies.

I think it goes to, and this is a point throughout too, of what are you optimizing for, and every single part or function of our government theoretically is optimizing for something very different, whether you're talking about U-S-A-I-D or you're talking about the Department of Education, and I'm not a government or policy wonk, but it feels like one that this idea doesn't necessarily apply in the same way to all of these.

COCO:

My question would be is there a purpose to be optimizing for? That's a very new phrasing, and that's part of the argument of my book is our language now is about optimizing for something, but a lot of the stuff that we do in our life and that our communities do, they're not, is the purpose of the Department of Education to optimize for student outcomes? I don't know, maybe. Or is the purpose of our utility companies to optimize for this or that maybe the purpose is to keep the lights on, or maybe the purpose of the defense department should be just to keep us safe. It's not optimizing for anything in particular,

ELISE:

And you write extensively about farming and whatnot in the book, but yeah, are you optimizing for a specific year's crop yield without necessarily understanding the impact of monocropping or whatever it is that we're doing or GMOs and the ancillary pesticide use that comes with that? What are we optimizing for? And again, how do you hold that part in the context of the whole, which we have just not been very good at, or we become much more parts focused to our own detriment potentially.

COCO:

Totally, totally. Yeah. I think that is the danger of treating things like an optimization function or it's one of the dangers is that we get very stuck in optimizing for one thing, the one thing that we've identified as our objective function and we forget about, it's a complex system that's often behind what an agriculture is a great example because it's when you're optimizing for something as specific as yield or revenues, you're losing track of other things that matter, whether it's the health of the community, the health of the soil, so on and so forth. So that's in my mind one issue with optimization and the larger issue is that we start to see the world as something to be optimized. So even if we were able to optimize unquote for the right things, even if agriculture said, okay, we're going to optimize for all these things that matter, we're still servants to this idea of efficiency.

ELISE:

And it's interesting you write a little bit about this idea of surveillance capital and how disturbing the concept is, and yet how far we are from actual, obviously we're concerned about the rise of the machines and the death of humans, whether it's in an AI context or an optimization run amok or this idea that our lives are being data mined and then given back to us in nefarious ways, some ways, of course, that are appealing, like, oh, these are all things that appeal to me, but that drive capitalism, et cetera. But you also write about how so much of this is, and it feels this way, we're at the most rudimentary when we're talking about optimization. It's like we're optimizing for one thing in this quest for this parts driven goal without optimizing again for the whole, and that we're sort of in our infancy and we're not actually that good at it. Can you talk a little bit more about why you think so much of the threat of something like surveillance capitalism is overblown?

COCO:

I'd love to talk about this. I think, and I'm very much a minority in this, at least in the circles I run in, but I feel like the common refrains that I hear at least and that one hears in the news media is either this boomer view of AI and it's going to save us. It's going to liberate us to spend our time in leisure. It's going to do all the work for us. It's beautiful, it's moving fast. It's going to get us to Jupiter and cure cancer and et cetera, et cetera, et cetera, or the domer view, which is it's going to annihilate us, it's going to tear us apart. It's going to be used for all of these nefarious purposes. Nuclear weapons will be launched by AI bots, et cetera. And the DOR view too is therefore AI needs to be controlled by us, the domer experts, the AI safety people, et cetera.

And I'm kind of like, meh, neither boom nor doom. People I know who work in this field of AI have pushed back on me and I've taken a little more time to explore some of these models, and there's some pretty cool stuff that's being done technically. I mean, I actually don't think the results are that cool yet from a product perspective. It's like, okay, we have basically a more advanced clippy that's probably going to eliminate some analyst type jobs that can synthesize information pretty quickly. It's also uses a huge amount of material resources that's a part of this that not many people are talking about, huge amounts of energy just to run these models and in my mind, pretty underwhelming results. We can synthesize information more quickly, we can make some cool graphics. We can write stories in the style of whoever Shakespeare, Hemingway, but I think number one, we're in a huge hype and money bubble around these technologies. And number two, I'm optimistic about humans honestly, because I think that once we realize that all this hype is hype, it will be freeing in a sense when the narrative shifts back to actually there's a lot these technologies can't do and that humans uniquely can do, which I think is very cool.

ELISE:

Yeah, no, I mean I should probably spend more time with ai, but when I've queried it for research purposes or when I've seen what it can do in writing an intro, I'm like, no, thanks. This is not at the this not very good. And I think we're all sick of AI art, so I respect that perspective, and I think as you age, you live through so many of these bubbles where it's everything and nothing simultaneously. And I agree, humans are amazing and maybe it moves us past this idea of needing to memorize rot facts, and we're clearly not the most adept calculators, but I think we're probably much better creative problem solvers and maybe have a bit more wisdom than knowledge.

COCO:

Oh, for sure. And we're rooted in the physical and biological world. One of the things that fuels optimization is abstraction or it's built on this notion of abstracting things into models. And it's depressing to me that this view has become so entrenched because it preferences the kinds of people that work well with abstractions. I think that's part of what we're seeing with the current phenomenon of Doge and so on and so forth. It's this of a kind of culture that works well with abstractions rather than kind of culture that's rooted in the physical material biological world. And I think as with everything right there are ebbs and flows, the pendulum will swing back, but I think that is a lot of what makes us human and that what machines can't do is that we're rooted in the physical world. We experience pain, we experience joy, we eat, we poop, we sleep, and we draw on that physicality to make art or to write books or to compose music, and that's something that computer doing calculations will never be able to draw on. It's this infinite amount of life, whereas computers only have just the accumulation of information bites to recombine in a limited number of ways

ELISE:

And that they're working with sort of what's already present rather than seeing what's to come or imagining what's to come.

COCO:

Yeah.

ELISE:

I want to ask you too about this idea, and I think this is something that we can all sense or that is the heartbreak of efficiency or running those tight ships, which is that you create all this fragility when there's no slack and the need for slack, just this, and obviously people are calling out for this and pushing against hustle culture and this always on productivity, and it's one of the big drumbeats of culture in that insistence. Can you talk a little bit more about how that ends? Because you can see it in your mind's eye. Of course, it breaks and you write about in the context of the decline of civilizations optimization.

COCO:

Yeah, I'm certainly not the first or the only to point to this, and there have been entire books about fragility of as you optimize and you see it. I have a few examples in the book about container ships crashing and shutting down supply chains or the electrical grid when it's overly optimized and we're overly reliant on it. You see it in biological organisms and the species that have optimized for their niche, their environmental niche are also the most susceptible to big shocks. So if you're living in an icy tundra and you optimized your food supplies and your way of building houses and et cetera for that particular environment, and suddenly the environment heats up or vice versa, you're living in a hot environment and suddenly you get a cold shock that can destroy a species, a culture or whoever we're talking about, whereas the more adaptive types of species or cultures that can adapt across environmental niches are less susceptible to shocks.

I think in the book, I'm more focused on the metaphor of optimization and the worldview of optimization, where that came from rather than the results of our overly optimized world, which I think both of those things are problematic, both that we've optimized too many things and that it's our sole worldview. I think we are seeing right now a lot of individual reactions to an overly optimized world. I was just reading about, I think it's called the Luddite Club. You are rejecting smartphones and Send and so forth. There are lots of examples like that, people moving off grid, but I don't really see any large scale movements right now to summarily reject optimization just because it's embedded itself or it's made things so convenient, right? Talking about specifically technological optimizations and we're all enjoying the fruits of all of these things that it's hard both on an individual basis and on a collective basis to say, we're going to reject this. And I think that's actually the nefarious brilliance of some of these companies. And you mentioned surveillance capitalism earlier, right? Everything. It's just becomes such an easy, mindless trade of little bits of our freedom or our data for convenience and luxury and material necessities that it's a very hard thing to say, Hey, we're going to boycott this or we're going to turn against it collectively. I don't know if you have thoughts on

ELISE:

No, I think it'll be interesting to see implications of the beginning of this year and as theoretically as global trade is interrupted as we start to not be able to get our avocados for 99 cents or whatever it may be, whether that actually resonates or I think we're getting a big civics lesson right now in terms of what is local government and state government and federal government, where do these things that I rely on come from and where does all of my waste go and who's taking care of this? I think we've taken a lot of these systems for granted, and so I don't know if there will be a forcing function. I also think that, and I feel this in my own life, and I don't know if everyone relates to this, and I think it's been going on for so long now, and people keep asking and we're all giving the same answer.

How are you? And you're like, well, objectively, I'm really good subjectively, everything's terrible. And that dissonance that I think is happening for a lot of people where they're experiencing a digital world that feels so incredibly chaotic and out of control, and then a physical material biological world right now where you're like, everything seems okay. I feel okay. I don't know. I don't know what will happen. It feels like we're at a fever breaking point and I'm probably more spiritually inclined than you are, but it feels like there's a fracturing or a fever breaking that will be quite healthy ultimately for us to sort of recalibrate against and maybe choose something differently once what we were choosing is no longer a given. But I don't know. I'm not sure what will happen either. And I recognize, I try and unhook, I recognize what institutions are important to me personally, so I don't shop at the big chain grocery stores. I don't shop at Whole Foods. I go to my more expensive but local grocery store, and I order my books not through Amazon, but through my favorite local bookstore diesel because absent those two institutions, my world would be much less vibrant and interesting. So it's stuff like that, but I agree with you. It's hard to then apply that across a massive swath of people, and we haven't seen insistent and sustained pushback against everything coming from one store overnight. Yeah,

COCO:

No, and I think you and I have this luxury, this ability to be able to make those choices, and I think everybody ultimately has the will to be able to make choices in their life, but it's much more difficult if you don't have that option of the local grocery store. You're without making huge sacrifices financially or time-wise. And I think that's why this creep of these optimized companies and systems and so on has been so insidious is that sort of slowly and in the light of day, they have for many people made that choice much more difficult to continue to do the thing that they believe is more sustainable for their community in the long run. I agree with you too. That to me it feels, my intuition is that we're at some kind of breaking point, and maybe it's a lack of imagination, but I don't know what it looks like, and I don't know if it requires kind of a breaking open or breaking apart for the tide to shift, so to speak. Or if suddenly this growth of these kinds of institutions that we're talking about both on the corporate side and the governance side, whether they will just run out of fuel basically, and suddenly that kind of collective action and individual action will bear more fruit.

ELISE:

No, it feels like we're on a course and we've to run to the end of the course, and then I think about it with a company like Amazon, and again, I'm over my skis. This is not my area of expertise, but where they got us all hooked on it by artificially deflating prices and creating this convenience culture, which is very compelling, and I'm not going to lie. I would never lie and say, don't order anything from Amazon. I do. And yeah, then when I actually go to order books, I'm like, oh, this is an expensive book, and I go to Amazon. It's the equivalent out here.

I think that there's less and less of that, and I think as we're experiencing more and more inflation, I think it's also maybe causing people to say, it's happened for me at least, where I'm like, but why do I think that eggs shouldn't be more expensive? When you think about what you're actually talking about, it's like, I don't know. It's forcing me to examine the way that I think about why should beautiful produce be so cheap, and yes, we need to expand access and there are food deserts, so I don't know if that then pushes people to say, I'm going to put a vegetable box outside of my door.

COCO:

Yeah, I mean it certainly has in the past. I think that's something really interesting. It's funny to be talking about the book now because when I wrote it, when I did a lot of the research and writing, which is like 20 20, 20 21, it's been continually astounding to me until very recently, how cheap our food is in this country

And still arguably compared to a lot of places, even with all of the recent inflation in the last three or four years, our food is so cheap, and we have such incredible abundance and variety of things that we can access for not very much relative to the cost of a lot of other things. And we associate cost with how much we value something for better or worse. And I do think if there is that recalibration that you're talking about, maybe we will start to not only shift our consumption and plant those garden boxes and so on and so forth, but also value food is such an important part. It fuels the cells in our body, and maybe we should be spending more money on it and less on other stuff.

ELISE:

Yeah, no, and I see it in my house with my kids too, where yeah, it's not highly valued and we waste too much food, and I feel like COVID was a little bit of a check on people where they were like, I need to just eat everything in my refrigerator and I need to savor these leftovers and so on and so forth. Yeah, we treat it as so incredibly cheap and largely disposable in part because we have access to so many calories in this country with ultra processed food, et cetera too. But it feels like that whole system needs a recalibration in part because I think we want our farmers to have more than a subsistence wage. I think about that whenever I buy, even if it's like an organic rotisserie chicken, I'm like, how is this possible that this is, yeah, 8 99 or whatever, where's the profit in this for anyone? How is this chicken's life worth so little?

COCO:

I mean, that's where optimizing, right? The only reason that so much of that, even when it is labeled organic, unless you're buying it from the farmer next door or a legitimately small farm, the reason an organic registry chicken can be so cheap is because the farms in this country have been consolidating, consolidating, consolidating. Without those economies of scale, we wouldn't be able to have the kind of prices that we're seeing for food.

ELISE:

As a progressive liberal person, you almost can't even contemplate it because you crash right into, I sound so economically privileged and not everyone has access to food. And we're talking about Montana. There are food deserts obviously, and many of these parts of the country where there's very little fresh produce and there's very little access to whole nutritious foods, and yet we're so glutted in other parts of the country and we don't really value it. We really don't. So Coco's book is Optimal Illusions. So this, I just want to read a part to you from the book, which I think is a really good summary of what we were trying to get at or pick at today. “Indeed, it's the same atomization that underlies much of computation in general held together by the glue of abstract expressions and formula. When we start to see the world in terms of self same units of zeros and ones of call center metrics of electrons that can be peeled away and pasted onto new substrates of a genetic code that can be spliced at will, we start to see a world that's ours to build across various disciplines.

This atomization of our material reality created a division between part and whole and cast to the side. Those things that weren't readily divisible. The result was a reductionist and blinders on science, economy, and culture. This chopped up view is a far cry from Democrats whose project in seeing the part was a distilled expression of the whole.” If you like today's episode, there are several ways to support the show. I produce it myself, so this helps me to continue to make it. First, please rate and review the show on the platform where you listen and consider sharing this episode with a friend. That's how it grows. It is so helpful. Second, please support my sponsors who make this show possible. And if you are interested in sponsoring the show, you can email me at admin [@] eliseloehnen [dot] com.