HEFFNER: I am Alexander Heffner, your host on The Open Mind.

I'm delighted to welcome our guest today, Coco Krumme.

She's the author of the new book, Optimal Illusions: The False Promise of Optimization.

Welcome, Coco.

KRUMME: Glad to be here.

HEFFNER: Coco, what inspired you to write this book?

KRUMME: A few things.

I have worked for many years as an applied mathematician in the field of optimization, and I began as a total romantic about mathematics and its ability to capture the world and a productive lens through which to understand it.

And the more I spent in these sort of centers of optimization, like MIT and Silicon Valley, the more I started to see the darker sides of that romanticism and some of the trade-offs we've, whether consciously or unconsciously, admitted into our lives as a result of optimizing everything.

So both on a professional and a personal level, I started to lose some of that romanticism, and the book came out of that.

HEFFNER: When we think of the word optimize, we think about increasing efficiency.

Is that a correct thesis or understanding of what that concept is, or no?

KRUMME: Yeah, I think it's actually a pretty sophisticated understanding.

I think some people, when they hear optimize, they simply think of the best, right?

When we see on a vitamin bottle, right, optimal nutrition, it's like that must be the best.

But I think, yeah, the sort of more technical definition is around increasing efficiency or maximizing or making something better.

HEFFNER: When you alluded to the dark side, did you mean it in the manipulation of data, or do you mean it in terms of a thesis that we've explored explicitly on The Open Mind for the last many months and years, which is the idea that efficiency in technology is actually tied to or correlated with deficient outcomes for people's livelihood.

So efficiency in technology is the net sum of that is creating deficiencies in our actual lives.

KRUMME: I think I know how you would answer the question.

But yeah.

Isn't that the dark side?

HEFFNER: Right, right.

For sure.

And expound on that because there's, again, there's the dark side of kind of manipulating data, and science of mathematical manipulation and how people are abusive with data.

But then there's the dark side of the outcomes of the optimization.

So I'm wondering.

KRUMME: I see the distinction now.

I think it's both those things, and what I talk about a lot in the book is, there's sort of an additional dark nuance to optimization is that as a way of understanding the world it's encroached and taken over.

So there's the process that might be dark.

There's the outcomes that might be dark as you talk about.

But there's also this belief that I think is pretty dark that optimization is the way we should be understanding the world, that we should always be seeking to make things more productive or more efficient, even if the results aren't dark right.

The, the results of a lot of optimizations are actually pretty bright and good and helpful, right?

But should we be despite that optimizing as much as we are, I don't know.

HEFFNER: You write in the introduction, optimization is a way of seeing the world that has overtaken most others in both modern America and much of the West.

In fact, it has the peculiarity and advantageous peculiarity in terms of its own ability to persist of rapidly crowding out others.

So let's just take optimization from your kind of day job in mathematics in a practical sense of what you mean by that, and then some more real life examples.

But in your own career as a mathematician, where were you pushed to optimize, and was the result kind of exploitation of people in some way or another?

KRUMME: I certainly don't think so and would hope not.

I mean, I have been careful about taking on work that, you know, I believe has some value and is in line with my own code of ethics.

I think there is always the, the potential of technologies to be misused or misapplied.

I think it's important to kind of restrict the domain, you know, when you're working in a technical field, when once you've decided, whether at the individual level or at a sort of corporate level, that an optimization is the right one, or that we should be engineering a system or optimizing a system, right?

The, the actual act of optimizing is pretty value neutral, right?

I think we can get into, I think there are these sort of emergent concerns out of AI ethics and so on and so forth.

That's a whole separate discussion, but the algorithms themselves, I guess is what I'm saying, are value neutral, right?

HEFFNER: It's interesting you mention AI because in a sense, with chat GPT, in the world of IA.

It's such an open source platform now that it isn't constrained by a lot of the human considerations.

Like if you're making a product that has been subject to what we were calling in the last few years, shrinkflation, like the material concerns that require companies, whether it's airliners or sodamakers to optimize, like with AI, I guess one of the benefits associated with it will be that it, it's kind of immune to the compulsory sort of the forced instinct of a lot of people to optimize and create greater efficiency.

You what I mean?

KRUMME: Not sure I know what you mean.

HEFFNER: I think like AI can do basically any task at any moment, um, not replicating how a human would do it with their sense of humanity, but, at least when it comes to like processing assignments that a human might give a computer, there, there are very few constraints, right?

When I think of optimization, I think of it as a way now that people are trying to, they may call it optimization, but in reality, what they're doing is trying to create more efficiency and a better bottom line in their company's output.

I'm just responding to your mentioning, because I'm thinking of it as a way that doesn't require us to have the same tunnel vision towards optimizing and creating efficiency.

KRUMME: So I guess I'm not sure what you mean by AI then.

Like, you mean GPT is, we sort of hit the button and it's running semi autonomously?

HEFFNER: Right?

And right.

And if you eventually, if you transfer that idea to an easier way to, to make products that are going to suit people.

Like right now, would you agree that the basic equation for optimization from, from the perspective of most companies in these last few years has been, they may be calling it optimization, but it's really about cost cutting.

KRUMME: Well, cost cutting is just one form of optimization, right?

I'm not sure what they're calling it.

These companies tend to call it whatever's going to resonate best with the public.

But cost cutting is optimizing for having more revenues and with less expenses or costs, right?

It's a form of optimization.

And there are other things they could certainly be optimizing for, like revenues.

They could be optimizing for some kind of softer metric like consumer employee happiness, or satisfaction.

HEFFNER: Oh, that's, that's a positive thought.

KRUMME: Some people, some companies do do that, whether it's a marketing tactic or, you know, they actually are on some level optimizing for those things.

You know, they're companies that claim to optimize for environmental outcomes among other things.

So I think that the point is that optimization as a way of organizing, whether it's a corporation or organizing the way that we understand our daily lives has, has really come to dominate our culture in the last decades and centuries.

HEFFNER: So take us through that history as you chronicle some of it in the book.

When did the idea of optimization really start trending in society?

When did people think of it in the way that we're thinking of it now?

KRUMME: If you're asking about trending or, or how we think about it now, I think that is fairly recent.

Um, I do, in the book trace kind of optimizations longer history, which, the word itself dates back to a French novella called Candide by Voltaire, in which at least as far as the etymologist can trace back that's the first use of something akin to the idea of optimization, which is l'optimisme, philosophy espoused by one of the characters in the book.

And it's this idea, it's sort of a buffoon of a character, this character named Professor Pangloss, who, every time the character, the other characters are met with tragedy, they encounter a bad situation, Professor Pangloss has kind of the optimistic view on it that, that things are living in the best of all possible worlds.

And this earthquake that killed scores of people, it was actually all for the best and personal tragedies that, that befall individuals are always for the best.

So that idea of l'optimisme eventually evolved into the English word we know today as optimization, which is kind of a more active version of that, which is we're going to make the best of all possible worlds through our ingenuity or engineering.

HEFFNER: And how did you find different sectors, different industries, different people over time to define what it means to create, because I think of optimize also as creating ideal circumstances for, conducive to growth, whether that's personal growth or economic growth.

But I'm interested what criteria have been set up in different areas that you write about to say, this was a successful case of optimization versus something that was less successful or unsuccessful.

Is there a certain criteria defining optimal conditions?

Like when you get to optimal conditions, you don't start at optimal conditions necessarily?

What defines when you get to them?

KRUMME: I think that's a separate question of how I chose the case studies in the book.

How do we define a successful optimization, right?

That's sort of baked into the technical definition of an optimization, right?

We set up an objective function, which is what we're trying to maximize.

And we set up some parameters or constraints.

And then in one of several ways that technology has afforded us, we solve that optimization.

I suppose a successful optimization is one that gets as close to the theoretical optimal as we possibly can, right?

So, a corporation that's optimizing for shareholder value, their success is defined by how much that shareholder value I improves over time, at least according to that narrow definition of success that they've set up through their optimization.

HEFFNER: The areas where you focus, you focus on the oil industry, ecology and the kind of contemporary farming, and then Silicon Valley.

What aspect of it did you enjoy thinking about the most, in in the various stories that you're relaying in this book?

KRUMME: I really enjoyed pretty much all the, the people I interacted with and who lent something to the story.

I just find people interesting in general, especially ones that are kind of embedded in a particular industry or on a particular quest.

So that had a lot to do with how I chose some of the stories to tell, right?

Whether I was compelled personally by these characters.

And then, you know, it was a question of whether they would lend something to the overall story I was telling about optimization and, and how it co came to, to be our, our dominant philosophy.

HEFFNER: But your ultimate thesis, I think you're concerned about optimization.

I don't want to say you think that universally is failing us, but what do you think after all this study?

KRUMME: I'm still fascinated by optimization, both as a technology and as a mindset.

I think I set out to try to answer the question of how it became our dominant philosophy.

I do think the book, you know, comes up with an answer or answers to that question.

I think the implied question is then, well, what, what comes next?

Or what do we do?

I tend to, both because I don't want to sort of perpetuate this, so there's one best answer, right?

I don't think there's a best answer for what can we do?

I don't have a prescription.

But I talk at the end of the book a lot about what I sense as kind of yearning in society right now, and I do sense a lot of yearning to move away from optimization and especially the trade-offs and the losses that this singular mindset has brought.

So I think we see that a lot across the board there are in many communities a desire to return to more local things, whether it's local foods or local makers and economies, to things that feel more tangible and, and real.

And, you know, I've seen that happening, um, everywhere from rural places to urban places where there's a desire to buy things and to interact with makers and businesses that are connected to the community rather than abstracted away in some corporate headquarters somewhere else or in some other country.

I think it's also just expressed in a desire to get away from this kind of overly fast paced global churning economy, right?

Whether that's in young people's choices of careers or choices of where to live or what communities to interact with or kind of the pace of life that they're choosing.

So you know, I don't know that that's a prescription, but I think it's sort of some strong signals about what might be on the horizon.

HEFFNER: No, that's helpful and thoughtful.

So your suggestion is that people refocus on maybe the, the quality of the business, the quality of the product, the quality of their life, not the speed at which they can achieve all those things.

That's what I hear you're saying, that's what you hope.

KRUMME: It's just what I hear people want.

And not everybody, right?

Some people love the speed, you know.

There are parts of it that I love too still, right?

I think it's fascinating that we live in this time, right?

Where we can order something from around the world and it can arrive at our doorstep in under 48 hours, right?

And there's all kinds of dark implications to that, right?

Like, what are the labor practices?

Where some of these materials come from?

What are the downstream effects?

So I'm definitely not advocating for this sort of universal return to some supposed quieter Golden Age where we were all subsistence farmers and lived in, in harmony with nature, right?

I don't actually think that moment exists.

But I do see kind as a trend or on average a step or two back from the over-optimized world that many of us participate in.

HEFFNER: I think one of the concessions or conceits or admissions in that is that I think we both, but I'll just speak for myself, I see optimization, the reality of optimization as outsourcing, right?

From the domestic American perspective.

It's hurt blue collar America, but it's hurt all of America.

That's why I asked about the criteria for optimization or the ethics governing optimization, because theoretically you can optimize in a way that's not going to force everybody to be in the millisecond by millisecond economy where you have to be there.

You can't have any reprieve, have any solitude.

I very much hear what you're saying, and, and I think part of the responsibility of stakeholder capitalism or compassionate capitalism or whatever you might want to call it, a socially conscious capitalism, is to refine and sometimes redefine the criteria, for optimization.

Because if it is entirely tied to the profit motive, and the ease with which you can, and the quickness by which you can achieve that profit motive, then I think things are not going change.

The culture is not going change despite what you say, which is the longing.

And I can testify that to that too, the longing for it to change.

It won't change if we don't set a rules of the road or ethics governing what it means to successfully optimize.

KRUMME: There's certainly that perspective, and I think we hear it from the corporate world.

We hear it from Silicon Valley.

I tend to be a little bit cynical when I hear things like that that's we're, sort of to paraphrase, it's like we've been optimizing for the wrong thing.

We've been optimizing for revenues, we should actually be optimizing for something more quote unquote compassionate, right?

Environmental outcomes or social outcomes.

And often, you know, I'm cynical because it's often simply a marketing ploy.

But I also think it distracts or it's an attractive kind of answer because it allows us to keep optimizing, right?

We're just changing what we're optimizing.

HEFFNER: Right.

And I think it's about quality versus quantity.

And in so many respects, we optimize quantity over quality and you know, and then we don't get the fundamentals when it comes to quantity, like cost of living or actual longevity, right?

Life expectancy has declined in the last several years despite the fact that billionaires are throwing so much money on living on Mars or the Moon forever, right?

So there are priorities that do not reflect the body politics' will, the longing that you referenced.

And just as a final point, understanding can finally crystallize what I meant when I said that about AI and ChatGPT before: It's like it's an open canvas, and that's what's attractive in the sense I'm going to have a five-hour date with ChatGPT to actually brainstorm ideas.

And when I refer to it being kind of in some respect the opposite of efficiency, because it's regenerative of our imagination, even if we're dialoguing with a computer instead of another human being, there's something about it that is counter-cultural in a high-tech world.

Because if it gets the human mind to spend more time and attention with one subject matter and think it through, I think that helps us.

KRUMME: That's really fascinating.

That's the first time I've heard that argument for AI.

I mean, I personally disagree.

HEFFNER: Now you get it right, now you get where I was going with that.

KRUMME: Yeah, I get it now.

I would personally disagree, but it's certainly a fascinating argument that AI is an open-ended diversion, or it can serve that way.

I guess my question would be, it's still just one part of AI, right, the ChatGPT side of it?

Um, I guess it wouldn't be my preferred kind of way to experience open-ended creativity and diversion.

But I certainly see that argument.

HEFFNER: Coco, you have this new book.

It's called, The False Promise of Optimization.

Optimal Illusions in the title.

Thank you so much for your insight and for writing this book.

KRUMME: Thank you.

HEFFNER: Please visit The Open Mind website at thirteen.org/open Mind to view this program online or to access over 1500 other interviews.

And do check us out on Twitter and Facebook at Open Mind TV for updates on future programming.

Continuing production of The Open Mind has been made possible by grants from Ann Ulnick, Joan Ganz Cooney, Lawrence B Benenson, the Angelson Family Foundation, Alfred P. Sloan Foundation, the John S. and James L. Knight Foundation, William and Flora Hewlett Foundation.

JoAnne and Kenneth Wellner Foundation.

And from the corporate community, Mutual of America.