I'm willing to make a bet this morning that bet is that you've never had a commute that was more than one hour each day, 30 minutes there, 30 minutes back. And I can make that argument because of a study done in 1994 by Italian physicists, Cesari, Marti, apologies to Vitor for that pronunciation. And he studied urban growth and he came to a realization, a pretty obvious realization that cities grow as a function of transportation speed. The faster the transportation, the larger the city will grow. And we can see that all the way back starting in ancient Rome, right?
Most cities, pre-industrial revolution were walking cities. And so as such, resources were concentrated in the middle of the city and people didn't wanna live further than 30 minutes away that they could walk a kilometer, kilometer and a half. So most cities pre-industrial revolution were only about three kilometers in diameter.
But this changed rapidly as innovation took place from steam rail in the early 18 hundreds in London to streetcars in Chicago, which led to the first suburban boom all the way to cities designed exclusively, mostly in America for cars, right?
And you can see the expansion over time and also the distribution of this transportation system. And what you see is that Martis constant holds as these transportation speeds increased, cities grew. What Martti was saying in another way is that transportation is the lifeblood of cities. And if you look at the growth from ancient Rome to sprawling Atlanta, you can see that the population definitely increased as you'd expect. But the geo space, the space these cities covered, exploded.
In fact, the population increased seven and a half times. But you can see the geographic size was like 2000 times bigger, right? Which should tell you a few things.
One, this is not to scale at all. The Atlanta would be the back wall covered. I tried it. PowerPoint would not let me do that. But also that the problem from urban planners and urban designers gets really difficult really quickly. If you're designing a transportation network for these kinds of cities, Rome is easy, right? Three kilometers, all roads lead the Rome. They said that for a reason. It's pretty a simple, straightforward approach. When you get to Atlanta, how do you give access to the places the resources people want to be and want to go and make it easy and optimize it?
The problem becomes intractable very quickly. And that's something that we should be paying attention to because the world we live in used to be like Rome, right? We had little empires of identity. We controlled everything. We were masters of our domain, so to speak.
But we don't live in Rome anymore, right? We live in sprawling, chaotic Atlanta. And so we have a similar problem. How do we design these identity systems? How do we approach it? How do we think about it? Not just to give people easy access to things, but to optimize that.
Thankfully there's an answer and the author, I'd like you to meet this guy or girl, or I'm not really sure. It's Pfizer Polycephalum, right? AKA slime old, something I like to call the blob affectionately.
Now, you only need to know about three things right now. It loves oats, the kind you find in normal oatmeal. It can sense that food from a distance and it can move around the environment. A and the reason I say slime mold is the author of our solution is because scientists in Japan in 2010 didn't experiment, right? They put a bunch of oat flakes on a, a flat surface, they released a slime mold, as you can see.
And it did what you would expect it to do, right? It spread out, it found all the food sources. But after about a day, scientists looked again and it had refined all those connections.
It had optimized that network, which is kind of what slime mold does, which is nice. But what really struck them was when they laid over a map of Tokyo and they realized that slime old had recreated the transportation system of Tokyo. What had taken engineers decades upon decades to think about and optimize an engineer slime mold did in 26 hours. And the real key point, there's lots to talk about with slime mold, don't get me started, but the, the part I want to really talk about today is how slime mold raises the place of the individual versus the collective.
It makes it an active participant in the process of creating these pathways to use transportation and identity.
And so I'd like to talk about that this morning in three different ways. How if we think like slime mold about identity will be empowering, individuals will be enabling to adapt to the situation and ultimately will be symbiotic with individuals. And by individuals, I mean humans. By the way, how many times have you heard humans are the weakest link? I'm here to argue against that today. Start with empowering.
Now, slime mold is not a monolithic organism. It's made up of hundreds of thousands of independent life forms, all acting independently. They throb across the landscape with protoplasm, that kind of pulses. It moves about a centimeter an hour. But where it goes and what it does is all result of the independent actions of each of those organisms, which that simple mechanism leads to some interesting results.
I showed you the network already, the transportation network, but at other experiments, like they put a slime mold in a maze, they put oak flakes at two different locations and they realize that slime mold through a large series of independent actions can actually pass one of the tests for intelligence solving mazes.
And the way it does that is on that edge, each independent actor has a localized context that it's aware of hyper-local environment that it senses.
And then it uses a distributed policy that's essentially agreed to, in this case, it's fine food don't die basic, but easy to follow, right? And then the slime mold empowers those individual cells to make good choices. It's kinda like a two year old running around finding, you know, oatmeal, crackers or whatever. Same kind of process.
Now, historically, we've held on tight, right? With centralized identities and centralized identity policies and centralized decision making. And you'll hear a lot this week about how that shifting. You already have, you already will, right? You'll hear a lot about localized identity context as well. And really what this does is we're gonna follow the same approach. I think we need to shift from being centralized to thinking about how we can empower the humans as agents in our environments.
We need to give them context.
We need to help them understand policy and we need to empower them to make a decision. Victoria just got done talking about verifiable credentials, which is one of the things you'll hear a lot about this week. And I like that thought process, the shifting of identity to individuals. It gives them agency to come in as different personas, different options, present different claims. And it also puts them in a mindset to think about the implications of their actions. What identity did I share here? Who am I to this business? What risk am I presenting to the business and to myself as a human?
We're slowly empowering them with understanding and being able to use that identity context. Secondarily, we need to equip them with identity policy.
Now, I'm still gonna say this is centrally def kind of agreed to, if you will.
There's still a need for governance, for accountability, for audit, proving you're doing what you say you do. But policy is in danger of being what I would call the new ai just to overload my terms here today, right?
With ai, what happens? Oh, because math, I have to trust that people often accept policy blindly or set policy for their users without interaction. And if you think about it, we should be interacting with our humans all along the way. Think about the process. You develop a policy, you figure out kind of what it should be, right? Then you have to distribute and communicate that policy. Distributing is just making it available. Communicating is making it in a language that person or that entity can understand. Then it has to be explainable. What's the reasoning behind the policy?
Why is this being enacted? And then finally, it has to be usable in context.
Now, that depends on what resource you're protecting, but think about the recent explosion of generative ai as an example. How many organizations, how many countries, how many entities just said, we're gonna cut off all access to generative AI because it's too dangerous? Or did your organizations ask people how they were using it? Figure out what the use cases were, talk them through good decisions, figure out the policy in conjunction with them. Explain it, communicate it, make it usable. There are options and should be a dialogue all along.
Instead of humans being the weakest link, they turn into a force multiplier for our effective policy. Now it's a conversation and just a quick drive by. There are places for design, obviously in this designing products, design solutions that guide people like Ian talked about last night into good options and good solutions. And there's always gonna be an education factor, training people, hey, for them to understand what's going on.
But we tend to spend 10 to 20 times more on technology than we do on communicating and educating and being with our people, helping them to understand.
And that is a mistake. Cause what we want is to let them have the identity context, help them understand the identity policy, and then sit back and trust them to make intelligent good decisions. And that's what slime mold does. If you put it in a area with five to eight different food sources, it does what you think it would do. It goes out and finds the first one, which is really close by, and then it finds the others as you'd expect.
But then, believe it or not, slime mold can differentiate and says, yeah, there's lots of food here, but that's the best one. That's what we want our people to do. We wanna be able to trust them to make the right decision.
The one that enhances privacy, the one that enhances security, the one that benefits both. So we want to empower our humans, but we also want them to be able to be adaptable to help adapt that policy. Slime mold, another experiment, they put slime mold on one side of a narrow track. They put food on the other side. They put a low salt environment in between.
Slime mold did what slime mold does, which most organ most organisms would do. It said salt equals death. Hard pass. I'm not going there. But after they repeated the experiment, two days, three days, four days, five days, the slime realized this low salt environment wasn't that dangerous. And it started to accelerate and accelerate and it passed. Another test for intelligence. Slime mold can learn, hey, it remembers where it's been and what it's done.
The local, remember those decisions, that interaction with the salt was only on the edge of the slime mold, but it passes back up to the core, adapting that access model.
And traditionally, that's kind of what we do these days, right? We have a steady environment, we have some new identity context that exists on the edge. And what do we do?
We say, no way. Prove more to me that you need access to this resource. Or you are who you are. If you're in Tokyo suddenly, well, why are you there? The first time I'm in Tokyo, it needs to challenge me and shut me down. But once I prove with MFA or other options, then over time that becomes the new normal. And if I stay in Tokyo, that policy adapts, which raises a question that I don't quite have a, a concrete answer for. How do we construct the protoplasm, so to speak, that communication mechanism, to let our humans be active participants in the modifications of policy.
One hand, it could be automatic, right? As things change, we should be able to detect activity change in the environment.
Oh, this is a new normal, like I described. A lot of organizations are doing that with ai, machine learning, identifying what's normal and automatically modifying things. But there also needs to be some way of using prompted change. And if our users, if our humans are already in a dialogue with us about the policy, for example, like chat G P T or something else like that, then they would feel more comfortable coming to us and asking us to change the policy. Saying this is a new environment, this is a new circumstance. I need this modified. They expect to be listened to.
They expect to be heard, they expect to be empowered. So policy may be a lot of things, but it also needs to be adaptable. Not only do we need to empower our humans, not only do we need to make them able to adapt our organizations and our policy to change, but we also need to be symbiotic.
Slime mold. If you take two slime molds and you put them in the same area, one that's been trained on the salt environment and one that has not, two things happen.
First, they reach out and connect, right? And become one slime mold. The other thing that happens though is after that connection is made, learning is transferred now instead of two slime molds, one that knows that low salt isn't really that big a deal and one that doesn't. Now you have one unified slime mold that shares knowledge and knows it all. This means that slime mold plays well with others with itself, obviously it transfers. Learning says, Hey, I know about this situation and you should too. And finally, scientists love it, right? It's easy to grow cheap food.
The scientists get research results. The slime mold gets to eat. They're symbiotic with each other up.
True fact, if anyone's interested, side note, let me know. You can buy slime mold online. You can have your own pet, right? Easy to grow. Like I said, all you need is a cardboard box. You're not allowed to take it through customs, sadly. So I didn't bring my slime mold with me. But what that says to us is that all too often our humans expect us to be joined together when we're not. Right?
One component that uses identity knows about a compromised credential or an identity, while the others have no idea. And that's why standards, like shared signals are kind of the epitome of us using identity symbiotically internally, right?
Therefore, if, if one of us knows something about an identity, then we can all share that information and be symbiotic. And then secondarily, we need to be symbiotic, of course, with our humans.
You know, we look at marches constant saying, oh, people will only drive 30 minutes a day each way to work, or these days to school or to the social event, or to grocery shopping.
And we see that as a limitation. We need to stop thinking about them being the weakest link and think about empowering them into a force multiplier. An ongoing dialogue involves 'em in the process. It makes us stronger, and the individual ultimately can make decisions that serve the collective. There's a cultural change that we need to embrace, and we do that through those three things, right?
Empowering our human users, helping them adapt and adapt the policy and ultimately being symbiotic. And this, the slime old approach of individual agents is, is being used in a lot of different areas of science.
In fact, all the, the videos and most of the images you saw in this talk today are using open source slime mold models that you can download off of GitHub. Also super fun if you don't wanna do it in real life. And slime mold is, like I said, it's been using, it's being used to solve other intractable problems like advancing cancer research or hydroelectric power generation, or even modeling the depth of dark matter. And it's tracing that dark matter throughout the universe.
And ultimately, if we want to solve intractable problems like this, like identity as we face it, then if we want to use identity in ways that are empowering, adaptable, and symbiotic, then this morning I suggest that we all take a moment, pause and see what we can learn from the blob. Thank you.
Thanks so much, Mike. Another inspiring, awesome presentation. You delivered the shoes, the hats, the content, the entertainment, and the image of the blob. I'll long remember your presentation a couple of years ago about blue. It taught me a great deal about a great deal.
I've much appreciated, I dunno how, how accurate the statistic here is here. But question from the audience is given that 50% of the population is below average intelligence, can we trust their decisions?
That's a great question. I know many of those people.
I, yes. I think ultimately we're going to have to, if we're talking about shifting identity into the hands of normal people, then we can design systems that help them along the way and people are more aware than they've ever been, right? I'm pessimistic about the core of human nature. We can talk about that later, but I'm optimistic about them making decisions that make sense, right?
I, last year I talked about facial recognition here, everyone wants protection against facial recognition. And so everyone's motivated for privacy, integrity and, and things that benefit us all.
Okay, great. I think everyone in this room would agree with you that that identity is the lifeblood of an organization, but outside this community, how well is that understood?
I think more and more and more, not, I mean lifeblood of an organization, maybe not, but I think they see it as their own lifeblood, especially as individuals. Now think about the chat. G p t example I gave was pretty real. Our organization at SalePoint use that approach where we said, how are you using chat G P T? And people embraced it because they heard the reasoning, they understood it.
Another example is spam phishing reporting, right? There's an on, whenever we have a phish test in our, in our company, there's always a large discussion about, oh, did you catch it? Did you catch it? How'd you use your identity? Where those kinds of things, those discussions matter, right? And raise the bar.
So,
Okay, great. Thanks. Another round of applause for Mike Kaiser.
