Building Mental Models and Considering Multiple Sets of Priors

What you will learn

• If information comes by, it must be important (01:51)
• Vectors and Bayesian probability in mental models (04:29)
• Creativity is noticing what other people don’t  (06:48)
• If your dog could talk, it would show you a whole new world (08:56)
• Only when you have all the pieces can you put the puzzle together… (09:35)
• …what is the art behind it… (16:47)
• …and does framework help you find the pieces (19:56)
• Selfish individuals versus altruistic groups (20:36)
• Crypto is not decentralized (22:28)
• Tim finds out somebody else built his idea (27:48)
• Tim compares himself to Cookie Monster! (28:57)
• You will succeed when receptivity and striving are in balance (31:03)
• It’s all about tackling the hard things (33:03)

Episode resources

• Lao Tzu
• The Way of Life According to Lao Tzu
• Wallace Stevens
• Eric Raymond
• Christine Peterson
• John Maynard Keynes
• The Man Watching by Rainer Maria Rilke

Transcript

Ross Dawson: Tim, it is a great delight to have you on the show.

Tim O’Reilly: Thanks for having me.

Ross: You have lived a life immersed in information and helped others in many ways to point them to, and digest that information. How do you think about that idea? How do you approach unlimited information and being able to make that into something valuable?

Tim: Well, to understand how I think about it, it helps probably to say that I don’t have an approach where I really try to keep track of information, or gather information. It means certainly some ways I do, but my working principle was expressed very well by Tevye, in Fiddler on the Roof, when he said, “Good news will stay, and bad news will refuse to leave”.

In a certain way, my approach is that things come by, and if they keep coming by, they’re probably important, and if they don’t, then maybe they weren’t. But there’s a bigger piece of it, and this is maybe explainable by reference to something like Google Maps, where people follow the map, and they stop noticing where they turn, versus people, who, in the old days, had to learn and observe the world around them to the level of where you think about the South Sea Islanders, who could navigate by watching the ocean currents, and the stars. They were their own GPS, and they were continually taking in information, and noticing how it was different from what they expected or the same as they expected.

I work a lot like that. I have a mental model that I try to build of the world, and that model is inductive. Basically, I’m taking things in and I go, Oh, this is different, this doesn’t fit. A lot of the work that I’ve done over the years has simply been trying to construct a map by looking around. You pay attention to things and the things that start… with a soft focus. This is an idea, I think from hunting and things like that, you watch with a soft focus. Of course, I’m not a hunter, but I once read a book called “The Tracker”, and I took a workshop with one of the founders of this tracking movement. You’re just receptive and you’re open, and then certain things just pop out at you as anomalies. That’s what’s interesting.

Ross: That is the same as my thesis around how it is that we build these models of the world. I think, when people talk about mental models, they often talk about discrete heuristics, in a sense, whereas, a mental model is really, I think, more holistic. It is a mental model of the entire world or the entire world of business, and how that works. How do you frame this mental model that you have built in, are building?

Tim: Well, first of all, I do have a set of frames for it. One, I wrote a little bit about this in a recent piece. I also wrote about it in my book “WTF”, all thinking in vectors. This idea, that there are forces that are moving things along, and a vector has both a direction and a quantity, so what you’re looking for in a certain way is acceleration in a particular direction. You’re looking for how those directions collide with other vectors, and what the resulting outcomes might be. In this piece, “Welcome to the 21st century”, I wrote about one of the big impacts of the pandemic might well be that we never go back to the office and sure enough, that’s turning out to be a possible future.

That’s the other thing, I have a mental model that comes from scenario planning about imagining very different futures. Then when they start to come true, you go, Oh, okay, that one was more right than the other one. You start to solidify. Again, it’s very Bayesian. It’s simply, you have a set… I think part of it would be to say it’s a Bayesian system in which you have multiple overlapping sets of priors that you’re willing to accept, loosely. Then they collapse differentially. I think that’s the thing I don’t think people think enough about in Bayesian probability systems. It’s not just one set of priors. It’s a set of overlapping sets of priors, that could collapse in different directions.

Ross: Ones where you’re, of course, continually chained to the things which don’t fit and could help you modify those frameworks.

Tim: Yeah, I do think a big part of it is building your own map, and there’s a creativity to that. So much of the model of creativity and our culture is, it’s making stuff up. I think creativity is noticing things that other people don’t notice because it’s ultimately a kind of scientific process. Just think about something that we think of as traditionally creative, like music, somebody, they’re making something up because they saw a possibility that wasn’t there before. They didn’t just make it up for the hell of it, they made it up because they were seeing what the new possibilities were in… maybe it’s in an instrument, maybe it’s in a cultural milieu, but basically we shape reality by what we notice and choose to pay attention to.

I’ve always been a huge fan of the poetry of Wallace Stevens because that’s what he’s all about. There’s this underlying reality, and yet, we shape it. He described reality as the quest for supreme fiction, something that we could all agree on. But the contention between people is for visions of the future, and you certainly see that in the political realm, but you also see it in, say, paradigm shifts in technology or physics. Somebody basically convinces other people that this is the right way to think about the world.

Ross: Yeah. I think that way of “we make the world by how we perceive it”.

Tim: That’s right. We have this bit of an illusion in the West, that somehow our quest is towards the one true reality. When I go, Yeah, but tell that to my dog. We’re out there looking at things, and it’s like, what are you people looking at? This is a really good smell here, you’re not paying any attention whatsoever? I think that’s one of the things, of course, in science fiction; It’s like, how do we know that there aren’t completely different and equally valid, and maybe even more productive ways of looking at the world?

Ross: Absolutely. In “WTF”, you mentioned, compiling the pieces of the puzzle, before you put the map together. How do you find those pieces of the puzzle? Or identify them or recognize that they are pieces of the puzzle?

Tim: Let me back up and give you a little bit of color on that analogy. The point is if you imagine doing a puzzle, and all the pieces aren’t there, you can’t actually finish it. Very often, when you’re dealing with something new, the pieces literally aren’t there. I thought of that very vividly around my work with open source software, because I was thinking a lot about the fact… the first thing you think back is, Okay, what did I notice that other people weren’t noticing?

I noticed that the Free Software Foundation didn’t talk about a lot of the software that I was selling really popular books about, they were also free software. They talked about Linux, they talked about the GNU utilities, they don’t really talk about Perl, they don’t talk at all about DNS and BIND, and all these tools out of the Berkeley Software Distribution, because they had a map that was all about the license, and it was particularly our license. I’m going, Wait a minute, they don’t include the worldwide web, which was put into the public domain, something is wrong with this picture, they don’t include Sendmail, they don’t recall the DNS, they don’t include all the TCP/IP protocol suite and implementations of that, so it goes, something is clearly wrong with this picture.

I thought, well, I’m going to bring all these people together to talk about what’s wrong with the picture. At that meeting, Eric Raymond says, Oh Christine Peterson came up with a new term three weeks ago, and she was open source, and we debated it. I was like, I identified a gap. I didn’t know that there would be a new name for it that was going to show up, but it showed up just on time. If I had done my meeting… and I kept going, why am I in… there’s an intuitive part to this, there was a part of me, like, why am I delaying having this meeting? Am I just being a slacker? Then you look at how it worked out, the timing was perfect.

Another one that was a little bit like that was when I started thinking about the licensing wasn’t the key to open source, it was really collaborative software development, it was a way the internet was enabling new kinds of collaboration, including software collaboration, and distributed computing, and things like SETI@home with distributed computation, and Napster with file sharing, they all started talking to me about, there is a new paradigm emerging. I kept following, tugging on that thread, and trying to integrate into some new map that would make sense to me.

I eventually came up with this notion that we’re building an internet operating system and it’s going to be based on data. As a result of that, I launch the Web 2.0 events, but then I launched one called Where 2.0, and I was like, guess what, location is going to be one of the big subsystems of this internet operating system. It was great. We’ve been promoting the event, and a month before we went live, Google approached me and said, Hey, we’re doing this new thing, any chance that we could introduce it at your conference?

That was Google Maps. I saw that there was a logic to this thing. I didn’t know anything. I didn’t have some inner intelligence on what Google was doing. I wasn’t really paying attention to the news, they just showed up because I’d actually built… this is what I mean by the pieces of the puzzle suddenly showing up and you go, Oh, there’s the piece that was missing, and you dock it in place, and everything starts to make sense.

Ross: The word “framework”, I think, is really relevant, because that’s exactly as, for example, the frame of a painting, what goes in and what goes out. The framework is the frame, where you can see what are the pieces, which fit within that frame. In that case, I don’t know whether it is you personally or it came up with the term Web 2.0…

Tim: No I didn’t actually.

Ross: But then there is this name, or is this something where you can have a label or something where you can then communicate about what that frame is?

Tim: Yeah, that’s right. I think that’s absolutely right. The name often… it doesn’t match. Web 2.0 had a lot of baggage. Here, I’m talking about the internet operating system, and it was just too geeky. Then Dale had already come up with the name Web 2.0 in a brainstorm meeting with Craig Klein. They were trying to come up with a way that our two companies could do an event together, and we were thinking, what could we learn from them, somebody else was doing events. Dale was like, well, it was really around the second company of the World Wide Web after the .com bust was what Web 2.0 stood for. Dale had developed a series of things that it was about.

Then I just really fleshed it out with a bunch of the ideas that had previously been calling the internet operating system, and the name took off. There was also an aspirational piece that I think was very interesting about that because Dale and Craig came up with that in 2003… I had been doing this internet operating system stuff since 2001, had this P2P conference, and then that emerged into this thing we call E-tech, the emerging technologies conference, which is really exploring this idea of how the network was changing the way we would all interact. I guess it wasn’t that we had thought everything through but we had a model into which the future shows up and it starts to make more sense.

Again, another analogy I use is a little bit like a hologram. You have the big picture there, but it’s fuzzy. The more data points come in, the clearer it gets.

Ross: You may have already answered this, but if you’ve got the pieces of the puzzle, and you start to see those map out, and then you can see the framework is there, then what is the art of putting the pieces of the puzzle together to form something, which is a whole?

Tim: I think the one piece of the art is patience. Lao Tzu, the author of the “Tao Te Ching” says, he’s talking about the qualities of the wise man, and he has a bunch of them, I say, winter fare on an icy stream, I forget what they all are, but then the last one is… but also this, royal does a torrent. Why royal does a torrent? Because sometimes there’s nothing to do but wait until the stream clears. I think that waiting quality has been a big part of what I’ve done, where I haven’t rushed to try to make a story where there isn’t one.

That progress for me from open source to Web 2.0 was a seven or eight-year process. If you look at the history of the talks I was giving, I was feeling my way towards this new paradigm. Then how did I get from there to the Gov 2.0 stuff, I started thinking about the lessons for the government from technology platforms. It was a set of conversations with people and somebody would show up with a piece of the puzzle and I’d go, yes! I still remember the conversation where I got off on government as a platform, but then continuing down that path that led me to all the work that I’m doing right now about marketplaces and anti-trust. I kinda threw this idea of algorithmic regulation, which is something else that I came up with.

Again, not all of these things catch, but the point isn’t… because I’ve had some things that have really taken off, everything sets the objective. It isn’t really, the objective is, just for me, to make sense of the world.

Ross: Absolutely. I think that’s the same in my own way, I create frameworks myself, and I happen to share them, and sometimes people happen to like them. But that’s the secondary aspect.

Tim: Yeah.

Ross: Coming back a step. Is this some kind of framing purpose that helps you find these pieces of these puzzles, these maps which unfold? Have you framed some way in which there’s an intent or around what it is within your purview of what you are considering?

Tim: Well, I guess there are definitely some big picture things, and they are more values than anything else. Again, I’ve just recently got some interesting new language for this, from evolutionary biology. David Sloan Wilson has been doing a lot of work on what he calls multi-level selection, which he sums up, I guess, it was he and Edward O. Wilson, who was one of his teachers, they’re not related, even though they both have the same last name, Wilson, which is, selfish individuals can outcompete altruistic individuals, but altruistic groups outcompete selfish groups. There’s a lot of fabulous research on these alternating levels of what he calls multi-level selection, where there are certain behaviors that are at the individual level and others that are at the group level.

That notion of the alternation of individual and group actually goes all the way back to some early learning I did in the 70s, in a context that I don’t want to unpack here, but that was very fertile ground for me, but I’ve always thought about cooperation, and what encourages cooperation. We have a metaphor in our society that says it’s about winning, that this capitalism is all about competition, and winner takes all. I guess I have a set of values that I, in some ways, I guess, I’m always trying to explore and justify why that isn’t true, which is why I was attracted, say, to open-source software. I was like, you say, it’s all about competition, but guess what, look over here, Microsoft was so competitive, they killed all the innovation, and all these people went off, and they just started fighting around with the internet and open source, because hey, they could, they could cooperate, there was this new model.

But even looking back, again, I guess, in some ways, one of the big shaping maps of my technology career was watching the alternation, where IBM was dominant. I came in, actually in the tail end of the minicomputer era, not in the PC era, and I watched the PC blow up, because there’s this democratization of access, then I watch Microsoft win, and replay that the tragedy.

Then an explosion of innovation and decentralization gives us the internet, open source, and then you watch Google and Amazon and the like, replay the centralization, and bit by bit abuse of power model, and I go, well, I’ve seen it now three times. I go, so value is going to go somewhere else because I guess, one of the basic… I have this big picture idea that if you take too much of the value, the system breaks down, and people find new niches. That’s actually a fundamentally ecological concept. Again, there’s a lot of analogies that are not exact but help shape your thinking and your map.

Ross: The more analogies you have access to, the more you’re able to perceive things, which can be useful for framing things?

Tim: Right. Oh, this is like that; Oh, that means that this is probably interesting. It can lead you astray. For example, because of my centralization versus decentralization narrative, I’ve probably been more skeptical of cryptocurrencies than I should be. Because everybody was like, Oh, this is all about decentralization, I go, Yeah, it re-centralized faster than any technology in history and it fits exactly the pattern where…

Here’s IBM, that’s got their centralized power of the computer industry by control over hardware. They don’t realize that this changes, when they have commodity hardware, they release the specs for the IBM PC, they don’t think it really matters, software is just something that goes with the hardware, but Microsoft basically makes a new explosion, a new power center, that center on software, they’re all about control of software API’s, they become dominant, they don’t realize this new thing, where software becomes commoditized by open source in the internet, then it’s all about data.

Now we’re in this paradigm shift around AI, we don’t quite know what that means. I see some things that I think are really interesting, where the models that are being released, often by the big companies, I think are going to undermine them in the same way that the PC undermined IBM, and the internet undermined… so I’m aware of that. Everybody was saying no, the next thing is crypto. Maybe, but what I saw was centralization.

Once again, the centralization is not in the original model. The centralization came through energy, not through… so that’s interesting, because of course, that starts to intersect with the whole world problem, the energy is going to be one of the critical pathways that differentiate whether we survive or not. There are some arguments that maybe crypto will accelerate that because people will go oh, well, this is a great way to make lots of money, but we need to have super cheap power in order to do it, but we’re not there yet. But you go, Oh, there’s something interesting though, this repeating pattern, maybe crypto is the way forward, but maybe because it centralized so quickly, that won’t happen.

Ross: Part of it is that the technologies are decentralized, but the economic manifestations of them are centralized.

Tim: Yeah, that’s always the point. The internet was also fundamentally decentralized, but the economic models became centralization. This goes back to this multi-level selection played out in culture, which is that we are seeing this dance between competition and cooperation.

Ross: Which I think is something which can help frame those models and whatever frameworks. Some of the powers of thriving and overload or filtering and focus, I get the sense that… at the beginning, you alluded to essentially allowing the wash to come over you and what comes, what you see, the things that you see are the ones which are the things which are relevant, I mean, are there any…

Tim: Yeah, I think that’s right.

Ross: But are there any particular tools, approaches, routines?

Tim: Well, first off, I will say that there is a real downside to my approach. Because I’m not very goal-directed or my goal is to make interesting things happen for other people, I miss a lot. I have this one very funny experience where there was a startup I heard about, that had been funded by a friend of mine, and I went, holy cow, that’s super interesting. I reached out to the guy who’d given them funding, and I said, Can you introduce me? And he introduced me to the founder and the guy was like, Tim, it was your idea that we built, we came to see you and we had a bad idea, and you didn’t like it, and you told us what we should be doing, and we went and did it. I go, Well, how do I manage to not then have them come back and tell me, we’re doing your thing. I would have invested in it, it would have been a good exit. But I just don’t think that way a lot.

I sometimes laugh about myself, I’m a little bit like this episode, I saw when my kids were little of Sesame Street in which the Cookie Monster won some game show and he’s now at the section where he gets to choose his prize, and behind door number one is a million dollars, behind door number two is a Chateau in France, or something like that, behind door number three is a cookie, and we all know what he chooses. For me, the cookie is just interesting people doing interesting work that seems to make the world go better. That’s why I tend to surround myself with people who pick up on things that I find interesting, and then pursue them methodically because I don’t.

Ross: Yes. Coming back to when I was asking about purpose, I mean, in a way, you’ve framed it just now, and the side effects of that purpose are wonderful.

Tim: Yeah, I just love to have… at my events, I like to connect other people. The things that I get so excited about is hearing that, Oh, you guys met and cooked up something amazing. Again, we did put something, for example, a planet, the satellites, it was sort of formed as a result of our science fruit camp and that led… we did actually invest in it through our venture fund, but it wasn’t because of me, I was just like, hey, let me put these interesting people together. These guys from NASA are making shoe box satellites, that’s super cool. Let’s invite them to this event. There’s a lot of things like that. That’s, as John Maynard Keynes used to say, my jam.

Ross: Yeah. That, as you well understand, is how the future is created.

Tim: Yeah.

Ross: Any final words of advice for anyone who is struggling with far too much information and trying to make sense of it?

Tim: Well, again, maybe I’m spoiled because, of course, I’ve been quite successful, and I’m not there like an early-career person who’s trying to make their mark in the world. But I think we try too hard.

Again, lots of my mentor says let life ripen and then fall, will is not the way at all. This is the wonderful Witter Bynner translation from the 50s or 60s, I’m not sure exactly when it was. But we need to have an attitude of receptivity for any of this stuff to work. That’s not terribly compatible with a certain kind of striving. Now, again, striving really works. I mean, there are people who are way more successful than I am, who are hyper-competitive, and they’re trying to make some particular thing happen, so there’s more than one way to do it. As Larry Wall, the creator of the Perl programming languages used to say, I can’t tell you that, my way is the best way, I can tell you that it’s been good for me, and it helps me align my work life and my personal life.

Again, you’re trying to make something happen, but you’re also trying to listen a lot, and it’s finding the balance between the two. I guess the other thing that’s been very shaping for me, is this wonderful poem of Rilke, called “The Man Watching”, in which she says, this is my rough translation of somebody else’s English translation from German, but saying that, Jacob, in the Old Testament wrestling with the angel, and Rilke says, what we fight with is so small, and when we win, it makes us small, what we want is to be defeated by successively greater beings.

I think there’s certainly a point where everybody thinks that it’s about success, and if you have the perspective that we’re all ultimately defeated, again, Rilke’s phrase, we come away stronger from the fight, you’re going to tackle hard things, and it’s not about winning. Rilke says winning does not tempt that man or that woman. It’s just like, you are about engaging with the world in a way that’s productive in the process, not necessarily in the outcome. Because ultimately, we don’t win. We can just leave things a little better than we found them.

Ross: Yeah, I think that’s absolutely relevant for everyone in these times because there’s no such thing as a win, in a way, as you say that.

Tim: Yeah, and that helps you select what to pay attention to, because you can start to say, Oh, my… if your fundamental goal is to make other people better off, then you value things like cooperation, you value things like making connections that allow people to do things that you wouldn’t be able to do by yourself, you realize that, Oh, you value telling a story that lets other people make sense out and see opportunities that you couldn’t pursue. That’s why at O’Reilly, one of our slogans has been “create more value than you capture”. We put together information that helps other people to do things, and that’s the heart of our O’Reilly online learning platform today still. It’s just like, how do we teach people to follow what they want to do? And it’s not directive, it’s enabling.

Ross: Absolutely. Tim, it has been a delight and an honor to have you sharing your insights. Thank you so much.

Tim: You’re very welcome. I enjoyed talking with you, too.