Keynote
Speakers: Joi Ito
Date: November 5, 2017
Transcript By: Bryan Bishop
Category: Conference
Media: https://www.youtube.com/watch?v=3pd6xHjLbhs&t=866s
Welcome back to Scaling Bitcoin day 2. I am sure everyone went home and tried Ethan’s wagering setup yesterday. Today we are going to be covering topics like layer 2, fees, and consensus. To get us started today, we have a very special guest, Joi Ito, director of the MIT Media Lab, and he will be talking about parallels to the internet and ICOs and he will cover a lot of ground. Joi?
I think everyone knows what the MIT Media Lab is. I am its director. I was involved in getting the digital currency initiative (DCI) going. I am trying to keep up with what’s going on. I am not in the details as much as everyone else. I have been doing this since we setup a digicash server in the 1990s. It feels like we have been doing this for a while.
Let me see if I can frame some ideas for you.
One of the things I think a lot about is complex self-adaptive systems. If you think about how our environment works, you have roughly but it’s changing now, stable temperature on a global level. Our body temperatures tend to be roughly around the right same temperature when we are in a healthy state. At different layers, you have protocol changes, species collapsing but what’s interesting about complex self-adaptive system is if they are developed properly then they can sustain change but have resilience by adapting to change.
If you think about how the earth has evolved, there was a point in history where a mutation happened and photosynthesis was created, and now we can take photons and carbon dioxide and create a currency which is glucose and oxygen. And once that started, other processes realized we had an abundance of glucose and water and then other processes evolved that took glucose and water as input and generated other things. If you look at the environment, you see many different currencies, such as water that holds energy, not everything can use the same currency there are different processes that take currency as input, do a transformation into another currecy. When there’s an abundance of a certain currency, sme processes start to eat that and turn it into it.
It’s all interconnected. You come up with multiple complex networks that interact. We have a body, we have a microbiome an immune system and other aspects. When your brain tells your body something it sends a signal but really there’s a T cell required to do a transfer into acetylcholine. Somehow your body decided to use the immune system to do a transformation.
With these complex networks you have vibrant self-regulating system and that regulation helps keep our body temperature the same and gives us pathological resistance to invasion and years of evolution have created a robust system that is under attack by the exponential growth that we have.
It also under attack by the fragile notion of having a single currency for trading. How can someone be smart if they are not rich? We measure everything in financial value. What’s exciting about currencies is when they are not used for exchange.
I used to play World of Warcraft a lot. I played 8 hours per day. We had a guild. One of the great things is that… you can buy gold but you’re not supposed to. Our guild had 100s of members. We would kick out gold members. If Neha and I are going to do a raid the next day she has to take care of her kids at night so let’s say I stay up all night mining and collecting gold to help her buy her cape that she needs for the raid. Well if she knows that I have been up all night preparing and so on there’s a lot of value in the transfer of the gold to her.. but if she knows that I bought it for $20 from a gold farmer it’s not interesting to her. The items that I had in the game I remembe rthinking this they are more important and valuable to me than anything in the physical world. It’s because you couldn’t transfer the value out. You could go buy gold but they would ban you if you did that. And 10’s of thousands of hours would disappear. The value of WoW gold as a currency was its inability to be fungible to cash.
When we talk about cryptocurrency it’s a design problem that people are thinking about fintech and liquidity. I think it’s because of thinking of the world as economists rather than thinking about the world as a complex system. Another piece that we need to thin kabout is machines and artificial intelligence.
Norbert Weiner was a mathematician at MIT and he is gone down. He thought that machines were coproporations of flesh and lbood. And if you think about rules in governance as a kind of autmation then beauracracy are really machines. When yuo think about corporations they behave like an AI that we don’t have yet. When we think about the future of AI I think we’re already there because we’re doing this manually. When you think about sort of how a legal entity like a corporation.. we are already treating it as a person, it can live beyond the life expectancy of a human being. As we think about how we integrate machines.. extended intelligence, it’s not artificial intelligence, I don’t think we will make AI I think we will take corporations and we will start to see machines in amplifying automation but the system dynamics of these entities are going to be a lot like what we think about in terms of corporations. If you in some cases we blame corporations for profit, we’re in Silicon Valley where the motive and goal and the operations of the corporations guide a lot of things that guide where resources are allocated and what people are incentivized to do. Corporations or machines wont make the system more wise just more computationally powerful.
In an ecosystem, we are tweaking it by adding different types of processes that are going t do different things that are to a complex system that are not well regulated right now because eah of the processes will become more computationally powerful. Someone sent me an emial the other day that the other machine learning people are computationalists rather than thinking about it as an interaction of networks. Thinking about it as an interaction of networks.. that’s how to think about it.
When you think about bitcoin, it’s the substrate on top of which a lot of this happens, is the market. You get capital from the market. You sell products. You pay workers. A lot of the exchange of currency between processes is done in financial currency. Just like machine learning and AI we will augment and extend “intelligence” and I think a lot of the tools today will create inability for these systems to transact in new ways.
I urge us to think about climate. We have a climate problem because we don’t like geoengineerin. We try to increase earnings and take more from ourselves and exploiting environments. A lot of the tools of post-industrial society is really about… if you look at post-war europe or post-war japan, it was about rebuilding these countries, it was about roads factories and maslow hierarchy low levels. Most of the problems we have today are obesity and climate and they are occurring between we are trying to create more stuff.
We should create ecosystems that flourish but aren’t about getting more stuff. I think the appropriate application of being able to process transfer of value is a system not just an exponential curve and worship of singularity and progress. You will make a more flourishing system.
I am going to pivot because many people have talked about this. I am sort of an internet guy. I look at the world from the lense of the internet. The success of the internet is there are a couple of key ideas that we can use. One of the internets success is that it is a vibrant of many products and ideas that connect to each other from different protocols. I invested in twitter, and people said it’s a feature not a company. As it began to develop, we found that even if it’s just a feature it’s still a thing. In the old days you had monoliths and then you had features and relationships between features allows angility and it allows the system to be dynamic and responsive. It allows innovation at the edges, instead of monoliths. Ecosystem model of innovation is really important.
If you think about the internet compared to previous systems, previously we had telecos and they did everything. The success of the internet was layers. There was ethernet which was designed by Bob Metcalf. And then TCP/IP which was DARPA and universities. And then HTTP which was Tim-Berners Lee at CERN. And then you had Netscape which made SSL and encrypted communication over the web. Each of these spawned a layer of companies. Metcalf created 3com for the switches. Stanford gave birth to Cisco which became the largest company in the world. Not just Amazon, but a whole bunch of for-profit companies got created when the web happened. We were able to do some payments when we got SSL.
When ethernet was being fought over and many of you still be old enough to remember, we had token ring, appletalk, novell networks, so many network protocols. Tokenring was like ethernet and TCP it was multiple layers like ethereum trying to do multiple layers. Where the layers are cut is not technologically determined. It is determined by communities and marketing.
It had the right people getting the right people at the right place. Similarly with TCP/IP there are groups of people getting together to solve problems. And then people wanted to connect things together. Before HTTP we had SQML and all kinds of markup languages.. and waves.. and gopher and it was the simplicity of the format and openness from an intellectual property perspective and how they recruited others.
While they were elegant protocols for each layer, to me it’s more of a community thing, it’s the community of people getting together and doing things and convincing everyone else that we will all use the same TCP/IP. The only TCP/IP thing worked because it was the only protocol. The IETF and ICANN I was on the board for 3 years.. it’s a mess, but you really want ICANN instead of IUC… You have smart technical people in ICANN. ITU said oh we should run this. And they wanted to assign numbers like postal code.. they wanted to give them to each country in serial, even the people who weren’t super technical knew this was way too dumb. So by having the core community of technical people who can figure out how to keep track of the routing tables, that gave ICANN the legitimacy to run things. We could have designed ICANN better. ICANN was physical meetings. Governments, IP, the registrars, the registries, people at large, we had a big room, open minds.. these meetings would go on for hours and people could come up and say whatever they wanted. By the end, we would come to a consensus, but everyone would disagree, and people would talk for hours, and then people would say we’re tired let’s just do something stupid. If you were in the room and you already had a chance to complain, there were some people much less likely to complain. So we let them talked in person, and they wouldn’t complain any more online.
I notice in the bitcoin community that online is not a good place to get consensus. You need to get them in the same room and have a back door meeting instead. The ICANN process is actually really important to me and I think we hsould move toward that. I think there’s something to be said for this. Each of these layers like IETF have their own process of doing consensus but I think that’s been key. And the reason why these are imporatntt…. Minitel is probably the most successful video text system… Everything from the wire to the software and so on, was managed. It was unable to innovate, it died. It was important that it was done. This is like R3cev to me. They are trying to become the new version of this fintech stuff, but as an incumbent trying to do the whole stack they aren’t fundamentally going to change the network. To me, AOL survived. If you ever remember the early days of AOL, it was dial up and portal and had whole stack. But they were smart enough to say that now that TCP is fixed, we’re not going to do access anymore. And once the web happened, they got rid of their portal and just did content and email. There are probably periods in AOL’s history where if you invested you wouldn’t have made money. To me, Coinbase is going to have to do full stack right now because the protocols aren’t fixed. They will probably survive because– I think hyperledger is similar, Brian unbundled layers, so that even if all the layers died, then Hyperledger might survive.. the companies that are willing to throw away and cannibalize layers will survive in some form. If you build something before you know the layers, then you have to make choices that by definition you don’t know the outcomes.
But going back to my earlier slide, you see the Ciscos of the world are the ones who re going to make the most money because they determined the protocols– but right now we’re investing as if it’s 1999 but the protocol is more like 1989. We’re a bit of head of our skis from a venture perspective. I think bitcoin will probably be the next layer. But I think, you know, ethereum is like novell networks, we have competing networks going on right now.
Another way to interesting to look at this is that you know you have these sandwiches, they are non-profit collaborative layers, good at bringing people together but important to remember they are non-government agencies and the competitive layers I would say they are good for scaling and execution. Non-profit communities are not good at execution but they come up with mindsets and ask questions like can we come to consensus on a protocol and then once you do that the execution comes in and they don’t worry about mindset and they just get shit done. You need layers of getting shit done between layers of protocols. So whether it’s Creative Commons and Flickr, you need a sandwich. The internet was unbundled and allow aggressive competing without screwing up your values.. you need some control of the mindset.
I think you know, we hopefully will figure out a payments layer, there might be other layers, who is going to be the stewards of those layers and how are they going to be sliced, I think that’s still open to discussion and innovation.
Lessig wrote a book in the 90s called code and other laws of cyberspace. Arhitecture is a technology. If you want people to not speed down a road, you can either pass a law for a speed limit, or you put in speed bumps, but they might have the same effect. Depending on what you are looking at and what your values are, you might use law, you might use technology. Social norms and market is the businesses and in the market. We do with self-driving car we do this survey in the lab and it basically said that most people surveyed we did millions of people, said you should sacrifice the passenger to save more lives. But they said, I would never buy that car. But everybody else should, and government shouldn’t regulate it. So the society is saying we should have passenger sacrifie for increasing security but that the market isn’t the solution and the law isn’t the solution. So who should be the custodians of different processes?
Think about initial coin offerings (ICOs)- you have the SEC, the market, tech, who should be in charge of figuring out that out… I think we have ICO fans in the room. I think someone retweeted that the difference between ICOs and raising money for a startup is that the entrepreneur can just walk away with the money in ICOs. The token holder doesn’t have rights against the company. To me the real problem that I have with ICOs is that the tech has a lot of pontetnial but going back to design its kind of I’m I get kind of emotional because the whole point of doing this stuff is that we think the financial systems are ripping us off because people in the cookie jar were stealing cookies. In ICOs we are creating a structure where we can steal money. It’s just bad. I think if you separate out the 3 elements of money- transactions, store of value, and unit of account. If you take miles, the reason why we don’t have crazy people getting ripped off with airline miles is because it’s pegged to something not floating in value. They exchange miles, but they don’t speculate because there’s no volatility. So you need someone to fix the value. Or make them non-transferable.
((.. missed a few lines…))
If you can figure out how to take cookies from the jar unfairly, then you’r epushing out the people that are going to make you look like an asshole. But instead people are thikning about becoming the next scammers. I think a lot of people are self-righteous they say it’s open-source so it’s okay if I steal money because I’m trying to do the right thing… but that’s what everyone says. The whole point of cryptocurrencies is to try to use tech to prevent people from doing bad things. But the minute you start thinking- I remember someone saying I hate how these right-wing media and use the media to tell the stories, so I’m goin to go buy a newspaper so I can tell left-wing stories. It’s something about how we are above being scammers just because we’re righteous.
I think there’s two things personally, I would never want to be involved in a company right now for an ICO even if it’s impossible to do something bad, because we’re right now we’re attracting all the wrong people… Even if you execute a perfect ICO and you don’t make any unfairly make any money, then I think it would still attract the wrong people that we are supposed to hate. This is the internet bubble, right? I was here in Silicon Valley during the bubble. You couldn’t make a good company even if you tried, because all of your staff would be talking to idiot investors trying to invest in pets.com. All of the people working for you would wonder why would I work for you if I can make money doing nothing instead? So it’s hard to get people to work for you. In 2003, there was no more money. The only people who were left were the people who loved the technology and didn’t know how to do anything else– so a lot of people became bloggers. Most of the best bloggers were a lot of people who lost jobs in the dot com bubble. Post-bubble was the most exciting period of the internet.
Right now we are in an ICO bubble era. I feel bad for you guys because you’re sitting around watching people makin a lot of money doing scammy shit. I think a lot of people will make money on the dot com stuff and walked away at the right time. but the majority of people wasted 10 years of their life and building something and getting it destroyed by speculators. There’s nothing worse for a CEO than overpriced stock because if your stock price goes above the real value of your stock then what happens is that the only thing that happens next is that someone find sout that your company isn’t worht that, and then you’re fired. So companies try to eep the ir stock price down, so that they outperform the expectations. So as soon as the speculators come in, you have to lie, steal and cheat to meet those numbers. I think right now is a bad time in the industry because things are overpriced.
But working on the internet during the bublbe… that’s when you build the relationships. The people I still hang out with were the people that I hang out with were the same people…
Landlord’s game was before Monopoly. It was to teach kids about how bad capitalism was. And it wasn’t that popular. Parker brothers changed the goal of the game but not the rules. You become a capitalist, drive your friends bankrupt and then you win. The rules weren’t as important as the goals. The goals dictated how the game was played. People say it’s about winning arguments… but the quadruple bypass burger is 8k calories. If you are 350 pounds you eat free. They know what’s going on it’s not an information problem.
Donella Meadows has this great thing where she is a systems dynamics person and she wrote this in the 70s about where to intervene in systems. These constants, these rules, these other things you can modify, but when you go down to the most important things to change in a system– like one is the goal like for Monopoly, but the mindset is important. Are you trying to make money and get more stuff? Or are you creating a beautiful network? Or are you trying to have a good time? Or wha tabout the power to transcend paradigms and think for yourself? A lot of us are still thinking like economists and not questioning fundamenta l value– is the goal to get happier and is money going to do that? No, I think that the digital currency should feel more like a left-wing conspiracy than a corporate conspiracy but we’re being usurped by the greedy yummy thing about making more money but the problem we’re trying to fix is that we have a stupid single currency that causes me to be wasteful and mean and greedy and I think what we want to do is figure out a technical way to break that mold and it starts by the power to trasncend paradigms.
I would finish by saying it’s important for conferences like this where people can feel free to question thsmelves, think for themselves ,and challenge authority. We need trust. The importance of IETF and ICANN was the ability to let people fight each other in person and hide that under the sheets instead of online. Right now we have splinter groups fighting over rules and tech instead of asking questions like what we want to build.. But the ICO problem goes away if we all agree it’s about stealing money, but in fact to create a method to do the system in a more fair way. And I think we’re not having that conversation, and we’re arguing about the details, and hopefully you guys will fix ICOs with Scaling Bitcoin. Thank you.