John P. Conley: Blockchain, Game Theory, and Empowering Humans Above Machines

Google+ Pinterest LinkedIn Tumblr +

John P. Conley recently spoke to Blockleader’s Fernando Sanchez about Game Theory, developing cutting-edge blockchain technology, and worldwide cuisine. The conversation was lively, educated, and it left me feeling quite hungry afterwards.

John is an incredibly smart man. Among many other accolades, he holds a Ph.D. in Economics from the University of Rochester. He has taught the subject at such prestigious institutions as the University of Illinois and Vanderbilt University.

His achievements include a year at Microsoft working on cryptocurrencies, biometrics, FCC spectrum auctions, the Internet of Things, and game theoretic aspects of WiFi and PCS radio protocols. And to round things out, John is the co-founder and first president of the Association for Public Economic Theory (APET), the co-founder and editor of the Journal of Public Economic Theory (JPET) and of an open-access letters journal called the Economics Bulletin (EB).

He talks with confidence about topics as disparate as gaming theory, blockchain, network theory, and that staple of Mediterranean cuisine, the almighty paella.

And he’s an affable, warm, and humble individual to speak with. His laid-back but attentive demeanor turns a conversation with him into a pleasant experience.

I caught up with John recently and we sat down to discuss anything and everything that his sharp mind could offer. Here’s the outcome of that conversation.

On personal and professional motivation

Motivation is a very personal, intimate, and complex thing. What drives us to do the things we do is propelled by a myriad of factors, both personal and professional. John spoke candidly about this and how the challenge of finding better ways of doing something drove him to work on resolving a number of inherent problems with current blockchain technology.

“Us economists suffer from a disease: We don’t like inefficiency. So when we see a thing that we think can be done better, it irritates us. It feels wrong. Fixing that sort of thing appeals to a certain type of person, that’s me.

When I first started to read about blockchain, I was taken both with the potential and the elegance of the solutions. As I got into it more deeply, I began to have more questions. Many projects were solving the wrong business problem, or not solving the problem at all. The heart of blockchain is the consensus protocols they use such as Proof of Work and Proof of Stake. I realized that they were ignoring certain kinds of strategic attacks that would invalidate the entire security mode. So, I thought to myself, I’m a game theorist, I can fix this.

“It turned out to be much harder than I thought! Some very deep problems had to be addressed, including certain fundamental impossibility theorems from computer science. Impossibility theorems typically say something like: if you want to have A and B, you can’t have C. We have these in economics too. The way to escape them is to realize you don’t really need A, since something very similar like; A’ will do, or that; C is more than you want, something a bit weaker or different from C will do. Generally, these theorems say that no more than 33% or 50% of your network can be dishonest if you want consensus on an honest blockchain ledger state. Our approach at Geeq was to construct the network and consensus mechanism in a different way that is not covered by these theorems. As a result, we have a security model that only requires a single honest node in the network. Bitcoin and Ethereum are 51% Byzantine Fault Tolerant, but we developed a consensus mechanism to increase this to 99% Byzantine Fault Tolerance.”

On game theory: the mathematical mechanisms behind rational decision-making and the Nash Equilibrium

We make decisions every hour of every day. What to wear, whether we want to buy coffee in a shop or save our money, where to eat lunch, and so on.

What we don’t usually think about is the complex set of mathematical rules and mechanics that affect the decision-making process. Game theory is the umbrella term for the science of logical decision making in humans. You’ll never think about your daily coffee routine in the same way after reading this!

“A game consists of three elements: a set of players, a set of strategies for each player, and a pay-off function that maps strategies into something they desire: money, rewards, prizes, or whatever it may be. That’s the basic structure of how agents interact strategically. Writing down the game helps you understand the problems that agents are facing. The next question is, what is the correct way to think about equilibrium in this situation?”

This leads to a very interesting part of the conversation. Within the realm of game theory, the Nash Equilibrium states that the game is in equilibrium if every player knows the strategies of the other players and no player stands to gain anything by changing their own strategy.

John explains this concept through a practical example.

“If all people within a given group speak the same language, why would I choose to speak a different language? No one would understand me. Speaking in the same language is a Nash Equilibrium, since no single player would benefit from choosing to speak a different language. This is called proof against unilateral deviation.

It is easy to see that this equilibrium is not proof against multi-lateral deviation. If 50,000 people started speaking a different language, that would also be a Nash Equilibrium. In general, there may be hundreds or thousands of different Nash Equilibria. Knowing that an outcome is an equilibrium is not the same as knowing it will be this particular equilibrium outcome we arrive at in the real world.”

“Figuring out which equilibrium outcome we are likely to see requires us to think about how players might coordinate, what information they have, what they expect to happen in the future, how they expect other players to behave, and so on. This leads us to new definitions of equilibrium in games that are more rigorous and complete than Nash Equilibrium.”

This has great bearing on blockchain, particularly on the consensus and security models currently used. Information, expectations, and different notions of equilibrium are almost largely ignored in modern blockchain implementation, according to John.

“Game theory helps us understand and incorporate all the information that people have, their beliefs, strategies, and how they would react in equilibrium. Sometimes it gives us answers, but more often it saves us from errors.”

The big question: Game Theory and blockchain

Game theory is a complex set of rules and concepts that are sometimes hard to relate to everyday life.

I was intrigued as to how game theory relates to blockchain technology, and John was able to illustrate this to me.

“Let’s take Bitcoin’s Proof-of-Work concept. Let’s say that we have 10,000 nodes that are validating transactions. How are they ever going to coordinate together and do something that’s dishonest? The hope is that the wide distribution of the validating network gives Bitcoin an inherent stability. At least half would have to be dishonest for the system to break, but the notion that 5,000 nodes would somehow coordinate to do something dishonest seems ridiculous.

“Let’s think for a minute about how protocol might be broken without coordination. Consider the following scenario. When a Bitcoin node, also called a miner, wins the right to propose the next block in the chain, it is allowed to create 6.25 new Bitcoins for itself as a reward. “Mining” these new Bitcoins is what gives the nodes an incentive to validate transactions and write new blocks. But suppose that if, instead of creating the 6.25 as well as the Bitcoins that the protocols allows, a node decided to create 50 new Bitcoins as reward. This is a clear violation of the protocol and blatantly illegal. But, the action itself does not steal tokens from anybody, and it doesn’t violate any other rules. The node simply decided to give itself a raise.”

What would other nodes and users do? Would they walk away from their bitcoins, or would they say, well, if all other nodes accepted this new block, why not? Other miners might think this is a great idea and start taking a mining reward of 50 Bitcoins instead of the 6.25. This leads to a Nash Equilibrium situation where if everyone decided to accept that 50 Bitcoins is a proper mining reward, no single miner would have an incentive to deviate. No single account holder would benefit from walking away. This is an example of an equilibrium that depends on the beliefs of how other agents will behave in a new situation.

Of course, this is just a single example among an almost infinite complexity of possible situations in the blockchain, so the implications of fault tolerance are enormous.

On childhood, foolish adults, and lifelong friendships

Who we are today is in many instances shaped by experiences and events from our childhood. Everyone’s past is different, but what a person goes through in their early years of life has a great bearing later in their timeline.

John opened up about his rebellious youth spent in the southeast of San Francisco, back in the 1960s and 1970s.

“I was rebellious, I was convinced I knew everything, and that adults knew nothing. I was totally certain about that.”

Then he said something that doesn’t quite tally with the man I’m speaking to today:

I was probably not a very nice kid. I would not have enjoyed myself as a child, I have to say. It was an unusual, situation because, in fact, I saw that a lot of adults were acting childishly. It was a time when people would not work but instead live for the day, and think about nothing but themselves, and somehow justify it as a philosophical journey they were taking to discover who they were. Their kids, their families, and ultimately, they themselves, suffered the consequences.”

“I guess that, intuitively, I thought that they were irresponsible and were not setting a good example. In reality, it likely would not have mattered. I was not a very nice person and I probably would have found flaws in anyone.”

Though John accepts the difficulties of his upbringing, he emphasizes that he picked up a lot of friends along the way, some of whom are currently working with him in his latest project.

“I met Lun-Shin Yuen when he was 12 years old. He was a Chinese immigrant. His father, who was the Taiwanese Ambassador to Turkey at one point, had recently passed away. He was a smart kid too, so we both went to Lowell, the public academic high school in San Francisco. We played Dungeons and Dragons together, which was what nerdy kids did at that point before the internet existed.”

“He went on to get a Masters in Computer Science from Stanford University, was the 13th employee at Intuit, and has been in eight startups, among other things. I called him when I started the Geeq project, and he’s now our Chief Architect.

“Another guy on our team, Eric Ball, was the Treasurer for Oracle for a long time. I met him in Grad School and he’s now looking after our monetary affairs and helping us make contacts in Silicon Valley.”

“It’s been quite surprising how many of my old friends have gotten involved in the Geeq project, and how much they bring to the table. It has also been a lot of fun.”

On cooking

We found common ground through the love of cooking. Discussing a good meal is always a good way to take a break from the heaviest topics, so we did just that. It turns out that John’s specialties range from Indian dishes to more refined Spanish cuisine like paella.

John then confessed that his culinary skills would be put to the test later that day, as his daughter was bringing eight of her friends to dinner. What’s on the menu? I asked. “Fish tacos.” I could almost smell the spices and cooking oil, and hear the praising words of John’s daughter and friends gathered around the table.

On blockchain

Blockchain technology is gaining further traction and adoption with every passing day. I was curious to know what drove John to it.

“I spent a sabbatical at Microsoft Research Building 99 in Seattle. This site houses around 500 Ph. Ds, mostly in Computer Science. Here, I could learn anything I wanted to. Cryptography, WiFi policies and protocols, Biometrics, you name it. An expert on almost any field would only be a few feet away.

“There, I started doing some work in telecommunications protocols, because these are basically games. Bandwidth is a scarce resource. If I am using a WiFi channel, then no one else can until I finish. How do we share bandwidth in a decentralized way where no one is in charge? This is what the 802.11 family of protocols does. Could we do this more efficiently? What if we had a centralized controller of bandwidth use? (This is what cell phones do with the LTE protocols.) What if we want to combine both protocols on a single WiFi channel. I know it sounds boring, but it affects everyone with a cell phone or a WiFi connection. The stakes are huge, so getting the game theory right is important”.

“I also became involved in a biometrics project to identify and aid refugees. Often refugees would have nothing, no documents, no possessions, only themselves. The United Nations High Commissioner for Refugees (UNHCR) uses biometrics such as iris scans or fingerprinting to provide relief, give continuity of medical care, and to give refugees an identity. The concern is that this data could be abused if it fell into the wrong hands and the use of biometrics creates significant privacy issues in general.”

Trusting intermediaries like Amazon, Microsoft, or even the Government, is a concern since these entities may not always be very trustworthy. Blockchain offers the potential to take control of information away from centralized authorities and return it to the people who own the data. It can create decentralized methods of interacting and cooperating that don’t require the permission of a central authority. Again, one has to get it right. There are a lot of details, especially in the consensus protocols. I love this kind of problem.”

On Geeq.io

John’s latest project is Geeq.io, what he refers to as an ‘infrastructure blockchain project”. It’s meant to be “Blockchain-as-a-Service.”

Geeq’s most distinguishing feature is the security guarantee, through this approach we call Proof-of-Honesty. We were able to recast the problem to get around the impossibility theorems, besides solving some of the game theoretic issues that were being ignored.

“The security comes down to something called coalition-proof equilibrium. In effective terms, this means that any coalition of nodes, even all the nodes potentially working dishonestly together, makes more profit from d behaving honestly than anything else they might think of. We have also added elements to Geeq’s protocol that allow it to recover even if all the validating nodes simply want to destroy the blockchain and don’t care about profit.

“If you want to trust your society to a blockchain, 51% Byzantine Fault Tolerance is not enough to guarantee the security of stock exchanges, bank accounts, medical records, supply lines, or electrical grids. If blockchain starts to do these things it is easy to imagine that hackers, hostile nations, criminals, or people simply wanting to create chaos, would try to bring the system down. Geeq is a blockchain that is immune to attacks like that.”

Time for a curve-ball

I have learned that John loves a good challenge so the machinations of my mind concoct the following question: AI. A blessing or a curse?

“AI is inevitable, there’s no other way of thinking about it. It is coming, and a lot of aspects are very frightening. Soon, every item in your house will be connected. Your toaster will have a microphone and a video camera, and you will never know if it’s on or off. The danger of AI is that it gives a single person, or a single entity, an enormous amount of leverage.

“So blessing or curse, it depends on your point of view and how it is used. Blockchain can be an element that protects us from some of this. Blockchain can be used to limit what any information system does and make its actions visible and auditable. I hope that we build some of these protections in before AI is fully released into the wild.”

Coda

Game Theory, Economics, and Fish Tacos. I discussed all these things and more with John Conley during our enjoyable conversation, which left me thinking about all those little decisions we take every hour of every day, and how there might be a mathematical mechanism governing it all. Pondering whether to have a Latte or an Americano never had so much depth and possible repercussions.

Still, a well-spoken, dignified gentleman, John would be the perfect companion for a good post-dinner discussion on the vicissitudes of life on the blockchain and beyond.

So let’s see. Latte or Americano. Hmmm. Latte today. I can hear the butterflies flapping in Tokyo…

You can follow John Conley through Geeq’s website, or his own LinkedIn profile.

Share.

About Author

Smart, dexterous. Tactile, intelligent, educated, and well-read. A little impatient, wholly intolerant to bad spelling and lazy grammar. Writer, poet, and lover of those who love the side that we don't see. Ireland-based, but my mind wanders the world. Adventurous, and non-smoker. Photographer of all that is worth being portrayed for its intrinsic beauty. Widowed father of two beautiful young creatures. There is a destiny, not a destination, in my life. There is an admiration for she who may be gone, but not forgotten. Through words we live, and through words we are remembered. Join me in my journey...

Leave A Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.