Cryptoeconomics for beginners and experts alike

This past week Koinify and the Cryptocurrency Research Group (CCRG), a new academic organization, held a 3-day event — the first of its kind called Cryptoeconomicon, an interdisciplinary private event that included a cross section of developers, entrepreneurs, academics and a few investors.  It was purposefully scheduled to coincide with O’Reilly Media’s own “Bitcoin and the Blockchain” conference which took place in the middle of it.

I attended what amounted to four days of seminars, brainstorming and networking sessions.  Below are my summarized thoughts.  Note: these are my opinions alone and do not reflect those of other participants or the companies I work with.  You can view pictures/info of the event: #cryptoecon and @cryptoecon

Rather than going through each session, I will just highlight a few areas that stood out to me and include outside relevant content.

What is cryptoeconomics?

According to Vlad Zamfir, of the Ethereum project, cryptoeconomics as a field might be defined as:

A formal discipline that studies protocols that govern the production, distribution and consumption of goods and services in a decentralized digital economy.  Cryptoeconomics is a practical science that focuses on the design and characterization of these protocols.

Zamfir discussed this at length (slides) (video) and rather than going too in-depth with what he said I wanted to reiterate his main points he gave:

Cryptoeconomic security as information security

  • Mechanisms are really programs
  • They can distribute payoffs
  • The programs have a certain behaviour in the Nash equilibrium case
  • The NE has a cryptoeconomic security
  • We can be assured that a program will run a particular way

He also argues that “cryptoeconomics” should be see as more economics for cryptography rather than cryptography for economics:

  • Economic mechanisms can give guarantees that a program will run in a particular way that cryptography alone can’t provide.
  • Incentives are forward facing, cryptography is a function of already-existing information
  • How do we provide custom cryptoeconomic guarantees?

The last part in relation to his talk that really stuck out to me was on the final day.  In his view (slides) the technical term that should be applied is, “distributed cryptoeconomic consensus” which would assuage concerns from the academic “distributed consensus” community that uses different terminology.  Under this definition, this means:

  • A cryptoeconomic mechanism with the Nash equilibrium of assuring distributed byzantine fault tolerant consensus
  • We should be able to assert and prove the cryptoeconomic assurances of any consensus mechanism
  • Distributed consensus mechanisms can create a pure cryptoeconomy. Even the execution of the mechanisms is has a measurable assurance.

Most interesting comment of the event

I think the most apt comment from the economics discussion came from Steve Waldman, a software developer and trader over at Interfluidity on the first day of the event.

While there will likely be a recording posted on Youtube (video), in essence what he said was that in the blockchain space — and specifically the developers in the room — they are creating an enormous amount of supply without looking to see what the corresponding demand is.  That is to say, there is effectively a supply glut of “blockchain tech” in part because few people are asking whether or not this tech actually has any practical consumer demand.  Where are the on-the-ground consumer behavior surveys and reports?

Again, if Bitcoin (the overall concept) is viewed as an economy, country or even a startup, it is imperative that the first question is resolved: what is the market need?  Who are the intended consumers?  So far, despite lots of attention and interest, there has been very little adoption related to blockchains in general.  Perhaps this will change, maybe it is only a temporary mismatch.  Maybe it these are the chicken-egg equivalent to computing languages like Ruby or PHP and eventually supply somehow creates the demand?  Or maybe it suffers from the Kevin Costner platform trap (e.g,. if you build it, will they come?).

To illustrate this contrarian view:

why startups fail

Source: David Norris https://twitter.com/norrisnode/status/561262588466839553

Maybe there is no real market need for these first generation concepts?  Perhaps the network will run out of block rewards (cash incentives) to the miners before these blockchains can gain mainstream traction?  Maybe the current developers are not quite right for the job?

Or maybe, blockchains such as Bitcoin simply get outcompeted in the overall marketplace.  For instance, there are currently 1,586 Payment startups listed on AngelList and 106 P2P Money Transfer startups listed on AngelList.  Most of these will likely burn out of capital and cease to exist, but there are probably at least a dozen or so of each that will (and have) gained traction and are direct competitors to these first generation blockchains.

Perhaps this will change, but then again, maybe the market is more interested in what William Mougayar (who unfortunately was not part of the event) pointed out a few days ago.  Simply put, maybe there is more room to grow in the “Blockchain Neutral Smart Services” and “Non-Blockchain Consensus” quadrants:

Crypto_Tech

We cannot know for certain a priori what market participants will decide.  Perhaps Bitcoin is good enough to do everything its enthusiastic supporter claim it can.

Or maybe, as Patrick Collison, CEO of Stripe, wittily stated in Technology Review:

“Bitcoin is kind of a financial Rorschach test; everyone projects their desired monetary future onto it.”

Now, to be fair, Collison (who was not part of the event) has a horse in the race with Stellar.  Fortunately there was not much emphasis on token prices going to the moon at the Cryptoecon event.  When incentives did come up, it was largely related to how a consensus mechanism can be secure through a self-reinforcing Nash equilibrium.

Perhaps a future event could discuss what Meher Roy (who unfortunately was not in attendance either) adroitly summarized and modeled in relation to how actors are betting on crypto-finance platforms:

meher roy table

Source: https://medium.com/@Meher/a-model-to-makes-sense-of-beliefs-and-associated-crypto-finance-platforms-f761a7d782cb

Back to the show

There were a number of startups at the event, probably around a dozen or so.  In my view, the most concise overview was from Sergey Nazarov co-founder of SmartContract.  The interface was clean, the message was clear and “issuance” can be done today.  I’m not necessarily endorsing the stack he’s using, but I think he has clearly talked to end-users for ease of use feedback (note: be sure to consult a lawyer before using any ‘smart contracting’ system, perhaps they are not recognized as actual “contracts” in your jurisdiction).  Also, drones.

It would have been nice to see a little longer debate between StorJ, Maidsafe and Filecoin groups.  I think there was probably a little too much “it just works” handwaving but thought that Juan Binet-Betez from IPFS/Filecoin gave the most thorough blueprint of how his system worked (he also showed a small working demo).

It was not recorded but I think messaging for Augur (a variation of Truthcoin) was pretty poor.  Again, just my opinion but I was vocal about the particular use-case (gambling) proposed as it would simply bring more negative PR to a space smashed with bad PR.  The following day other members of the team discussed other uses including prediction markets for political events (similar to what Intrade did).  I am skeptical that in its current form it will become widely adopted because futures markets, like the CME, already do a relatively competitive job at providing this service for many industries and these decentralized markets could likely just attract marginal, illicit activities as has been the trend so far.  I could be wrong and perhaps they will flourish in emerging markets for those without access to the CME-like institutions.

Things that look less skeptical

  • There were about 10-12 people affiliated with Ethereum at the event, all of them were developers and none of them seemed to push their product as “the one chain to rule them all” (in fact, there was a healthy debate about proof-of-stake / proof-of-work within their contingent).  I’ve been fairly skeptical since last summer when their team looked gigantically bloated (too many cooks in the kitchen) but they seem to have since slimmed down, removing some of the pumpers and focusing on the core tech.  This is not to say they will succeed, but I am slightly less skeptical than I was 3-4 months ago.
  • I also had a chance to sit down with a couple members of the IBM ADEPT ‘Internet of Things’ team.  They held a ~3 hour workshop which was attended by around 20 people.  The session was led by Henning Diedrich (IBM), David Kravitz (IBM) and Patrick Deegan (Open Mustard Seed Project).  Again, even though I’ve paged through the ADEPT whitepaper, I was hesitant to believe that this was little more than marketing on the part of IBM.  But by the time the session was over, I was a little less skeptical.  Perhaps in the future, when more appliances and devices have secure proplets, they could use a method — such as a blockchain/cryptoledger — to securely bid/ask on resources like electricity.  B2B and machine-to-machine ideas were discussed and piggybacked on.  Obviously there are all sorts of funny and sad ways this could end but that is up for Michael Bay to visualize next year.
  • This also intersects with another good comment from Stefan Thomas (CTO of Ripple Labs).  In a nutshell, on a panel during the first day, he thinks there is some confusion and conflation of the terms “automation,” “decentralization,” “smart contracts” and “blockchains.”  That is to say, while blockchains are automated, that is not to mean that it is the only means to achieve automation.  Nor is decentralization necessary for automation to be achieved in every use-case.  Nor are smart contracts the only way to control automated devices.  When the video is posted I’ll be sure to link it (video).
  • Ethan Buchman, lead dev for Eris, was both witty and on top of his form, noting that in practice users don’t need a new browser every time they go to a new site, so they shouldn’t need a new client to view a different blockchain.  Let’s keep our eye on Decerver to see how this germinates.
  • Lastly, the two investors that attended the VC panel on Wednesday included Shahin Farshchi from Lux Capital and Pearl Chan of Omidyar Network.  What I liked about them is they weren’t pushing a certain binary viewpoint.  They were both upfront and honest: neither had invested in this space, not because they hated it, but because they were taking their time to see what opportunities actually fit within their mandate.  Perhaps they will at some point.  One joke that Farshchi mentioned was that back when cellular telephony was growing, “everyone and their mom” was selling base station equipment and chips.  Similarly there were over 300 companies creating thin film solar cells before bankruptcies and mergers.  So the type of euphoria we see in the Bitcoin-space is not necessarily unique.

Room for improvement

Perhaps if there is a next event it could include representatives from Blockstream, Bitfury and other Bitcoin-centered projects.  It would be nice to have some perspective from those deeply concerned about with maintaining secure consensus and the Blockstream team has some of the most experienced engineers in this space.  Hearing their views next to what Peter Todd (who attended and had some interesting calculations for the estimated costs to attack a network), could help developers build better tools.  Similarly, developers from Peernova, Square, Stripe, M-Pesa and Western Union would also likely be good resources to provide empirical feedback.

Additional clarity for what a decentralized autonomous organization (DAO) actually is and is not could be spelled out as well.  And how do these intersect with existing legal jurisprudence (can they? as Brett Scott might ask).   For anyone who has read “The Cookie Monster” by Vernor Vinge, both Matt Liston and Vitalik Buterin made some not-entirely-unreasonable points about machine-rights and whether or not machines should trust humans (e.g., humans expect bots to provide truthful information, but can the reverse be expected?  And what happens if a bot, like a DAO, is deemed too successful or broke a law in some jurisdiction — does it get “carted” away in a truck?).

Lastly, I think by the time there is another event, there will hopefully be more clarity for what a “smart contract” is.  One panel I moderated, I tried to get the participants to use the word “banana” instead because the term “banana” is overused and often conflated to mean many things it is legally not.  Primavera De Filippi from the Cryptolaw panel made some good comments too about whether or not “bananas” are actual legally binding contracts; she previously did a workshop with Aaron Wright (also in attendance) at the recent Distributed Networks and the Law event held at Harvard/MIT.  Steve Omohundro also spoke realistically about these scenarios on the final day, where does liability start and stop for developers of DAOs?

[Note: I would like to thank Kieren James-Lubin, Vitalik Buterin, Tom Ding, Sri Sriram for organizing the event, Robert Schwentker for acting as emcee/photographer, and CFLD and Omidyar Network for sponsoring the event including the delicious food.]

7 thoughts on “Cryptoeconomics for beginners and experts alike

  1. Thanks for the nice and balanced review, Tim.

    Some comments of my own:

    1. I don’t think it’s fair to accuse everyone in the space of not spending enough time analyzing consumer interest and use cases. There are many levels of startups in the sector, and some are necessarily more abstract. As an extreme example, I see ethereum as being similar to nodejs or c++: we are so abstract and generic that we do not really need to care about what consumers want, we need to care about what developers want. Prediction markets I would say are similar. They’re a tool that’s potentially valuable in many places, and it’s actually better to build it out as abstract infrastructure first, and then figure out what can use them later (note emphasis on “what”; I think the primary customers of many dapps will be other dapps). But as far as more close-to-end-user stuff like IoT is concerned, I would agree.

    2. “Cryptocurrency maximalism” I think isn’t quite the correct way of phrasing what non-bitcoin cryptocurrency projects are trying to do. We’re not even trying all that hard to push the cryptocurrency angle, we just like the idea of people being able to use digital services inside of pure cryptoeconomic ecosystems. Now I do think there’s a line between ripple/stellar and ethereum/tendermint etc, but it’s a rather subtle one; I perfer describing it as “objective” vs “subjective” or “pure cryptoeconomic” vs “hybrid”. One is targeting the idea of services that try to be cryptoeconomically pure in some sense, living on a sort of ethereal layer without any dependencies on any kind of existing social trust networks, and the other tries to integrate more. Particularly, note that the belief that pure cryptoeconomic systems deserve to exist is NOT the same thing as a belief that they should be the primary or only digital infrastructure of society; Robert Sams might be a great example here, advocating that pure cryptoeconomic systems should stay pure but supporting hybrid projects at the same time.

    3. What I particularly liked is actually the growing awareness of a divide between the “pure cryptoeconomic” and “hybrid” as one of the primary differentiating factors between projects in the space and a debate between the two approaches (eg. see the identity and reputation panel). I hope to see more analysis of the exact nature of the distinction and strengths and weaknesses of each approach in future discussions.

    4. Having more people from blockstream would definitely have helped, esp on the technical panels. Having not enough PoW boosters makes for a boring circlejerk, even if it is one I agree with 🙂 If you are reading this and are from blockstream, feel free to contact us so we can include you next time around.

    5. As far as formal cryptoeconomics goes, one thing I would like to see is a formalization between the different forms of security assumptions; for example, I see a strong difference between “51% are altruistic” (DPOS / naive PoS), “51% will not collude” (Bitcoin) and “51% will not collude and destroy many of their own currency units in the process” (Tendermint/Vlad-style). Additionally, failure modes are important; Bitcoin’s “miner with 51% can revert a bit of history” is substantially different from Blockstream’s “miner with 51% of a sidechain can steal everything from that sidechain”.

    6. I now think Ethereum popularizing the term “contract” so much was a mistake. Something like “agent” or “program” would have been better, as it emphasis the fact that it’s a sort of sui generis beast that just automatically passes messages and money around, not something meant to be a perfect parallel to legal contracts.

  2. @Vitalik: You’ve nailed it with the second comment.

    I appreciate this distinction. Thing is, my target audience is readers attempting to bet on the future primary financial infrastructure of society.

    It is in this target audience that the distinction is unclear.

    I wish I could build this distinction in that same blog, but it would have made it inelegant and long. Let me dedicate another blog to this distinction.

  3. The ADEPT platform has some very interesting ideas but in order to be successful needs to address quite a number of significant security and incentives questions. These sorts of IoT angles are really tricky and have been a major problem in that space pretty-much forever. In this particular case I believe they’ve got to solve some real questions on incentivization to secure their blockchains, while security is all too often a secondary matter for companies who want to create “smart” devices as they’re usually transitioning from non-connected devices and have very little comprehension of the security headaches that they’ll face.

    I’ve seen other, very large companies, attempt to make bets in this same space and all have floundered for similar reasons.

  4. Useful summary, thank you.

    I do get the sense this technology is 3 – 5 years from profitability in the western world, I also agree with the main thrust of the article that if you’re not solving a problem then you’ll lose money.

    Building tools for developers has proven to be a good strategy where the demand for those tools is high. Where the demand is unknown it’s a different question.

    I’d like to see a really good public debate about the cost / revenue model of commercialising a blockchain and / or distributed consensus and / or cryptoeconomic consensus capabilities, and where which technology applies best.

    In my mind it’s a 4 dimensional table with no clear winners in each category. The trick at the moment appears to be go generic, and get the developers and cross your fingers.

Leave a Reply

Your email address will not be published. Required fields are marked *