We are drowning in a sea of misinformation. A new type of blockchain-based system, called token-curated registries might help humanity pass through this post-truth age. In this age, information is so cheaply made and transmitted that noise drowns out meaningful signals. Google is a mirage, causing us to eat sand: even if we don’t click on the sponsored links, many of the seemingly objective articles have their own agenda. This is inherently a problem of centralization: by massaging media creators and optimizing for search engine algorithms, marketers are able to direct us to ideas and products that aren’t necessarily the best.
Token-curated registries (TCRs) align a community around building a quality list. Members can use their tokens to vote on whether a submission is added to the list; earning additional tokens if their choice coincides with the majority’s decision. The end result is a semi-permeable membrane; making it difficult for bad submissions to be listed.
In April of 2018, the first implementation of a TCR system was launched on the Ethereum mainnet. adChain is a list of websites that are deemed by the token-holders as being a legitimate place for advertisers. I spoke with the project lead, Mike Goldin, at the third Ethereal Summit.
Interview with Mike Goldin - project lead
Tristan: In your terms, what is a token-curated registry?
Mike: They’re means for creating lists in a decentralized way, with intrinsic economic incentives for the people doing the curation, to do it well. So they’re like closed loop systems that create lists.
Tristan: What problem is the first TCR, adChain, trying to solve?
Mike: Sure! So, programmatic ad tech is about a $200 billion a year industry. […] The industry's own lobbying group, [estimates that] ten percent [is lost to fraud]. So the industry's own lobbying group is saying $20 billion. [The problem is:] what is a human versus what is a bot on a webpage?
And so what you do if you're in a country that is never going to extradite you for something so silly as bot fraud in the United States... Just pin up a website; Coca Cola will pay you a penny to serve an ad, it costs you half a penny to buy a bot to view that ad, and you’re just going to do that all night long. And the only way to detect that this [...] detective work and mathematics, where people look at statistical patterns and the impressions and try to figure out which of the suppliers are fraudulent.
So, it’s an incentive problem in ad tech. What adChain does is it introduces these third party token holders, [whose] only interest is in curating a list of high quality, generating this white list and then building on top of that […] tools for advertisers and publishers to talk to one another in a fully peer-to-peer way. So basically the registry is on chain for peer discovery, and [the goal is] fully peer-to-peer ad commerce.
Tristan: So, MetaX had this idea called ‘adChain’ to curate a list but they didn’t have the implementation?
Mike: Yeah, so at the outset, [when] we started on the project in December of 2016, we knew that we wanted to solve fraud in advertising technology; a $20 billion dollar a year problem. Didn't know how we were going to do it, and we spent four months [...] designing a system which [was] fundamentally flawed. […] I was very depressed. […] I was about to go on vacation when we decided to throw all that work away.
But then my former coworker Ameen, who left to do SpankChain ... had an early copy of the white paper from the Aventis project [which] had this mechanism [...] for a state challenge mechanism, and [this] was massaged into TCRs.
Tristan: The system has been running for a couple weeks, and recently an application to whitelist Facebook on adChain’s registry was rejected. Why do you think this happened?
Mike: Great question. So this was such an interesting thing to witness and if the TCR idea takes takes off, it will be historic. And I think it's interesting because the kind of basic purpose [...] of adChain is to create this registry of non-fraudulent publishers. Now there is certainly some evidence that there is a lot of fraud on Facebook, but what I think that vote was about, was about values other than the primary [focus of AdChain].
At the time Facebook [was] in the news a lot for being a creepy company. Ad token holders are sovereign, they can [vote with their values - even if it goes against the stated purpose of the registry.] [They valued preventing] what they view as a creepy company from participating in this ecosystem [more than the] commerce that Facebook would have brought in.
Tristan: So, how was this split, how was the vote split? Was it a little bit more contentious or --
Mike: So, no, to be realistic and transparent, this is the early days of adChain and not many people are voting yet. That particular vote was overwhelmingly against Facebook. Now we've also had some votes that I was like personally unhappy with - the result of Newyorktimes.com was challenged, and the person who brought the challenge said that [...] because the New York Times published stories that WMDs in Iraq and [helped coax] the United States into the Iraq war and they should not be in the registry. I respect their subjectivity but it's not my own [...] So I've been careful with how many tokens I vote with because I want to see how other people are participating and I don’t necessarily want to like scare people off [by outvoting others with less coins to stake as group norms are built]. But New York Times, they actually lost that [challenge], which like is offensive to me as an adToken holder. So, now I'm personally going in much harder with my own tokens. But it's interesting to see the curation process play out, to be a witness to it.
Tristan: With AdChain, at what points do on-chain and off-chain data cross?
Mike: Yeah, so in deciding how to interact with the registry, whether that’s should I buy this domain, or should I challenge this domain, which side should I vote on. Hopefully, you’re doing like Off-Chain consensus-making processes with other adToken holders would be like to provide before too long is, someone could say, If I did this, would you do that? And just have like an Off-Chain polling system, you get assigned a message, provide like signals as to how you’re going to respond to certain actions. Let people get some information about what token holders would do before actually putting their money on the line.
Tristan: So, right now there is no KYC?
Mike: There’s no KYC to participate in the curation process.
Tristan: What other TCR projects are you watching?
Mike: Yeah, so there's so much stuff that happens in TCR world now that I can’t keep on top of. So, I’m going to use this question to shield my own projects. Another project I work on is called the Delphi, so I’m working with Mark Beylin from the Bounties Network on that and -- well there aren’t, you know Bounties Network at glance whatever. You put down some bounty, paint my house. You come and paint my house and then I say, thank you very much, I’m not going to pay you. This is the problem on all of the existing bounty systems.
So there are a few projects that I’m working on, On-Chain Dispute Adjudication Services. The approach we're taking with Delphi, part of the approach is, have a token periodic registry of you know, credible arbitrators. The system for like these arbitrators making-- for doing adjudications… This TCR could just be like a focal point for people selecting, okay, I want this person to be my arbitrator, I want this person to be my arbitrator. But another thing you can do is say, I'm not even going to pick a specific arbitrator, I'm just going to -- anyone in the TCR can vote on the outcomes of the disputes and kind of like let people you know, make like an easy decision like you know, how disputes will be arbitrated with that to a lot of research.
Tristan: What are some attack vectors for TCR’s generally?
Mike: So, good question. So, there's the majority validator attack […] It's an extremely expensive attack; it's easy to recover from [by forking the registry].
[More] interesting are various bribing attacks that can occur [like] “P + epsilon” [in which the] attacker [wants] a poll [to go] a certain way, kind of irregardless if it’s good for the registry. The smart contract says my vote for the option that I want. If the option that I want you to vote for wins, if it wins the poll, you’re just going to get your normal voting rewards, but if it loses, you're going to get -- I'm going to pay you -- the contract is going to pay you, it’s normal voting reward. The normal voting reward plus epsilon, plus some additional amount...
If you're a relatively disinterested yet rational actor, its logical for you to participate in this type of bribing attack. […] I kind of feel like we will be in 2018 where the decentralized exchanges were in 2017 [with a couple problems to still solve].
There are theoretical paths to solving P+ epsilon, we are not there yet and it's another thing where we'll see what happens in practice. Maybe it's just not something that will happen in practice, I don’t know, we are in the wild right now.
Tristan: What is your background and how did you get into TCRs?
Mike: I got into crypto totally by accident. I'd been an anthropology student at Bennington college; I winded up transferring to Columbia where I did a Computer Science degree the summer after my junior year. I was looking for summer jobs, I was applying for a job with the NSA but I failed my polygraph because I was smoking a lot of weed and at that time I thought I had a patriotic duty to conceal that information.
Tristan: They found out?
Mike: I’m a terrible liar so I failed the polygraph. […] Summer was about to start and I didn’t have a job. I found this random company called ConsenSys; this was back in 2015. It was like 15 people, I walked in. It wasn't a very rigorous interview process at all and they took me on as an intern, even though I didn't have experience at all with blockchain before that. But I became fascinated because blockchains are very much in the intersection of computer science, economics, and philosophy to some extent.
[…] I was in the enterprise team. [...] I got put on this project called AdChain and -- AdChain was like the first TCR and if you read the AdChain paper, it’s not exactly a TCR but you would recognize that that's like the primordial thing that turned into TCRs. As we were [developing AdChain] -- we had this inkling that maybe that system was generic and would apply to other things as well, we had an inkling of it.
What was interesting is that it was actually in the course of developing it, turning the idea into code, we discovered a few optimizations that would make like the code nicer, that would also make the concept much cleaner, and that summer was when I realized like oh shit, this is [generic enough to] carry any list and that’s when I wrote the TCR paper.
Tristan: How are the developers and investors rewarded for this venture?
Mike: So, there's a billion Ad token, 500 million were sold to public, 100 million were allocated in private arrangements and 200 million are time-locked for MetaX, 200 million are time-locked for ConsenSys. So, ConsenSys and MetaX each have an upside in seeing the AdChain ecosystem become adopted.
Tristan: If TCRs do proliferate, what impact do you see that having on society?
Mike: So, I'll say this, real talk. [...] I am not and I don't think I have ever been one to blindly say the blockchains are going to save the world. I think there are a lot of potentially social and negative outcomes that could come from blockchains. And like -- it's kind of like I'm here, I had like the doom or the luck of being here, and if I walk away, it's not going to stop, it will just be people who aren’t me-involved. I don’t know what TCRs are going to do in the world. And I'm lucky to be in a position where like you know, a lot of people want to work with me and I choose to say no to any project that you know, clearly or evidently would lead to bad outcomes. I don’t know, you can come up with the utopian vision for what the world looks like but, also a dystopian vision.
So, what I'll say on that front is, I make no guarantees that what I’m doing will be socially-positive. It could be, but I wouldn’t guarantee it.
* * *
There is some irony here. TCRs may eventually make advertising more difficult, by directing our attention to good content without the need for journalists and social media algorithms. However, the first implementation is intended to help those who seek to pay for our attention.
A meetup on TCRs was held during Blockchain Week 2018. What had intended to be a couple dozen people sharing pizza and beer quickly became an overflowing mass of people excited about this new technology. It seems like TCRs, or at least technology similar, will be used to rank everything under the sun. They may not provide objective truth, but it is a step closer than the centralized systems provide.
For some criticism of TCRs generally, check out this Medium article.