Spotting Economic Opportunity in an SSI World (3/3)

Spotting Economic Opportunity In An SSI World (3/3)

Session: 11I

Convener: Dave Huseby

Notes-taker(s): Scott Mace

Tags for the session - technology discussed/ideas considered:

Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:

Link to slide deck of presentation, provided by Dave Huseby:

http://docs.google.com/presentation/d/1uljRTBy8HqGSd8sfxUxMF1oAFswxzHYVsHu0AlPjc9s/edit?usp=sharing

Nine problems of distributed systems

• Discovery
• Introduction
• Coherence
• Public Service
• Trust
• Privacy
• Coordination
• Membership
• Persistent State

SSI can solve introduction, somewhat trust & privacy by combining with blockchains for the persistent state.

Spotting economic opportunity, where some of the nine problems are not solved.

Adhere to the principles of user sovereignty. Offer paid edge services that do not violate user’s trust.

My first victim today, how this model is applied to existing distributed systems and how they create economy opportunity.

Linus Torvalds relied on external systems like email to build Git. How do i find another developer to collaborate with? Failed to solve discovery. Introduction, no built-in PKI infrastructure. where I started the DID-GIT project for built-in PKI in a Git repo. Git relied on external solutions. Coherence, how do I reconnect in Git to submit my patches? They didn’t solve that either. The original design was to email patches. A royal PIA. Membership, no obvious way to do this in Git. Automatic governance is possible but not in Git out of the box. Coordination, Git is about tracking changes to files. Doesn’t have the other necessary doing collaboration for software development. GitHub was sold to GibHub for $7.5 billion. You don’t want a company like Microsoft owning the revision control for all these open source solutions. Microsoft is starting to take advantage of this. Merging with Azure services. You could just run Windows. Recapture developers as Windows users. Existential threat to Linux Foundation? Wouldn’t say it is, but wouldn’t say it isn’t.

Bitcoin is not decentralized. How can we harden it? Discovery and coherence is lacking. Code has a hard-coded list of IP addresses. Used to be done through IRC, at least more decentralized than IP addresses. Discovery, no mechanism for that. Bitcoin itself does not provide a mechanism. Very little privacy in bitcoin. Can trace. A compromise due to regulation. But wasn’t in the bitcoin design. Coordination, no built-in to transmit a secure bitcoin address to a sender. How do I know I’m talking to who I think I’m talking to? Misdirection of funds is a big problem. Authorized push payment fraud, you get the wrong address. Knowing they didn’t provide those solutions? Coinbase.com provides a centralized solution, also an easy way to send bitcoins to someone. Market share is big, like 50% of retail investors. Coinbase is bitcoin like AOL was the web back in the 90s in casual users Coinbase is now valued at $8 billion. Friends say this model predicts centralized solutions attacking decentralized tech.

Secure Scuttlebutt is not decentralized. Disc & coherence uses “pubs” at a specific DNS address to discover other users and get updates. That alone right there makes Secure Scuttlebutt subject to corporate capture. Supposed to be a distributed Twitter. Throw in a bit of Patreon, RSS. Build it up until it’s a threat to Twitter and sell it to Twitter for a billion dollars. It will happen.

Before I worked at Hyperledger, I worked at Mozilla tasked with looking at surveillance and privacy. Worked a lot with the Tor project. I didn’t know why the Web was constantly being recentralized. The Web originally had no discovery mechanism, so the search engine industry was inevitable. Originally no privacy, created the CA system. I want to know when Riley updates her blog, a rare case where the community came together. RSS. But Blogger.com was already making tons of money, launched the Atom protocol, corporate-owned, to have paid subscriptions for bloggers. Atom supports all that. This kind of coherence solution was hampered by RSS vs. Atom. Widespread adoption of RSS was slowed by that fight. RSS almost lost. Also no public service solution, Intermediary solutions like Blogger.com. And no persistent state. Money made by databases & middleware as a result.

Maybe dinosaurs need to start making oil. Whatever we do next needs to be designed to resist corporate capture from day one. We need to start from scratch. No way to fix the browser. Too much legacy that can attack our privacy.

Pick an existing legacy system. Say auto maintenance records. How would I get the maintenance records from the mechanic over to the dealer? Now that we know how SSI works, build companies that can exploit these. Make money teaching developers how to solve these problems. That’s a great idea for a business. Any questions?

Scott Mace: Content payment problem.

David: The toughest problem. Let’s put a pin in that.

Liam McCarty: Trend away from decentralization. Unless transaction costs are extremely low in a decentralized system, maybe we will have this trend toward centralization.

David: Hold PII is more and more risky.

Sam Smith: Transaction costs: triangulation, transfer and trust. Right now, much are hidden costs. A lot of the corporate takeover systems provide implicit transaction costs. We don’t have decentralized trust systems to do trustful transactions. FB & Google have OAuth and a sign in. They take care of all that trust exploit stuff, making sure nobody’s hacking their servers. When decentralized systems can solve the trust problem, then we can become competitive to centralized systems in terms of transaction costs.

David: Back to micropayment. One criticism I have of writers, happening in financial writing. You need to give away stuff so people will read more from you, then charge for things that have extra value. For instance, I could have a subscription service. I don’t think anybody is going to be paid to tweet and to write an opinion blog. There has to be this asymmetric value.

Scott: Wondered how this could do it?

David: Could do micropayments without incurring too large a transaction cost. That’s what Lightning Network is happening.

Robert Mitwicki…: Data monetization is a popular business model. Aggregate & sell data to maniupulate behavior. Would we get rid of those business models? Create a better next version of the Internet or still stuck with old business models?

David: Won’t end that business model. Will make it respectful. User aggregation is a concierge servcie. Sign up with edge service that I feed them all the info about it, they become the intermediary. Show me all the Ford Truck deals.

Robert: Do we need to solve all 9 problems, or no matter if we solve them or not?

David: The surveillance economy will be here as long as we haven’t solved these problems.

Johnnyfromcanada: FB & Google kill companies arbitrarily when they change their policies. His point, once we build this internet computer, you’re coding against one computer. Just the problem you’re trying to solve. How we just moved the goalposts. “Definity” in the chat.

David: Metastability, since you’re always able to connect to at least one other node, you only need 10-15 nodes in a cluster, you get an island of stability in a world where people are frequently disconnected. Protocol called Last Known Wherabouts. Makes a statistical connection. You could discover the whole network. Even though none of us have a central server anywhere. Blog post by Constantine about a fictitious collaborative tool for software development. He’s talking about a fully decentralized solution.

Vic Cooper: Is telecommunications (phone networks) an example of a distributed system?

David: Original system was decentralized. Modern telecomm is IP based with central control. Say I want a radio not tied to a provider. The FCC is about to open up the 6GHz spectrum. A lot of nonintereference rules will allow big commercial providers to squat on public property. A fully decentralized solution would be a ubiquitous fabric. Let’s start building our own handsets.

Marc Davis: In a truly user-sovereign data economy, it’s not just individuals but collectives that want to counter data aggregators.

Scott: Milton Pedraza is working on third-party representation of the rich to people with whom they want to do business. He was at Monday’s Me2B Alliance day.

Sam Smith: You share stuff and they make a legal receipt that says they accepted that data, you can watermark it. They’re liable for it. Differential privacy, the inverse is differential watermarking. Run an algorithm on it. Prove they distributed your data set without your permission.

David: That’s captured in the informed consent principle, fees into trust & privacy.

Sam: Liability is a stronger word, a weapon against surveillance.

MarkL (smartopian): Trust means assurance, means liability, which is what Sam is talking about. Decentralized has to have this concept it’s outside a system. We’ve been working on that for a long time with notice and consent standards. Notice just made it to ISO last month. Most companies pretend to have consent records, but they actually don’t. That’s been coming for a long time out of IIW.

David: I appreciate the feedback. This is just a work in progress. I’m the security maven at Hyperledger, plotting our way forward. A lot of navel-gaziong, should we host the code repos? How do we know who has contributed to allow them to vote in our elections? A lot of foundational problems we’re looking for better solutions to. Home to a lot of SSI stuff. Brian Behlendorf wants to make a lot of our solutions more robust, private, decentralized, eat our own dog food. Also to predict centralization. It’s a useful tool. Appreciate any more thoughts.

Zoom chat from this session follows:

12:30:35 From mitfik : TDA (Trusted DIgital Assistant) is a new browser :)

12:30:47 From rileyhughes : Is this Robert? ^^

12:30:54 From Michael Shea : of course

12:30:55 From mitfik : sure thing :)

12:31:04 From rileyhughes : ðŸ‘

12:42:13 From dsearls : Hate to say I'm going to be going in and out of this one.

12:45:56 From Robert Mitwicki *THCF* : They follow unix philosopy

12:47:27 From KathrynHarrison : Is there any way to make the same order of magnitude of profit (as a centralized system) through decentralized systems? Do the incentives exist to solve these problems in a decentralized way (esp when there is so much money to be made the other way)? Sorry to be the capitalist but...

12:48:57 From Robert Mitwicki *THCF* : we don't care much about technologies which are web related as we want to get rid of them ;) Let M$ take them.

12:51:27 From johnnyfromcanada : Dominic Williams (of Dfinity fame) is _very_ opinionated about _eliminating_ the need for centralized cloud platform providers (like Amazon, Microsoft, Google, Facebook), via an Internet Computer (which they are building). http://dfinity.org

12:52:52 From johnnyfromcanada : In the DLT world, Bitcoin is increasingly used as the counter-example to the principles of decentralization.

12:53:10 From KathrynHarrison : Coinbase’s marketshare is mostly US though. Not nearly as global

12:53:39 From KathrynHarrison : (Which just emphasizes your point)

12:54:09 From KathrynHarrison : So then why are decentralized solutions so frequently built with so many holes??

12:54:25 From Benedikt Olek : I think when dfinity succeeds and becomes very dominant, it potentially starts to threaten users sovereignty. Or other players enter, filling the gaps and becoming this threat - if I follow Davids argument correctly

12:56:37 From rileyhughes : FWIW, Greg Kidd was one of the first investors into Coinbase & blockchain.com. Coinbase uses a custody model and blockchain.com uses a self-custody model. Apparently, they each have similar market share globally

12:57:09 From Liam McCarty : Is it possible that there’s a “Coase’s theorem†of centralization? Coase’s theorem says a firm forms when transactions costs are higher in the market than in the firm. Maybe the idea of full decentralization in the cases David’s describing is as untenable as everyone in the market being an independent contractor. Thoughts?

12:58:13 From Cam Geer : and search needed a great customer experience — simply and effective

12:58:17 From rileyhughes : Just posted a blog post today, BTW :D

12:58:21 From Timothy Ruff : @kathryn: there is one way that top-end revenue potential player can be as large in a decentralized marketplace as it is in centralized marketplaces today: if the market grows significantly (as it did with Uber). Most of that effect is unpredictable, but there is at least one economic predictor: the degree to which transaction costs are reduced. The more efficient a market becomes, the bigger it becomes. With user sovereignty there will be many more players of all sizes, not just one dominant one in each category, because users can so easily move to the latest/greatest, as it should be.

12:59:33 From dsearls : That was Blogger after Google bought it. Ev & friends made no money onBlogger before that. Atom was created to flatten Dave Winer, and RSS won because it supported podcasting.

12:59:42 From johnnyfromcanada : Bitcoin mining is centralized in practice, since expensive mining encourages “poolingâ€, two or three pools of which currently control more than 50% of node work. So not BFT (also no finality).

13:01:32 From Dee Platero : Preach!

13:02:30 From johnnyfromcanada : Maybe we need a new economic model

13:02:41 From dsearls : Amen brother.

13:02:53 From Robert Mitwicki *THCF* : :+1 no way to fix browser !!!

13:03:19 From KathrynHarrison : I think the answer lies in incentives. If people building the initial system could have made more money maybe there wouldn’t’ have been so many holes for corporates to capturre.

13:03:21 From Steve Todd : Web standards are developed by organizations that need centralization to survive.

13:03:39 From KathrynHarrison : Right somebody has to fund those orgs

13:03:44 From Cam Geer : +1 kathryn

13:03:59 From Gabe Cohen : +1 how to change the incentive structure?

13:04:02 From johnnyfromcanada : One possible way to shift the economy is to stop locally optimizing the stock market. See the Long-Term Stock Exchange (LTSE): http://ltse.com

13:06:27 From Cam Geer : http://en.wikipedia.org/wiki/Coase_theorem

13:06:33 From rileyhughes : Kathryn, it’s incentives and speed. I totally agree with you about incentives. The other thing is that open development and standards work is painfully slow. Developing in a closed, centralized environment is super fast.

13:06:52 From johnnyfromcanada : Hedera folks talk a lot about microtransactions being a way to re-monetize industries that have been centralized, including media.

13:06:53 From dsearls : Dave is moving toward a theory of decentralization. At ProjectVRM we have one: free customers are more valuable than captive ones—to themselves, too sellers and to the market. We await proofs in the territory Dave is laying out here.

13:07:59 From Cam Geer : +1 doc for free customers

13:08:02 From dsearls : Sam http://www.amazon.com/Tomorrow-3-0-Transaction-Cambridge-Economics/dp/1108447341

13:09:10 From dsearls : See emancay: http://cyber.harvard.edu/projectvrm/EmanciPay

13:09:35 From dsearls : It's about micro accounting, not micropayments

13:10:22 From dsearls : But it invites a fresh decentralized look at better economic signaling between demand and supply.

13:10:23 From KathrynHarrison : @Riley— totally agree on speed!!! Plus, most open source devs are being paid by someone to contribute.

13:10:25 From Cam Geer : like enterprise open source — free starter kit // then premium services

13:10:45 From MarkL. @smartopian : One issue here - is that the word Trust is actually Assurance - (in decentralised speak).

13:10:47 From KathrynHarrison : Also much easier to make difficult decisions when there is centralized leadership

13:10:48 From dsearls : We came up with it 13 years ago, but have been awaiting a perspective like Dave invites here.

13:11:00 From johnnyfromcanada : The primary concern with decentralized systems for enterprises is that there is no one to sue when things go wrong. That may not matter to individual people. So are we talking about people or enterprises? Or even possibly the elimination of the need for enterprises altogether?

13:11:23 From MarkL. @smartopian : Privacy is = Trust in Decentralised concepts

13:11:32 From Marc Davis : People are paid to post on Instagram today as “influencersâ€

13:11:49 From Matt Norton : By firms ^

13:12:04 From KathrynHarrison : Lots of people are paid to create content today— see you tubers who get paid to play fortnite

13:12:21 From KathrynHarrison : +1 Matt and Marc

13:13:26 From johns : And there is a shift with video content creators who now rely on Patreon, merchandise and Tips during live streaming away from advertising

13:14:50 From johnnyfromcanada : Of course David is in Vegas! What happens in Vegas, stays in Vegas!

13:15:06 From Marc Davis : The “concierge†sounds a lot like John Hagel’s “trusted infomediaryâ€: http://en.wikipedia.org/wiki/Infomediary

13:15:38 From MarkL. @smartopian : In the Pan-Canadian Trust Framework -for digital identity ecosystem - this (concierge role is) called Notice and Consent Processors -

13:15:39 From Cam Geer : in fact that is an intent that Dave cast with the conceirge for his interest in going to see a show

13:15:52 From johnnyfromcanada : “Concierge†is a good analogy also because they do not ask who you are!

13:16:02 From johnnyfromcanada : They focus on “what†you need.

13:18:39 From Marc Davis : The self-sovereign “Personal Data Store/Server†concept can in theory outperform traditional “data aggregators†because it can have permissioned access to ALL of a person’s data and thus provide a more holistic and valuable set of data and intents than any single data aggregator.

13:18:42 From dsearls : do we have a note taker here?

13:18:53 From scottmace : Yes I am taking notes right into the Google doc in real time

13:19:07 From dsearls : very cool, scott, thanks!

13:19:34 From johnnyfromcanada : Ya, a CMM for sovereignty

13:21:26 From johnnyfromcanada : Your network will be uncorrupted if no more than 1/3 of nodes can be compromised (BFT).

13:23:05 From KathrynHarrison : I think even the most idealistic people (developers) are looking for leadership and the most popular “decentralized†systems all have them as you’ve shown (though of course they aren’t decentralized) — Satoshi, Linus, Vitalik…

13:24:39 From johnnyfromcanada : Indeed, do you trust benevolent dictators?

13:25:02 From dsearls : The Internet itself is distributed. Not absolutely, but pretty close.

13:25:29 From Laura Jaurequi : Sounds like a utility...

13:25:30 From johnnyfromcanada : But not trusted

13:26:14 From johnnyfromcanada : Trust in a trust-challenged environment is pretty much the whole point of DLT / Blockchain.

13:26:34 From Bill Wendel1 : @MarcDavis & @JohnnyFromCanada… Does John Hage’s “Trusted Infomediary†goes beyond acting as a concierge to acting as an advocate, ie. a fiduciary? Here’s an example of a buyer’s agent in the auto industry: http://www.authorityauto.com/about

13:27:02 From johnnyfromcanada : So like power of attorney?

13:27:21 From johnnyfromcanada : Or one form of guardianship?

13:27:48 From dsearls : peaking at 51 people now here. very good.

13:28:22 From johnnyfromcanada : The laws are critical…. unless you really are going to defy the law.

13:28:58 From Sam-Smith : Dfinity uses a DCID (Data Center ID) that is managed by the Network Nervous System. Users must request and be issued a DCID. Its root of trust is not as decentralized as it should be. Entropy is the only root of trust we need.

13:29:02 From Bill Wendel1 : Johnny, I’ll answer your question from my experience as a buyer agent in real estate. n the past, conversations about fiduciary duties in real estate have focused on agency duties — essentially requiring the RE agent to disclose which side their on, buyer or seller. These principles Dave is describing point to where the agency conversation needs to go in the future in real estate — the role of Information Fiduciaries. More specifically, now real estate brokerage companies use data to help buyers, sellers and increasingly homeowners.

13:29:42 From Cam Geer : exactly! RISC -V!!!

13:30:18 From Cam Geer : would be great to have a RISC-V / IIW meetup

13:30:47 From Cam Geer : they are the silicon side of the SSI conversation

13:31:35 From Robert Mitwicki *THCF* : +1 for what Marc jus said!

13:32:22 From Marc Davis : @BillWendel1 Similar to John Hagel’s “trusted infomediary†concept, legal scholar https://en.wikipedia.org/wiki/Jack_Balkin has the idea of the "information fiduciaryâ€.

13:32:24 From johnnyfromcanada : @Sam-Smith - Indeed, however the Network Nervous System in theory at least is an effective DAO via a “liquid democracyâ€.

13:32:30 From KathrynHarrison : Someone has to pay!

13:33:00 From dsearls : http://www.luxuryinstitute.com/

13:33:01 From Cam Geer : my bad sorry

13:33:30 From Robert Mitwicki *THCF* : The way we think about collective data hubs is that you strip out the PII's and then you can do criteria search and do what Marc comment on.

13:35:19 From Bill Wendel1 : Question about Data collectives that would have value for consumers. As the real estate industry shifts from a seller’s market to buyer’s market in a post Covid-19, would it be possible to organize a data collective for homebuyers to offset the seller bias, lack of transparency in organized real estate? If so, the market opportunity is HUGE!

13:35:23 From johnnyfromcanada : For any decentralization to be BFT, there must be some kind of “control†to ensure that 1/3 of work / stake / etc. does not get corrupted. That “control†is (currently) either a consortium (governing council, hopefully sufficiently decentralized) or a crypto-economics (with a high enough valuation to block capture of 1/3 of the value).

13:35:25 From Marc Davis : Link to my keynote from pii2010 that articulates some of the benefits of a personal data economy based on what we then called “user centric†(but might today call “self sovereignâ€) models: http://vimeo.com/14401407

13:38:10 From dsearls : good one, Marc. Maybe now we can finally make it happen.

13:38:53 From dsearls : i'm going to punch out and go to the session Brian is at...:-)

13:39:10 From KathrynHarrison : Despite my critiques, I think the work is awesome and I’d love to help find the right incentives and models to make decentralized systems competitive.

13:39:23 From johnnyfromcanada : Decentralization is an infinite regress. Do you trust the chip maker of the digital wallet that contains & computes on your behalf?

13:40:15 From johnnyfromcanada : Perhaps Trusted Execution Environments (TEE) and homomorphic encryption might offer solutions.

13:40:21 From Cam Geer : great job Dave!

13:40:23 From Cam Geer : thanks

13:40:29 From KathrynHarrison : Awesome conversation!

13:40:33 From Marc Davis : Thanks @dsearls! The work on SSI discussed at IIW looks really promising!

13:40:35 From Laura Jaurequi : Thanks!

13:40:46 From Alex Blom : Great conversation, txs