0:00 Samuel Smith ｜ State of the KERI Suite ｜ KERI Conference 2026
0:02 All right. I want to introduce Henk
0:07 van Cann. Right, the KERI
0:09 Foundation (https://keri.foundation) would not exist without him.
0:12 He came out of retirement. He likes to
0:15 play guitar and sing on the streets in
0:18 the Algarve in Portugal, and he could be
0:21 doing that instead of being here with
0:23 you guys. So,
0:24 give him a big round of applause.
0:28 And he's also a financial sponsor of
0:31 the KERI Foundation. So, there's no
0:33 amount of
0:34 thanks that
0:36 is too much for him.
0:37 I want to thank everyone for coming.
0:43 This is KERI Conference 2026.
0:47 Get a t-shirt cuz there'll
0:50 never be another first-ever KERI
0:51 Conference. So, you know.
0:54 And I think they're downstairs, right?
0:55 Yeah. All right.
0:57 And then…
0:59 one more shot for the
1:02 for the sponsors.
1:03 So, I'm going to talk about the state of
1:05 the KERI Suite,
1:07 the journey to adoption.
1:10 and
1:13 I look at 2026. We've been doing KERI
1:16 for a lot of years now,
1:19 and it feels like this is going to be
1:22 the year of KERI. So, mark
1:24 your calendars. This is where the world
1:26 starts to understand and
1:28 and know that KERI exists.
1:32 Some early feats
1:33 that have already happened in 2026.
1:37 The KERI Suite of protocols released
1:39 as Linux Foundation open standards. That
1:42 was a four-year process starting in
1:45 2020.
1:47 And we've got TrustoverIP. Drummond [Reed, red.] who
1:49 is the chairman is here today.
1:52 And then we have a couple of members
1:53 of the steering you know, several member
1:55 at least three members of the steering
1:57 committee here today. They've
1:59 been helpful; Karla, Niko and Drummond.
2:01 We started a funded project to build
2:08 open infrastructure for KERI.
2:10 We received a grant from the Utah
2:13 Governor's Office of Economic
2:15 Opportunity to build a digital
2:17 identity utility for SEDI,
2:20 based on KERI.
2:23 the Telecom Industry's OVC Alliance is
2:28 doing proof of concepts to vet use
2:31 of the KERI Suite of protocols and the
2:33 verifiable voice protocol. And later
2:35 today, you'll have Daniel Hardman
2:37 is here. He'll talk about that.
2:40 We have Utah State passed the SEDI
2:44 implementation law
2:46 funded law to build SEDI.
2:50 There is
2:52 production implementation of
2:57 using those for stopping call spam.
3:00 Healthcare's vital
3:02 Initiative. We'll have a talk later
3:04 today about that with the vLEI, KERI
3:09 and using the KERI for healthcare
3:12 data provenance.
3:14 We just completed the SEDI
3:18 Policy summit the last 2 days. How
3:22 many people are here that were here the
3:23 last 2 days?
3:25 All right, almost everybody. Fantastic.
3:27 So, you so I don't need to tell you any
3:29 more about what happened there.
3:32 And first-ever annual KERI conference
3:35 So, that's just this
3:38 first 4 months of 2026. We're just
3:40 getting started.
3:43 One of the things I want to
3:45 look at is
3:48 what are the chances of broader
3:52 adoption? What are the sort of
3:54 rules? And so, I went back to
3:57 Malcolm Gladwell wrote a book
4:00 called The Tipping Point.
4:01 And he had rules for
4:05 when adoption reaches a tipping point.
4:07 And this is a curve.
4:11 And the tipping point is this inflection
4:12 point down he- down here where
4:16 you sort of go from
4:18 flat adoption to where you start to
4:22 get a significant grow to
4:25 a steeper curve. And the
4:28 the reality is down here, it's not a
4:31 smooth curve. It's really noisy down
4:33 here. And so, really the tipping point
4:35 is when the noise starts to dissipate
4:38 and you start to see
4:41 month over month, week
4:43 over week, year over year growth. And
4:47 where does that tipping point
4:51 happen? How do we get to that
4:53 tipping point? So his three rules
4:57 are the law of the few, the stickiness
4:59 factor, and the power of context.
5:02 So, the law of the few. What does that
5:04 mean?
5:05 Well, if you have a sufficiently
5:07 influential set of early adopters, those
5:10 are the few.
5:12 That means you have to have people
5:14 that recognize the technical
5:16 suitability. And then they have to try
5:19 to get other people to adopt it.
5:22 And so, you guys
5:24 are the few. You're the ones.
5:28 So, let's see what's happened.
5:30 You have to first overcome the seven
5:33 stages of adoption resistance.
5:36 I know there's some people
5:38 in this room that [are in there:] shock,
5:40 denial, anger, bargaining, testing, and
5:42 acceptance, right? Most of you are
5:46 well along. There's still a couple that
5:48 maybe they're in the <i>bargaining</i>
5:50 stage yet. They're still trying to well,
5:52 do we really need KERI or you know,
5:54 does it really solve what we want? We'll
5:55 see.
5:58 What are some adoption vectors
6:03 that the few start; sort of like an
6:05 epidemic, right? What are some
6:07 of those? Well, I sort of hinted at
6:09 those. We have the telecom sector.
6:11 That could be a massive adoption
6:14 vector.
6:15 We have the healthKERI sector. That
6:18 could be a massive adoption vector.
6:20 And we have state entitlements,
6:23 which
6:23 Could be a massive adoption vector.
6:25 Any one of those
6:28 could
6:29 cross the tipping
6:31 point. And since I'm a fault-
6:34 tolerant person, I believe in fault-
6:37 tolerant systems. I got to have three,
6:39 right?
6:40 You got to have three massive adoption
6:42 vectors to make sure one works, right?
6:44 So, that's where we're at.
6:46 Widespread bootstrap vectors. So, the
6:49 difference here is
6:52 in order to have adoption that is
6:55 we had a
6:56 saying in Florida, "What
6:58 comes in over the transom goes out
7:00 over the transom.
7:01 You know, as a wave washes fish in, it's
7:04 junk fish, you don't want it, right? But
7:06 what we want is to have adoptions
7:09 that we didn't plan on. We didn't know
7:12 that they were going to happen. And they
7:13 they just spin up
7:16 spontaneously. And so, we need something
7:18 that will bootstrap those without us
7:20 forcing them. These up here
7:22 are the result of years of work.
7:25 We started working
7:27 with SEDI, three years
7:29 ago. I know how long Provenant and
7:32 healthKERI have been working in their
7:33 sectors. It didn't happen overnight.
7:36 It took lots of time, right? But we want
7:39 to also have vectors that happen
7:41 quickly. So,
7:43 we have two bootstrap vectors,
7:45 organizational Root of Trust, that's
7:47 GLEIF.
7:48 And that's been the thing that has
7:50 been enabling these other ones, right?
7:52 They went to production with
7:54 the vLEI in 2022. That's now
7:57 been 4 years. Seems like it was
8:00 yesterday, but it's been 4 years.
8:03 We're starting to see the
8:06 fruits of that.
8:08 But one of the things that is new this
8:11 year and in the next year is a personal
8:13 Root of Trust. That's never happened
8:15 before, not an organizational root of
8:16 trust, a personal Root of Trust. That
8:18 means lots of businesses
8:22 that wouldn't normally consider using
8:25 digital identity, centralized digital
8:27 identity,
8:29 are going to get for free bootstrapped
8:33 identity assurance that is highly
8:36 trustworthy.
8:37 That means every industry
8:40 has the potential to leverage
8:43 that Root of Trust, which is Utah state.
8:46 And then all scale enablement vector,
8:50 that's open standards,
8:52 open source, and open infrastructure.
8:55 I'm going to talk about what I mean by
8:56 open infrastructure a little bit later
8:57 in this talk.
8:58 So, there's the law of the few. So, I
9:00 feel like
9:02 we've got that.
9:05 The next one is the stickiness factor.
9:06 What makes it stick?
9:08 Well, it needs to be compelling and
9:10 contagious.
9:12 So, we have to think about things that
9:13 make it, right? So, we need to have some
9:16 sort of pathology that's endemic
9:20 that isn't going to go away and it's
9:22 going to stick
9:23 and it's going to keep driving adoption,
9:26 and that is fraud.
9:28 The world is filled with fraud. Every
9:30 business is suffering the cost of fraud.
9:33 And fraud includes identity theft. It
9:36 includes,
9:37 all of the security,
9:39 exploits,
9:42 ransomware.
9:43 But, it includes just normal day-to-day
9:47 business, factoring fraud, PO
9:49 fraud [Purchase Order, red.], those sorts of things.
9:52 And they're getting worse. And one of
9:54 the things that are making them worse is
9:56 the fact that the tooling that enable
9:59 fraud has now reached the point where it
10:02 is better
10:03 than the mechanisms that protect against
10:05 fraud.
10:06 And so,
10:08 that's going to continually push people
10:11 over the edge. They're going to
10:12 say, "Do I really need KERI?" And you
10:14 go, "Well, can you fix fraud without
10:17 KERI?"
10:18 And they go, "Well, no, not really."
10:19 Okay, then you really need KERI.
10:22 Regulator imperatives to
10:25 repress fraud and abuse. Those
10:28 are happening in several industries. For
10:30 example, in the telecom industry, FCC.
10:33 In the healthcare industry ..,
10:34 you see Scott [Scrimshire, red.] nodding his
10:35 head.
10:39 In state endorsed
10:40 digital identity.
10:43 You're you now have regulators that are
10:46 saying, "Hey, we need to we need to
10:49 we need to we have an imperative that
10:51 says we've got to get rid of fraud and
10:53 abuse." And by abuse, I mean
10:55 exploitation of citizens in the case of
10:57 personal identity.
10:59 The malaise of digital exploitation.
11:02 Do any of you feel like you're sick
11:06 of being exploited online?
11:10 Only a few of you. Okay.
11:12 Do you know of anybody that
11:14 feels like they're sick of being
11:16 exploited online? Okay. [laughter]
11:18 Do you have a desire to
11:21 control your digital presence? Yes.
11:24 Absolutely! The
11:26 So that's a stickiness factor. Is
11:28 that desire going to go away?
11:31 Do you know anybody
11:32 wants to have less control?
11:39 I know some.
11:40 Yeah,
11:40 There are few. But most of the ..,
11:42 everybody in this room, I'm pretty sure,
11:43 wants more control over their digital
11:45 presence.
11:46 The other thing that's happening
11:49 is that the technology
11:51 that enables KERI
11:54 is key management.
11:56 And that means the devices, the
11:58 software, the systems, operating systems,
12:01 have to start moving in the direction of
12:04 educating people
12:06 to manage their own keys. And that is
12:08 happening.
12:10 You know, when we started KERI and
12:12 when I started doing digital
12:14 identity in 2015, 2016,
12:17 the people were
12:20 really saying, "There's no way
12:22 you're going to get
12:24 people to manage the private keys."
12:27 The blockchain community had
12:28 already gone there. They had to
12:30 manage private keys to be on the
12:31 blockchain. The technology
12:34 got a got a big boost. Now you're
12:36 seeing it in every system.
12:39 This endemic pathology is pushing
12:42 people to better key management.
12:45 As people become familiar, the
12:48 friction to adoption reduces.
12:52 So, I think we have stickiness.
12:55 The power of context.
12:58 What that means is timing and
12:59 externalities. Is the timing right?
13:06 I think the specter of AI is making
13:09 an unprecedented
13:12 context that says the timing
13:14 is right.
13:15 We are faced with
13:19 the potential for the annihilation of
13:21 trust. There's an impending doom. I
13:23 don't If you guys don't see it, if you
13:26 aren't, go read OWASP 100-page
13:30 report they just
13:31 published
13:33 on how to deal with AI threats to
13:36 security.
13:38 And it is chilling cuz they say
13:41 it has never existed before, it is a
13:44 different type of threat, and we don't
13:46 have a cure. All we can do is mitigate
13:49 some of the symptoms.
13:51 And they're only going to get worse. And
13:52 it's 100 pages long.
13:54 Right. And it's just starting.
13:56 They're going to add 100 pages
13:59 on a regular basis because the threats
14:01 are increasing at an exponential rate.
14:05 So, impersonation,
14:07 been around for a while. AI
14:08 impersonation, deep fakes,
14:11 those, they just barely
14:13 happened.
14:14 When was the ChatGPT moment?
14:19 Yeah, 2023. It's only 3 years around,
14:22 and we're already
14:23 seeing those effects.
14:26 Exploitation, you thought somebody
14:29 could exploit you? Wait till an AI
14:31 exploits you.
14:34 Surveillance, you thought you were being
14:36 surveilled online?
14:39 Just ask an AI to surveil you.
14:42 You know, I've done this test for
14:44 people where I went out and said, "Hey,
14:47 I'm working with X,
14:50 and he needs somebody to manage his
14:52 online presence and reputation, and I'm
14:55 doing that for him."
14:57 And it says, "Okay."
14:59 And after about three prompts, it says,
15:02 "So, would you like me to create an
15:04 interactive dashboard that tracks all of
15:06 the things they say and do online?"
15:09 And I always say "No."
15:11 [laughter]
15:12 It takes
15:14 literally 10 minutes for anybody in
15:17 the world to say, "I want to know
15:18 everything about everything that anybody
15:20 that they know of that they have a name
15:22 for has done.
15:24 And that's just
15:25 starting.
15:27 We're not even close to the
15:28 capability yet.
15:31 Sufficient political will to legislate
15:33 new civil rights. I think you saw that
15:35 the last 2 days.
15:38 Individual control over digital ID data
15:40 and context. That is
15:42 what we need is legislation that gives
15:45 us back control over our internet
15:47 presence.
15:48 <i>Duty of data loyalty</i> is probably the
15:50 most significant legislative initiative
15:54 that we can put in place to enable
15:57 control over
15:59 over our digital ID presence online.
16:02 We can't do it with anything less. It
16:05 inverts the liability equation. And for
16:07 those of you who don't understand the
16:09 difference between
16:12 consumer harm legislation and fiduciary law,
16:16 the standard of proof is different.
16:19 if I have consumer protection,
16:22 I can't
16:24 punish the person abusing, me unless I
16:26 can show harm. The standard of harm is
16:28 pretty high. And it's pretty hard
16:31 to show harm in a diffuse
16:34 internet online where your data gets
16:36 spread out all over the place and you
16:37 can't point to somebody and say, "You
16:39 harmed me." They go, "No, I gave
16:41 the data to this person. They gave it to
16:42 that person." And then you can't show
16:44 harm.
16:45 But fiduciary duty of loyalty says,
16:49 "If the fiduciary
16:51 is self-dealing,
16:53 then you can punish them because your
16:55 interests have to come before the
16:58 fiduciary's interests." That means
17:00 everybody that collects any data about
17:03 you online in Utah now has to
17:07 understand that if they self-deal,
17:11 if they benefit from your data and you
17:14 don't,
17:15 you can go after them.
17:17 That has never happened before. That is
17:19 the way place where it has to be.
17:22 So, that's the stickiness factor cuz
17:24 that's now in law.
17:26 And that stickiness, that's the context.
17:28 And then the last one, the inevitability
17:30 of KERI. You have to have true
17:32 believers who are going to fight through
17:34 the fights, who are going to
17:37 go to the meetings and put up
17:39 with the arguments and make the
17:41 case. And I think we have cadre of
17:45 people that are willing to do that. So,
17:48 people that are willing to do that.
17:49 And the thing is, we have a severely
17:52 constrained solution trade space.
17:54 And what I mean by that is, I spent
17:56 years
17:58 doing the trades. What can we do to
18:00 solve these problems? What tools do we
18:03 have available? And everything every
18:05 time I go, "Well, I could do this, but
18:07 wait, if I do that, then
18:10 here's the tradeoffs, and I don't
18:12 like those tradeoffs. And I do that,
18:13 here's the tradeoffs. I don't like that
18:15 those tradeoffs." And once you do the
18:17 tradeoffs, you realize that there's only
18:20 a very small
18:22 space for a solution, which means, as
18:25 far in my opinion, if it's not KERI,
18:28 it's going to be something that looks,
18:30 smells, walks, and quacks like KERI.
18:32 They just changed the name.
18:34 Because the under the hood, there just
18:36 isn't any other solutions. We have to ..,
18:39 cryptography only gives us
18:41 a handful of things that we can use.
18:43 That means that
18:46 the timing is right. So, anyway
18:51 Persistent problem cascade that
18:52 continually exhausts anything less
18:53 capable.
18:55 It's just every time you look at some
18:57 other system, you look at <i>DNS/CA</i>, you
19:00 look at <i>OAuth, OIDC</i>, you look at these
19:02 systems,
19:04 and they keep failing for the same
19:06 reasons.
19:07 People are starting to recognize that.
19:11 So, what does that mean?
19:14 It's a matter of when, not if. If
19:16 it's not this year, it's next year. If
19:19 it's not next year, it's the following
19:20 year.
19:21 Because none of this is going away.
19:37 I'm going to predict that before the end
19:40 of 2027,
19:40 KERI adoption will reach its tipping
19:42 point.
19:46 [applause]
19:49 All right.
19:53 The journey begins. So, what I wanted to
19:55 do on this first ever conference
19:58 is give a brief history of how we got
20:01 here.
20:02 Sort of like the foundations.
20:05 I think most of you have been there
20:07 through most of the journey.
20:09 But, I know there's some people that
20:10 haven't been, and this talk,
20:13 will be made available online for
20:15 other people. So, I thought this would
20:16 be a good opportunity to sort of
20:19 talk about, where it came
20:21 from.
20:22 I started down this path
20:25 back in 2013, 2014. I wrote a protocol
20:28 called RAET.
20:29 It's reliable asynchronous event
20:31 transport. It is an
20:33 end-to-end, peer-to-peer,
20:35 encrypted, signed, and it was used in
20:39 production at scale, and it was open
20:41 source.
20:42 So, this was years before most of you
20:45 have even heard of Signal. Back when
20:48 Signal wasn't called Signal, it was
20:49 called Whisper.
20:50 But, it was the start of that. So, what
20:52 was the lesson learned?
20:55 Well, I could solve all the problems but
20:56 key management with RAET.
20:58 [laughter]
20:59 So, I said, "Okay, that's not a
21:01 solved problem."
21:03 So, I already knew that
21:07 there was a big problem that had to be
21:08 solved sometime.
21:10 <i>OpenReputation</i>
21:12 I wrote a white paper.
21:15 It started using self-certifying
21:17 identifiers.
21:20 It had the concept of identity ledger
21:23 and identity graph.
21:25 It defined reputation as a modulator for
21:28 trust for online interactions.
21:31 And it
21:34 brought up the idea that we wanted to
21:35 have individual control over the context
21:38 in which we interact online.
21:40 Our internet presence.
21:42 And lesson learned from that is that I
21:44 needed persistent identity. I can't have
21:48 a good reputation without a persistent
21:49 identity, and that's the hard problem of
21:51 identity because nobody had a solution
21:53 for it back then.
21:56 <i>Identity System Essentials</i>, this
21:58 white paper. This is the first draft.
22:01 Introduced cryptonyms, built on the
22:03 SCID model. It provided a formal
22:08 identity model, primary secondary IDs.
22:11 Talked about identifier management
22:13 with HD key chains, three degrees of
22:15 privacy, and it introduced the concept
22:18 of self-sovereignty for identity. And
22:20 the draft paper was
22:23 shared in the community, and lots of
22:24 people amplified and built on the
22:27 concept of SSI, but as far as I know,
22:29 that was the first. And I stole it
22:31 from Phil. Because Phil Windley had
22:34 written some blog posts on
22:36 self-sovereign data, and I said, "Well,
22:38 we don't just want data to be
22:40 self-sovereign. We want our identity to
22:41 be self-sovereign." That was
22:43 back in 2016.
22:47 The lesson I learned, there's a
22:49 hard trade space. If you want to trade
22:51 off security and privacy and
22:53 self-sovereignty and control,
22:55 you're making some pretty hard trades.
23:00 Sovrin Foundation.
23:02 That really changed the environment
23:06 for decentralized identity.
23:08 The idea was there's going to be one
23:11 ledger for digital identity.
23:14 One ledger to rule them all.
23:16 And it had persistent identifiers.
23:19 It had prophylactic [preventive, red.] key rotation,
23:22 not recoverable key rotation. That
23:24 was going to
23:26 come a while later, but it built on
23:28 but it established the foundation of
23:30 managing your key state
23:32 so that you'd have persistent
23:33 identity.
23:35 Hard problem, shared governance.
23:38 That was a hard problem.
23:40 Spent a lot of time trying
23:42 to solve that problem.
23:44 2017, the ledger wars.
23:52 Okay, a lot of you fought in those wars.
23:54 What happened is
23:55 SSI was such a great concept
23:58 that everybody wanted it. Every ledger
24:00 wanted to have their own SSI
24:03 Identity.
24:05 So there was every time somebody would
24:07 have an application
24:10 that they wanted
24:12 decentralized identity for,
24:14 the very next thing was, "Okay, which
24:16 ledger are we going to put it on?"
24:19 All the people for all
24:21 the ledgers would come to the table and
24:23 say, "No, my ledger. No, my ledger."
24:25 And that's what it looked like
24:27 for several years.
24:28 And then
24:30 GDPR came out in 2018 and that made it
24:33 difficult to have identity on a ledger.
24:37 So the whole idea of persistent identity
24:40 using a ledger is a problem for GDPR.
24:44 And so
24:46 the ledger wars, so
24:49 the insight was that blockchain is too
24:52 centralized
24:54 for truly portable identity.
24:58 That was a really hard lesson to learn.
25:00 It took several years of fighting the
25:01 ledger wars to realize that.
25:04 The way of KERI.
25:08 Really the first paper that sort of
25:12 opened the path was this one in 2018.
25:16 The three R's of key management and one
25:18 of those R's
25:19 is Rotation and this is where I
25:22 defined the concept of pre-rotation. So,
25:24 it's in that paper.
25:26 It says, "How are we going to do key
25:27 rotation?
25:28 If we do it the normal way,
25:31 where we use our signing key to rotate
25:34 to a new signing key, if our signing key
25:38 is compromised,
25:39 then we lose control of our identity and
25:41 that's all of blockchain, by the way.
25:44 And that's all of blockchain
25:45 identity. If you lose control of your
25:47 signing key, you lose control of your
25:49 identity. That's
25:51 kind of a problem.
25:53 The other approach is
25:55 have a rotation key that you use to
25:58 rotate your signing key.
26:00 But then what happens if your rotation
26:02 key gets compromised?
26:04 Well, then you have another rotation key
26:05 to rotate the rotation key that rotates
26:07 the signing key. And I had several
26:09 cryptographers says that's best
26:10 practices. Well, what happens when that
26:12 one says two is usually enough. You only
26:14 need two rotation keys. They'll
26:16 never get there.
26:19 Good idea, but
26:20 maybe we could do better.
26:22 Pre-rotation is one-time use only
26:24 rotation keys.
26:26 You rotate your rotation key every time
26:28 you rotate your signing keys.
26:31 Pre-rotation
26:33 and then that also introduced the idea
26:36 of having a log of your key state.
26:41 Didn't call it a KEL event because
26:42 we hadn't coined the term KERI, so it
26:44 couldn't be a KERI event log,
26:45 a KERI Key Event Log, but it was a log.
26:49 Had multisig individual source and
26:52 provenance data.
26:55 The idea is make key management
26:57 fault tolerant. So, I started down that path.
27:01 In 2018, a little bit later,
27:03 paper called. A DID for everything.
27:06 It introduced the idea of <i>Verifiable</i>
27:08 Provenanced Graphs of Cryptographically
27:10 Chained Data.
27:11 That's the core idea that is now in
27:13 the ACDC standard.
27:19 Yeah.
27:21 In 2019, KERI version one white paper.
27:24 So, based on those, I said, "Okay, let's
27:26 see if we can figure out a way to do it
27:27 without a blockchain."
27:29 And the idea is that you have a
27:31 cryptographically verifiable append-only
27:33 event log for each identifier.
27:36 There's no shared governance.
27:39 You get rid of shared governance, you
27:40 get rid of the ledger wars.
27:42 Introduced CESR,
27:45 key rotation, witnesses, deconstructed
27:48 blockchain.
27:50 It solves the hard problem of cross
27:51 trust domain transfer.
27:54 Lesson learned:
27:56 we don't need a blockchain.
27:58 And I spent a couple of years arguing
28:01 with people that said, "No, you have
28:02 to have a blockchain. You can't do it
28:04 with a Key Event Log."
28:06 But now I hardly ever have anybody
28:10 tell me that you need a blockchain for
28:11 identity. Does anybody Is anybody
28:13 arguing that still in the room?
28:16 You don't have to raise your hand.
28:19 [laughter]
28:20 I took that paper on
28:24 the road. I went to IIW several
28:26 conferences, shared with everybody that
28:27 I knew. Tell me what I'm doing wrong.
28:29 Got lots of really good valuable
28:31 feedback. There's a list of
28:36 of acknowledgements in the back of
28:38 the KERI V2 white paper because there
28:40 were some really significant
28:42 meaningful enhancements that came from
28:44 that. One of those is the watcher
28:47 network,
28:48 CESR streaming, hierarchical
28:50 delegation, layered threshold
28:52 structures
28:54 in more detail.
28:56 Basically, it solved all the hard
28:59 problems of key management and so it
29:01 was ready to go and that was 2020.
29:05 And then the same year,
29:09 while we were still refining
29:11 the version 2 white paper came
29:14 across the idea of quantum-secure DIDs.
29:16 I co-wrote a white paper on this.
29:19 Using hashes to protect the
29:22 pre-rotated keys means that we've
29:23 pre-migrated along with the cryptography
29:25 agility. We already
29:28 have a story for post-quantum secure
29:31 and Daniel [Hardman] and I, Daniel did most
29:33 of the work, co-wrote a recent paper
29:35 about the post-quantum
29:38 position for KERI. If you're not
29:39 familiar with that, go read that
29:41 white paper. And then
29:43 you can maybe do a reference for it.
29:45 2020, GLEIF
29:49 Got religion. They said, this
29:52 works. We don't like the ledger wars.
29:54 I think that was your main motivation,
29:55 Karla? Yes. The ledger wars. They
29:58 didn't want to fight those anymore.
30:01 so lasting peace from the ledger wars.
30:05 important insight. It solves
30:07 organizational ID. Organizational
30:09 identity is a new concept. Came from
30:11 that because now you have a
30:12 organizational Root of Trust. It's a
30:14 bootstrap for any organizational ID
30:16 problem.
30:17 And now people in this room are starting
30:19 to get that. I wrote a white paper
30:22 "Universal Identifier Theory" just to
30:24 extend the identity model, unified model
30:27 for identifiers, bootstrap trust, be a
30:30 reputation by reference, multi-valent
30:32 key management, starting to talk about
30:34 how we build infrastructure at scale.
30:37 That went into more detail.
30:40 And then in 2020 we formed a
30:43 task force at Trust over IP to start
30:46 standardizing this stuff and
30:48 ACDC
30:50 became a thing,
30:51 at that point.
30:54 Authentic Chained Data Containers, and
30:56 I think,
30:57 We'll talk a lot about those in the
30:59 next 2 days.
31:03 And then 2022, GLEIF vLEI in production.
31:08 I won't go into
31:10 most of you are familiar with that.
31:12 So, now where we're at?
31:16 There was a story
31:18 missing in KERI. KERI is about
31:23 authentication, proving that you control
31:25 an identifier.
31:27 But, there's more to the story
31:30 and had lots of conversations about how
31:33 do you deal with privacy and
31:36 confidentiality, and
31:38 wrote the white paper, the SPAC white paper
31:42 which is now a task force
31:45 within Trust over IP working group
31:47 at ToIP, and it's like draft version 3
31:50 or something,
31:52 To solve,
31:54 protected communications
31:57 online.
31:59 It uses a three-layer tunnel,
32:05 and the idea is, that if I have
32:07 layered confidential contexts,
32:12 then I can solve
32:15 surveillance and privacy using layered
32:18 confidential contexts. And so,
32:21 that's the basic idea. And what
32:24 I mean by "It's exploitation, stupid,"
32:26 what I'm saying is that the real
32:28 question is, how do we protect people
32:31 from being exploited, not how do we help
32:34 people hide better. Hiding is not a very
32:37 fun place to be. If you've ever tried to
32:39 really hide,
32:41 you don't have a life.
32:43 I know people that live off grid. They
32:44 don't have a life.
32:47 You want to be able to live with a life
32:50 and not be exploited, and that means
32:53 having control over the context of your
32:55 data, not having a context, right?
32:59 That's the main difference.
33:01 Do you have a context that you control,
33:04 that you share data within, or do you
33:06 not have any context?
33:08 Most of the community
33:10 says, "Don't have any context."
33:13 That doesn't work very long
33:15 because
33:16 you live longer than a week, and you can
33:19 maybe have no context for a week or a
33:21 month or 3 months, but you're going
33:24 to run out of places to hide
33:26 after a while.
33:27 So, you need
33:29 a fortress that you can hide in.
33:31 You need multiple fortresses you can
33:32 hide in and have context.
33:35 So, 2023, Utah State Senate work begins.
33:38 Wrote a white paper called <i>Sustainable</i>
33:39 <i>Privacy</i>, talks more at length about the
33:41 regulatory things.
33:43 Fiduciary data loyalty
33:47 is in there.
33:48 The idea is you need a comprehensive
33:51 approach. Technology by itself won't
33:53 solve the problem. You've got to do it ..,
33:54 That meant as a
33:56 technologist, spending a lot of time in
33:59 the room with
34:01 regulators and politicians,
34:05 which were not my favorite
34:07 people, but now they become my favorite
34:08 people. I'm looking for Chris Bramwell,
34:10 but I don't see him.
34:11 He's late today.
34:20 ToIP, we got our own working group,
34:21 which helped speed along the things. We
34:22 had some
34:24 Cardano Foundation made a big investment
34:26 in open-source software.
34:29 Utah State started passing legislation.
34:32 KERI Foundation we established last
34:35 year [Dec 2024, red.]
34:40 No, that was that's wrong. That's
34:42 2024 12 02. So, it's in just the end of
34:45 2024, but it's really last year, to
34:48 try to foster
34:50 KERI throughout the world and get adoption.
34:53 And then we're back; we're now at
34:56 2026.
34:57 And we already did what happened in
34:59 2026. So, that's the history.
35:03 And you know what? I left a whole
35:05 lot of stuff off. There's been a huge
35:07 amount of work by people in this
35:08 community pushing KERI along. But the
35:11 core ideas, where they came from, and
35:13 then
35:14 you know, the milestones that got us to
35:16 where we are, I put on there.
35:19 So, what's the mission of the KERI
35:21 Foundation? Our mission is to foster
35:23 open infrastructure.
35:25 What does that mean?
35:27 It's the I in KERI. I had somebody ask
35:30 me, "Why Is that a typo? You made
35:32 the I in KERI red." I said, "No, I'm
35:34 trying to emphasize that I stands for
35:35 infrastructure."
35:37 KERI doesn't exist without
35:38 infrastructure. It's not a system that
35:40 works without infrastructure. And if you
35:42 don't have infrastructure, people can't
35:44 adopt it. They can't practically use it
35:47 because nobody wants to build their own
35:48 infrastructure.
35:50 So,
35:52 KERI is a decentralized key management
35:55 infrastructure.
35:56 Right? It's It was designed to be web
35:59 scalable
36:00 so that we get an ultra low cost floor.
36:03 Right? That's why blockchain ledgers
36:07 were a problem.
36:10 The idea is .. [tokonomics, red.],
36:12 and I spent time working in the
36:16 blockchain world, I think there was
36:18 somebody here from the Wyoming
36:20 delegation yesterday that used to work
36:21 for Consensus when I worked for him.
36:24 But he had a beard and you know, had a
36:25 different lifestyle. He was all
36:27 clean-shaven and had a suit on and I
36:29 didn't recognize him and he said, "No,
36:30 it's me, really. We worked together." And
36:32 I go, "Oh."
36:33 You know, it took me a minute.
36:36 But, the idea is <i>Tokenomics</i>. You build
36:41 incentives into the governance structure
36:43 so that people work together to build to
36:46 to make your infrastructure alive.
36:48 Another term for that is a
36:50 <i>cooperative model</i>. You can be in
36:52 competition. You can make money.
36:54 Everybody's making money.
36:56 But, they're helping everybody else out
36:59 to build the infrastructure that makes
37:00 it all possible for everybody. And
37:02 that's the idea.
37:04 So, we want to incentivize multiple
37:07 vendors.
37:08 No single vendor lock-in. We don't want
37:10 that. That's bad for KERI.
37:16 If you think about it,
37:19 if you have pre-rotated private keys,
37:24 you can and we have this in the code
37:28 in the Keeper, you can
37:29 reconstruct
37:32 all of your keys if you're using HD
37:34 keychain from any copy of your KEL
37:37 because the HD path is in the KEL. It's
37:40 the sequence number and the offset into
37:42 the into the key list. You don't need
37:44 any other information. You actually
37:46 don't have to worry about losing
37:50 all of your information
37:52 as long as one copy of one legitimate
37:55 valid copy of a KEL exists somewhere in
37:57 the world, you could reconstitute it.
38:00 We don't want wallet vendor lock-in.
38:02 We want multiple wallets. So, we now
38:04 have several wallets that are open-source
38:06 and we're going to demo
38:09 open-source stuff. So, the KERI
38:11 Foundation team, like they've
38:14 only been doing this since January, but
38:16 you can ask them whether
38:18 whether it was hurting their brain to
38:20 try to learn KERI in 3 months.
38:23 Some of them had a head start.
38:26 We've added another wallet. We've
38:29 got a mobile wallet.
38:32 You'll see some talks today of
38:35 people that are building wallets and
38:37 and infrastructure for KERI so that
38:39 we have multiple vendors because
38:42 adoption requires multiple vendors.
38:44 No big adoption wants to have a
38:47 single source of the technology. They
38:49 want to see they want to see multiple
38:51 vendors.
38:54 We have in KERI the concept of
38:57 <i>Non-Cooperative Key Custodianship</i>.
39:00 You can have key custodians that
39:03 are custodians of your signing keys.
39:05 They can have the keys.
39:09 And they can go away. They can hate you.
39:12 They can lose them. And you can still
39:15 reconstitute control over your
39:16 identifier
39:18 because you just have to retain your
39:20 pre-rotated private keys.
39:23 And so the major friction
39:26 of people managing their keys and
39:28 managing their signing infrastructure,
39:30 you can
39:32 rent out.
39:33 But you're not locked in because at any
39:36 moment in time you can pull your
39:38 pre-rotated keys out of
39:40 air gap cold storage, publish your
39:42 rotation,
39:44 signing 'See-ya' to your custodian.
39:47 And I think there have been
39:50 lots of money lost in the blockchain
39:53 world from
39:55 key custodians who have .., what do they
39:57 say? "Not your keys,
39:59 not your coins." Not true in KERI.
40:02 Not your keys, <i>still</i> your identity!
40:04 Not steal, still.
40:08 [laughter]
40:11 Users benefit from witness pools that
40:14 have no common mode failure sources. So
40:17 if your witnesses all come from the same
40:18 vendor, the vendor is a common mode
40:20 failure source. So, you really want to
40:23 tell your customers,
40:25 you know, have witnesses
40:28 supported by multiple vendors.
40:31 The watcher network, every watcher
40:35 that is an honest watcher
40:36 benefits from other watchers sharing
40:40 what they see in <i>duplicity</i>.
40:42 There's a huge incentive for watchers to
40:44 cooperate.
40:46 What we're doing for, what we're
40:49 calling <i>open infrastructure</i>
40:52 is "The Five Ws"
40:54 or <i>One-click KERI</i>. The idea is people
40:56 would come and say, "Okay, I want to use
40:58 KERI. How do I do that?" Well
41:03 Come to some meetings,
41:05 go read these repos, go do this, build a
41:08 bunch of infrastructure, then you can
41:09 use KERI. I'm sure I see
41:12 some people in here that have
41:14 walked that path. That's not a
41:16 really good story for adoption. The
41:18 idea is
41:20 we want to be able to tell somebody
41:22 One click,
41:24 it'll install KERI infrastructure for
41:26 you. You got a wallet, you got
41:27 witnesses, watchers, and now you can
41:29 start to build stuff with it. So, that's
41:31 the goal. So, the five Ws,
41:34 wallet,
41:35 witness,
41:36 watcher,
41:38 web,
41:39 so it's all on the web,
41:41 and wizard. And what wizard is the
41:43 catch-all term. We want <i>user interfaces</i>
41:46 that hide the complexity of key
41:48 management from the users.
41:50 And KERI gives you the ability to do
41:52 that. I mentioned some things that KERI
41:54 has built in that allow you to do that.
41:56 People building user interfaces need to
41:58 understand those so they can build user
42:01 interfaces to take advantage that. The
42:03 other wizard is <i>agents</i>.
42:07 You want interfaces to agentic
42:10 AI, and you want to control those
42:12 interfaces, you want to use KERI to do it.
42:15 So, this is what KERI infrastructure
42:16 looks like.
42:18 We've got a wallet on one side. Whoops,
42:21 go back. [organizing slides]
42:33 We have entity A has a wallet.
42:44 Entity B has a wallet.
42:47 Entity A has their witnesses they
42:48 control. They have watchers they
42:50 control. Entity B has witnesses they
42:53 control, watchers they control. There's
42:55 a shared watcher network. And now entity
42:58 A and entity B
43:00 can communicate
43:02 authentically.
43:05 But, they can't do it without
43:07 the watcher network and the witnesses.
43:12 (Now I can move on.) These are
43:17 early screenshots from
43:19 Mobile wallet that
43:22 that you'll see
43:24 demo-ed today. If you look at
43:26 the sessions, KERI Foundation team is
43:28 doing demos of the KERI stuff.
43:33 HealthKERI had developed
43:36 Locksmith wallet,
43:39 watchers and witnesses
43:42 for their proprietary products. And
43:46 earlier this year, they donated those to
43:48 the KERI Foundation and allowed us to
43:50 debrand them and rebrand them so that
43:53 everybody that wants to build, can build
43:56 on proven production-quality
43:58 infrastructure. And we're hosting those
44:00 repos and we have spun up -
44:04 we'll be providing some
44:06 witness and watcher networks on our own
44:10 servers so that people who want to test
44:12 it out can test it out running in the
44:14 cloud. That's one of the big
44:17 things that we're doing for the
44:20 community and
44:22 and I hope that makes a difference.
44:26 [applause]
44:33 There's two other pieces of
44:34 infrastructure
44:35 that are new;
44:36 that we're building this year:
44:39 Registrar and Observer. And the reason
44:42 these are important is that
44:45 they are how we do this concept what I
44:49 call
44:51 control over context.
44:54 So, I have issuances like SEDI
44:58 issuances.
44:59 They're ACDCs
45:01 and you have a registrar
45:03 and registries for like revocation.
45:06 And you have observers they get bulk
45:09 updates. The key here is the bulk
45:11 update. Bulk update means that the
45:14 observer
45:19 and the issuer can't correlate back to the point of
45:23 validation
45:24 because that's the point of use the
45:26 observer sees it
45:29 but a change in state by the issuer
45:32 can't be correlated forward because of
45:35 the bulk update and the observer's
45:38 not allowed and doesn't need
45:42 to communicate to the issuer because
45:44 they get a an update of all the
45:46 information that they need to do
45:48 verification. The verification
45:50 happens at the observer,
45:54 and observers can't collude
45:59 between each others if you're using
46:02 bulk issued credentials. So, we'll talk
46:03 more detail about that.
46:07 So, this enables bulk issuance. It takes
46:10 advantage of Sparse Merkle trees.
46:13 And so, that's some new
46:14 infrastructure. Another new
46:17 Use case, and these are in the spec,
46:20 these are in the ACDC spec, are things
46:22 called a <i>user presentation registry</i>.
46:25 So, one of the challenges
46:27 in verifiable credential world is that
46:30 if I have a proof
46:32 that a holder has that they're
46:34 the subject or the holder of a
46:36 credential,
46:38 and that proof gets stolen, and they
46:40 steal the user's keys, then that user
46:43 can be impersonated, and they can
46:44 present that credential, and the user
46:45 can't know that's happening or
46:47 doesn't have any way to know it. But, we
46:49 can use the same ...
46:52 (Go back.)
47:00 We can use the same
47:02 registrar mechanism
47:03 so that a user in the ACDC, when they
47:07 get it issued, can
47:09 designate
47:10 that their own registry
47:14 Is going to keep track of
47:17 blinded presentations.
47:19 And that means that a verifier won't be
47:22 able to verify the presentation unless
47:25 the presentation itself is anchored
47:28 in the registry. And this is all
47:30 blinded so that it's not correlatable
47:32 presentation to presentation. But, that
47:34 means that
47:35 issuees or users
47:38 who want to be ultra secure
47:42 can have fault-tolerant detection of
47:44 compromise of their keys with regard to
47:47 their own credentials. And they'll
47:50 know, just like:
47:52 you know that your keys for your KEL
47:55 been compromised because you can watch
47:56 your own witnesses. You can watch your
47:58 own registry to see that somebody else
48:01 is presenting your credentials
48:03 and you didn't do it. And so now
48:06 now you can take
48:08 recovery action.
48:11 So, what is KERI really?
48:13 Well, you've heard the phrase <i>Keys at</i>
48:16 <i>the Edge</i>.
48:21 <i>Security First, Always.</i>
48:24 <i>Minimally Sufficient Means. </i>These are
48:26 mottos, phrases; [another:] <i>Signed Everything</i>.
48:30 Sort of ways to look at the elephant from the
48:33 multiple blind men. [Don't assume one viewpoint captures the entire situation, red.]
48:36 So, all of these are features of
48:38 KERI, but if you only
48:41 thought of it as any one of these,
48:44 it wouldn't be KERI.
48:46 And lots of people have copied parts of
48:48 KERI,
48:50 but they've but not all of KERI. And if
48:52 you don't have all of KERI,
48:54 I guarantee you there are
48:56 vulnerabilities that you'll be
48:57 exposed to.
48:59 Heard a lot about that yesterday.
49:06 So, no more half measures, guys. It's
49:08 It's all KERI or no KERI.
49:10 [laughter]
49:12 KERI is
49:15 truly an identity revolution. There are
49:17 concepts in KERI that don't exist
49:19 anywhere else in the world in identity.
49:22 And I'm going to give you a short I
49:24 love creating acronyms because
49:27 I'm old and if I don't have a mnemonic,
49:30 I won't remember what I thought of
49:31 yesterday. So, this is how I remember
49:33 what I think from day to day.
49:35 Reputable Autonomic Pseudonymity. That
49:38 That encapsulates KERI in the smallest
49:41 set of words. Right? An identifier's
49:45 a pseudonym.
49:46 Is an autonomic pseudonym. What does
49:48 autonomic mean? Well, all of the
49:50 features of true individual control
49:53 over you know autonomic means self
49:55 governing individual control over your
49:57 identifier that's secure that's
49:59 persistent that's perpetual all of those
50:01 things right and reputable
50:04 means it doesn't go away.
50:06 There's something to stick to a
50:08 reputation isn't good unless it sticks
50:10 to something it has to stick to an
50:11 identity right? And then the other
50:15 mnemonic Control over Confidential
50:18 Context in which you use your identity.
50:20 You want to control the context that you
50:22 use the identity and that means that you
50:24 want the context to be confidential. Now
50:27 in the dictionary there are 20
50:29 definitions of the word privacy
50:31 or private.
50:34 About a third of them are synonymous
50:36 with confidential.
50:38 So when I say I'm sharing a secret with
50:40 you and I say it's just private between
50:42 us right?
50:43 That's not private that's confidential
50:45 right?
50:46 Because there's other meanings of the
50:48 word private that don't mean that.
50:50 The core private word means
50:53 intimate means only you know it.
50:57 Right? So a private key is never shared.
51:01 A shared secret isn't private.
51:04 Shared data over the internet is never
51:06 private in that meaning of the word
51:08 private but it could be confidential.
51:10 You could make the information
51:11 confidential.
51:13 So if you're going to interact with
51:14 people you want to interact in
51:16 confidential context and so you want to
51:18 make those contacts you want to have
51:20 control over that confidentiality. That
51:22 means control over what you disclose
51:25 control over what they do with it once
51:26 you disclose it to them.
51:30 So here's the thing RAP-C3
51:33 and if you say it really fast
51:35 it's a rhapsody* and that's KERI. [*an enthusiastic, emotionally expressive outpouring, red.]
51:46 That's it.
51:53 And we have a minute for questions.
51:56 – I think it'd be useful to explain
51:57 the difference between watchers and
51:58 observers.
52:00 So, watchers watch
52:03 the KEL. That's all they watch. They're
52:05 looking for duplicity in your KEL.
52:08 Observers are watching the state of
52:11 issuances that are anchored to the KEL.
52:14 So, KEL?
52:15 Via the TEL. Yeah.
52:17 So, it's a TEL observer. Yeah.
52:19 Yeah.
52:20 – Why do they need to be separated?
52:23 Well,
52:25 they have different purposes.
52:28 So, separation of concerns means you
52:29 design protocols so that everything has
52:32 Yes, it's layered. So, it's a layer,
52:35 right?
52:36 They depend on watchers,
52:39 but they aren't watchers.
52:41 Any more?
52:42 I guess we're out of time.
52:44 No, one more?
52:47 Four more minutes. Okay, I guess we
52:49 we started a little bit late. More
52:51 questions?
52:55 – Please be gentle. I'm Moises.
52:57 Be gentle. I'm a newbie at KERI. So,
53:01 Watchers witnesses. You
53:03 mentioned the term
53:05 KEL or TEL. Key event log.
53:08 No, TEL. But TEL is transaction event
53:11 log.
53:13 – Since there's an exchange of two parties
53:15 of information and keys,
53:17 how do the watchers attach to
53:20 confidential a confidential transaction?
53:23 Well, watchers aren't part of a
53:25 confidential transaction. Your key event
53:27 log is public.
53:29 You're publishing your key state so
53:31 everybody can verify
53:33 any attestation or assertion you make
53:36 relative to an identifier. And to do
53:39 that, you sign the
53:42 You sign things, and you have to verify
53:44 the signatures, so you need to know the
53:47 public key to verify the signature. So,
53:48 the Key Event Log publishes your
53:50 public key and then the witnesses and
53:52 the watchers, the witnesses make the
53:54 public key state highly available and
53:57 also give you a detection mechanism.
54:00 And if you were you at the SEDI
54:02 summit?
54:03 – No. Oh, okay. One of the things that
54:05 happens is that you want to be able
54:06 to detect that your keys have been
54:08 compromised. So, the witnesses are a
54:10 detection mechanism. Because if somebody
54:12 wants to publish something on your KEL,
54:16 they have to get your witnesses to
54:17 witness it, which means you get to see
54:20 it because what the witnesses witness is
54:22 public. And so, you can see that
54:24 somebody's compromised your keys in
54:26 order to publish to the witnesses. And
54:28 then the watchers are watching the
54:30 witnesses to make sure that the
54:31 witnesses don't cheat.
54:33 – Where is it published, if not
54:35 in ledger?
54:37 Well, a KEL is a ledger. It is a
54:41 micro ledger controlled by one
54:44 identifier. So, it's not a shared ledger.
54:46 It's not a shared ledger. It's a
54:48 non-shared ledger. But, technically
54:50 under the hood, it is a
54:52 blockchain. Yeah.
54:54 – But, how is it made public? I guess
54:56 that's my point.
54:58 I probably can't answer that in this
55:00 room. Okay.
55:00 But, if you're here for the conference,
55:02 you got 2 days of people who'll answer
55:04 that for you.
55:05 With that open-source infrastructure
55:07 Okay.
55:08 Yes, that's right.
55:11 Okay, thank you. One more Yes.
55:13 – Maybe you'll answer this later, but
55:16 witnesses are selected by the
55:19 controller,
55:19 and then the watchers are
55:21 selected by the verifiers. What's
55:24 your vision of how a controller actually
55:28 selects a witness?
55:30 Usually they will hire .., They'll either,
55:32 If they're technically savvy, they'll
55:34 spin up their own witness, but most
55:35 people won't do that. They will just
55:37 rent witnesses from witness services,
55:39 just like you would rent
55:41 a cloud host for your website.
55:43 – Okay, do you ever see individuals
55:46 being or employees being controllers, and
55:51 would they have to go hire someone?
55:53 The vLEI, if your work for company
55:55 has a vLEI and you get issued,
55:58 an ECR and OOR credential, you'll have a
56:01 wallet and you'll select
56:03 witnesses that you
56:05 probably your company spins up for their
56:07 employees to use.
56:09 – Okay, and then at the individual level,
56:10 do you ever see grandma having to go
56:12 select a witness?
56:13 One of the things
56:15 in the discussion that we've had
56:16 with the state of Utah is how do you pay
56:18 for disadvantaged populations to have
56:20 infrastructure?
56:22 And so, so there would be that would be
56:24 an entitlement that somebody would
56:26 pay for your infrastructure. But, it's
56:28 going to be the cheapest form of
56:30 infrastructure, which is web
56:31 infrastructure, not really expensive
56:33 blockchain infrastructure.
56:35 – I'm curious to know more
56:37 about the transactional friction
56:40 there or the just the friction
56:42 there.
56:42 a really good conversation to have
56:44 because
56:44 another conversation.
56:45 Yeah.
56:47 – A real quick one on the registrar
56:49 observer network, do you see benefit for
56:52 the registrar putting random
56:55 noise into the stream so that they can't
56:57 tell it's a correlatable event on a
56:59 revocation? Absolutely. The blinded
57:03 state TEL, which is the thing
57:07 that we use, not the current
57:09 version 1 does not have that,
57:12 allows you to do updates because they're
57:14 blinded that don't change the state,
57:17 they're just there to introduce
57:20 to whiten the signal. So, you can as a
57:22 registrar, you can whiten your signal.
57:24 You can you can issue random updates to
57:27 to registries that don't change the
57:30 state. And so, that means that a state
57:32 update doesn't necessarily mean the
57:34 state changed. It just means you did a
57:35 state update and most of them are ... and so
57:38 you could create
57:40 you could roll out 4 million
57:43 registries for birth certificates
57:45 and update them. And then when
57:47 somebody's born, the fact that there's
57:50 an update to a registry doesn't
57:52 correlate to the fact that they were
57:53 born and you then started using that
57:56 registry for their birth
57:58 certificate.
58:00 Thank you very much, Sam.
58:01 Time's out. Running out.
58:05 Do you want to introduce the next
58:06 speaker?
58:07 Yes,
58:08 That's Daniel.
58:11 Do I need to I need to get to move my
58:12 Please a big hand for Sam.
58:15 [applause]