Sergey Nazarov and Ari Juels: DECO Acquisition and Ari Juels Chief Scientist for Chainlink Labs

(bright upbeat music) – Hi, everyone. Welcome to our conversation with Ari Juels and some of the great work
we'll be doing together. I'd like to once again, welcome
Ari as our chief scientist, and to say thank you to
him for working with us for these many years, and for
taking his immense intellect and experience and focusing
it now on the set of problems that we're working on. We're all very grateful to him and it's been great
working with him so far. And I think his ownership
and thoughtfulness being put into Chainlink even more is gonna really take us to the next level. So I'm excited to chat with him here. And the rest of the community, and with the rest of the community about some of our plans and
the things we're building. And it's, I think it's
gonna be very exciting. So, sorry. Thank you again for,
for working with us and the kind of the first
initial question I think is what is the nature of the
problem we're working on? Why is that an exciting
and worthwhile problem to be working on in your opinion? – Well, you know, I think the best problems
in computer science are the ones that are easy
to explain but hard to solve with rich layers of complexities
to you dig into them.

Consensus, of course, is the best known example
in the Blockchain community. At first to think how hard can it be to get a bunch of servers
to agree on a message or a sequence of messages? It seems pretty simple on the face of it, but it's exercise the minds
of top computer scientists for decades now. And then of course, in 2008, Satoshi Nakamoto dropped
this completely new approach on the heads of scientists. The Oracle problem in my view is similar. How hard can it be to relay
a correct piece of data from a few servers to a Smart Contract? Can't you just bang out
a person in 10 minutes.

But it turns out that
truth in computing systems is almost as challenging as in real life. What if the web servers disagree? What if the Oracle now disagree? What if there is an adversary present? You know, there's some overlap here, in fact, with the problem of consensus, and then there are a whole
bunch of other challenges, of course, like what if
we want confidentiality? How about the ability to
connect with Legacy Systems? Is high performance possible? As you know, my PhD, former
PhD students, Fan Zhang, who's now gonna be spending
the year at Chainlink productizing DeCo. I'm happy to say, basically wrote an entire excellent thesis on this very question. As for Oracle's importance. Now, when I first came across
Ethereum several years ago, it was clear to me that Smart Contracts, in order to do anything interesting would need access to external data.

With price feeds, DeFi has taken off. So just imagine what will
be next as we have a richer data ecosystem. And those Smart Contracts can draw on the full power of the internet. And this vision is the
reason that you and I, and Steve connected back in 2016, 2017. I think you had a similar view. And I think our views have
continued to evolve a bit, but in the main, the
story remains the same. – Yeah, yeah, I completely agree. I think I really see it as an evolution in that people have started to create systems that prove things
at a much higher standard in the format of a Blockchain.

And those initial incarnations
in Bitcoin and Ethereum were very important as far
as getting all kinds of value and initial use cases working. But then I think the place
that we really came together and even in our initial conversations, was the importance of how
these systems interact with various data that you would then
write a contract about. So I remember that that
was a very clear thought that both you and I shared initially. And I think it's a very logical
thought of yes, you know, these systems are very important
for storing contract state or sort of certain transactions, or maybe retaining certain
private key signing records and creating a globally shared record.

And I've always agreed with that. And that's a very good starting point for how this technology
kind of comes into existence and begins a foundation into which you would put certain data. But the question then does become, how do you expand beyond that, right? How does the space evolve
into something that works for the financial system
that works for global trade, that works for the insurance
industries of the world, which are all really
fundamental industries to how our daily lives are able to be the way that they are.

Just, I don't think it's
super clear to most people how dependent they are on all
these contractual systems. I think it will become more
clear to them over time, especially if some of them don't work the way they're supposed to. But it's definitely a worth problem. And in that sense, and I think that there's
definitely gonna be serious demand for the type of definitive truth that we're seeking to create, and the type of assurances that Blockchains and
Oracles together a combined can provide, right. I think that's a whole
new level of assurance about the state of the world, about the state of the
relationship that people have to data, to assets, to systems that inherently have very
strong impact on their lives. And once again, I'm really thrilled that
we're working together on it.

So, thank you once again. I guess the second question then would be what led you to become, you know, our chief scientist and
to take a leading role in our research program and to how and around how Chainlink will evolve. What led you to come to
working with us on all of this? – Well, I really enjoyed my
time working with Chainlink over the past few years. And particularly
appreciate the seriousness with which Chainlink
takes academic research. And the fact that Chainlink strikes what I think is a good balance between meeting the urgent needs of customers and investing in the critical, but all too easy to put off until tomorrow need for innovation. For me, you know, leading
the research program, along with Lorenz Breidenbach
is unprecedented opportunity, really to bring research
ideas and principles that I've been exploring in
academia to the real world. So thank you for that opportunity Sergey.

– Yeah, absolutely. It's our great, great pleasure. I think from our point of view, I think I've said it three times now, but we are just grateful to
be working with somebody of your caliber, and there's only a few people in the role that I think that can really
reason through these things, the way that I've seen you and
the people that we work with on the research team
really thinks through this in a detailed way.

I think the other thing
that really excites me about the research program is that, Chainlink has always really
been based on research. On some foundational
ideas that came out of the computer science world. And a lot of that was influenced by your thoughts on all of these points and how all of these
standing on the shoulders of giants ideas can now be composed and then expanded on and improved, right? And I look at the way that people build foundational infrastructure in our space, I really think that's the right approach.

I think the approach that I see sometimes where people either try
to make a web service, and say, Oh, it's fine. Don't worry about it. I'm just gonna make a web service from some web architecture that I know and call it a highly
secure cryptographically kind of proven system, right. And that's what I'm gonna do. Or sometimes I see people
copy things like Blockchains and just turn, try to turn
them into something else. I think a principled kind
of foundational approach to what is the problem we're solving and a rigorous system for
reasoning through that problem while standing on the shoulders of giants and combining their work
with new new additions we can make, is really what every kind of real technology
company should seek to do in my opinion.

And so I think the research program is going to really be a
combination of these things we've already been doing. And just a great expansion of
it and kind of a scaling up of implementing the
best possible approaches to enable secure data
transmission, data validation, various proof about data at this new, extremely high
standard of definitive truth and kind of the standard
that I think the world is gonna need basically
for relying on data and relying on systems.

And that standard is thankfully now due to a lot of the work you've done and other great people
like you can now be made to hit a much higher level than before. Which I think is coming exactly
at a time when people need, need reliability and need assurance that the systems they're
using actually provide them what they promise to provide them. And that's a very, I think, yeah… So it's a very exciting,
important thing to research and work through. And I'm thrilled to be
working on it with you. So Ari in the context
of the research program that we're working on together, what are the some of the bigger problems that we're working on and
working through at the moment and in the future.

I know we can't cover all of them 'cause there's a lot of
different ones we're working on. But let's maybe just cover a few and then focus on some of
the more immediate ones related to some of the
topics we've announced here at the conference. – Yeah, yeah. We're certainly working on a
variety of research challenges, but let me mention maybe
the top two or three. The first of course is scaling. Now the question of how
we get high throughput for Oracle queries and reports, and also how we achieve low gas costs. The community is working on
layer two scaling solutions for Blockchains, and there
are quite a number of them, diverse array. One of our challenges is to figure out how we're gonna draw on and interoperate with exactly these advances.

Second big challenge is
this notion of staking or crypto economic security
as people like to call it. I kind of prefer the academic
term mechanism design. This is another problem,
like the consensus problem, like the Oracle problem, that seems straightforward
on the face of it. It seems as though you
could just have nodes and make a deposit. And slash them when they
provide faulty reports. But it turns out to be much
more complicated than that. Previous works in this area for instance, haven't modeled adversaries
strong enough for our purposes. They haven't modeled for instance, the full range of
possible bribery attacks. So we're working with
experts in mechanism designed to ensure that our staking solution, once we roll it out will be carefully and thoughtfully designed and robust. Another problem that is really critical. And one of course, I'll be
talking about later today is the problem of confidentiality. People are today, I think so
sort of mesmerized with DeFi that they're willing to overlook
its lack of confidentiality in some critical regards. Addresses, trades, all kinds
of things are exposed on Chain. And I think users are gonna come to demand stronger confidentiality.

For enterprises, it's
an absolute requirement. Oracle systems need to
be part of the solution because Oracles of course
are handling sensitive data, both queries and reports. So those I think are the top
three research challenges were working on and we'll
continue to work on. – Yeah, yeah, I agree. I think scalability is something that we're gonna have to solve more and more because even if we solve scalability for putting data into layer one systems, we already have many layer
twos that want Chainlink data inside of them as certain
applications migrate to them. And then you start to see also a number of mission critical systems
from the web world and the existing enterprise infrastructure that have extreme speeds already, right? Because that stack is very well developed and they have a whole
different set of assumptions about speed and responsiveness.

We are seeing a number of use cases for Chainlink validated data beyond even layer one
and layer two Blockchains and into mission critical
web system processes, right? Whether that's at technology
companies or enterprises. And so I think that the
scaling of of Chainlink is it's progressing very well now. And it's progressing to meet the demands of our most immediate users, which are the DeFi users on layer one, soon DeFi various other
users on layer two, and then relatively soon
also certain tech companies and other web grid systems
that also need validated data. So scalability is gonna
be an ongoing thing that we need to continue to improve upon because the amounts of data that we're gonna want to
validate and kind of pass through into a state of definitive truth, into a state of being reliable
is gonna continually increase as the demand for
validated data increases.

So I completely agree that
that's a highly, highly needed, continued set of improvements. That I'm very glad we're
doing incrementally to meet the demands of our users. Crypto economic security and staking. I think you're right, that
it is a very nuance problem. And that it also needs to
scale with the value secured. So I think there are versions of staking that can work for certain sizes, right? And it's something that we
need to roll out in a way where not only do we provide the
technical foundations of staking, which I think are very
well along and clear in how some of those things will work. I think we also need to provide
certain amounts of guidance.

And that guidance then needs to evolve against market dynamics for how stake will be put up
against certain scenarios. And just like Oracles vary in the types of reports they give and the types of contracts they work with. The various staking
guarantees that they provide will also need to vary according
to different use cases. And so when you're trying
to make is in my experience, a framework that can provide
these for a number of use cases can also scale to meet
the security demands of those use cases as the
value they secure scales, and can do that in a way that's immune to all kinds of adversarial
kind of behaviors. So it is something that I think we'll be making work initially in a straightforward way.

And then it'll continue
to advance and evolve together with the help of our community and various users that help us even, even help us define what
are the proper requirements that people should seek to give in regards to stake and deposits and the mechanism designed
that they're relying on, right? Because a part of it is also
the user's comfort level and all those types of things. In terms of confidentiality of data and compatibility with existing systems. I think that's absolutely right. I think another place
where you and me really do come together very, very
clearly on all of our thinking and have had for many years now is the need for privacy preserving systems that enable the usage of
various valuable data.

And the interaction
between enterprise systems and Blockchains. And I think that really
has these two components of you need some kind of technology that creates confidentiality or a set of technologies. And then you need a system like Chainlink that's integrated into all
these enterprise systems and integrated into all the data sources to make that technology usable. And that's I think a very
challenging high standard to meet, but I think it's very worthwhile. 'Cause once you create privacy, I think you generate a lot of new data, premium data that can
go to DeFi contracts. You can make all kinds of user
credentialed data accessible without really revealing
it to Blockchains, but you can still have
it useful on Blockchains. And then enterprises are just generally very privacy conscious. Like all the enterprise
implementations I see their first selling point is privacy. We provide you privacy, right? And so if you can make
existing enterprise systems compatible with Blockchains while allowing them to
maintain their privacy, but still prove that the
different data that they have in their systems or the
different data they rely on from data sources is a certain way.

You give them the benefits of Blockchains while allowing them to use
their existing systems, which is exactly what
all of them want to do for the trillions of
dollars they have secured on those systems. I think this might be a
good segue to consider how you see DeCo playing into this. And to briefly give us
an overview of how DeCo could fit into this goal
of both confidentiality and kind of compatibility with things like enterprise systems. – Well, the way I view DeCo
fundamentally what it does is extend the reach of Oracle systems safely into private web data. This is important for enterprises. It's also important for users. I'll be giving a couple
of examples later today, but it's worth perhaps
mentioning those briefly now. One would be decentralized identity, this ability of users to create and manage their own credentials.

These credentials that
users are going to create are going to rely on
privacy sensitive data. For example, if you want to
prove that you are of age to sign a legally binding contract, you want to prove you're over 18. You're going to have to draw on an authoritative record
of your birth date. And that's clearly something you don't want to reveal
on the Blockchain. A confidentiality preserving Oracle system can translate a birthdate
into an attestation saying you are of age to sign a contract without revealing your
birthday to Oracle nodes without revealing it, of course on Chain.

And I think in general, it can be very helpful for bootstrapping decentralized identity. Or decentralized identity ecosystem. So that's one really important example. Then a whole slew of
examples revolving around connection of enterprises
with privacy sensitive data to Smart Contracts in finance, in trade. Talk about a supply chain
example later today, premium data to which users connect using access controlled system, using API keys, for instance, can be leveraged in an Oracle system only if confidentiality can
be provided for the API keys and possibly even for the premium data.

It may be possible only to publish digests of premium data on Chain, given the contractual
obligations that node operators have with the sources of data. So these are a couple of settings in which I see confidentiality as critical for enabling enterprise systems and users to rely on Blockchains for myriad needs that are unmet today. – Yeah, yeah. I completely agree. I think it's quite quite exciting when you have new categories
of data being made available to the internet or in
this case to Blockchains. I think historically what we've seen is that once you create a certain, either security or privacy
around certain types of data, it makes its way into applications.

And people can then use it
to build all kinds of things. The kind of canonical example
I always have in my mind is before HTTPS, I don't think people would
have really had e-commerce. I don't think you would have
really been sending around credit card numbers. People wouldn't have been inputting their personal credit card numbers that if compromised the data, they would be able to just
steal money from them. And then once you did have
encrypted connections for data, once you were able to maintain
a certain degree of privacy, you did have a huge kind of explosion in the applications that could
then use credit card data. And I don't even know
if people foresaw that. I don't know if people
first saw that, you know, HTTPS means credit card data,
means e-commerce, right. And I think this is a
similar type of dynamic for initially for Blockchains,
because Blockchains, I think the thing that
people don't understand when it comes to Blockchains
and data transmission, is that Blockchains
inherently seek to make this single source of truth transparent.

And therefore the
transparency seeks to create a public nature to the data, right? It seeks to create an auditable
public transparent record. And that actually extensive
and beyond public Blockchains and into these semi-private
consortium kind of Blockchains used by enterprises. They still have the same problem where they essentially
want to create a record, a shared record among all the parties. But the transparency that
that creates sometimes is worrying to certain members, right? But they still have data, or they still have some
relationship they want to have to that record. So now I think with DeCo, the fact that you can take entirely inaccessible, previously inaccessible categories of data, whether that's premium market data, whether it's data from enterprise systems, whether it's personally
identifiable data frame, AML KYC around the use of
something like DeFi, right.

By institutions or by other entities. With this data now being made available, I really do see a lot of use cases for it, both in the public Blockchain
and in the semi-private kind of consortium chain environments. I really think that people don't always, they don't always
appreciate the connection between how, when data
appears in a system, suddenly you can do things
that you couldn't do before. And I think this is what
really a devolution of DeFi is a testament to. And I think the ability to
put high end market data, enterprise data, personally
identifiable data, not on Chain, but into systems
that can then prove on Chain that the data is a certain way so that you maintain
the privacy of the data, but you also gain the use
of it in these Blockchains. You really do arrive at a much
larger landscape of things that people could do.

And I think that's really the goal that you and I have always had. And the goal that Chainlink labs has. And the goal that we all really have is how do we take this unique
method of defining truth about state or about
data or about the world and how do we combine it
into something that is unalterable, tamper-proof
and truly reliable for a global system of contracts. Ari, thank you once again, for
working with us and all this. I think it's a truly worthwhile goal that we're embarking
on fully solving here. And I'm excited we're doing it together with you and all the other great people on the research team. Thank you, thank you very much. – Thank you Sergey. It's been a real pleasure
working with the Chainlink team for the past few years. First-generation in the system. And I'm looking forward to the chance to work more closely with you
and the rest of the team to thinking through the next generation. (bright upbeat music).

You May Also Like