Latest / The Starting Block / Open source AI the new crypto? ft. Nansen CEO Alex Svanevik

Transcript

Open source AI the new crypto? ft. Nansen CEO Alex Svanevik · The Starting Block

0:06
Welcome back to The Starting Block.
0:07
Happy Tuesday to you wherever you
0:09
are joining us from today.
0:11
It's a great pleasure to have you
0:12
with us. If you've never watched the
0:13
show, this is The Block's new daily
0:15
show. We'd be going for just over
0:17
a month, started in
0:18
RBW a few weeks ago with a
0:20
very special interview with CZ.
0:22
Again, if you haven't watched that,
0:24
you can go and check that out on The
0:25
Block YouTube channel.
0:26
And while you're there, please
0:27
subscribe to our channel.
0:28
We're working very hard to get
0:30
you the best content in the
0:32
internet. And of course, if haven't.
0:33
Been paying attention to social
0:34
media for the last couple of
0:36
days. It's been a crazy few days
0:38
for the show.
0:39
We had Charles Hoskinson last
0:40
Friday, and that's definitely caused
0:42
a bit of a buzz with a lot of other
0:44
accounts sharing bit of an
0:45
altercation that we had.
0:46
Go check it out for yourself.
0:48
Make of it what you will.
0:50
We're continuing doing what we're
0:51
doing. And today I've got a
0:53
very, very exciting show.
0:54
A guest of mine who probably won't
0:57
give me gears if I ask him for his
0:59
founding story.
1:00
Alex Svanovic will be joining us a
1:01
little bit later. But before we get
1:03
into that.
1:03
I'm gonna jump into the block
1:05
briefing, which is of course our
1:07
big rundown of the top news that
1:09
you need to know in the last 24
1:10
hours.
1:11
Bringing up our first headline
1:12
today, Circle just made a big bet on
1:15
intellectual property.
1:16
The stablecoin issuer has acquired
1:18
core assets from IBM's blockchain
1:20
patent portfolio.
1:21
In a deal, Circle says, makes it
1:23
the largest blockchain patent holder
1:25
in the United States.
1:27
I saw this news come out late
1:28
yesterday.
1:29
And already asked the Circle
1:31
team if we can get an entire
1:32
list of that blockchain patent
1:35
because over
1:37
680 patents are in
1:38
that family and
1:40
there's been nearly a thousand
1:41
issued worldwide covering everything
1:43
from foundational blockchain tech
1:45
to banking, insurance, supply chains
1:47
and secure cloud.
1:49
Circle says the expanded IP position
1:53
backs its push to push
1:55
an internet financial system.
1:57
Including USDC, the Circle Payments
1:59
Network and its ARK blockchain.
2:02
In short circles, not just issuing
2:04
dollars, it's stacking patents.
2:06
I do think it raises a very
2:08
interesting question around the
2:09
ownership of these patents and what
2:11
they're taking off the market and
2:13
doubling down as IP
2:15
that they themselves own.
2:18
Obviously, a lot of people out there
2:19
on the internet think that
2:20
everything should be open source.
2:22
This is the direct opposite of
2:24
that. So definitely a story we'll
2:26
be watching in weeks to come.
2:28
Then a deal in the wallet space,
2:30
Kraken's parent company, Payward,
2:32
is acquiring Magic Labs embedded
2:34
wallet business.
2:35
It's structured as an
2:37
asset sale, and it comes as Magic
2:39
Labs restructures and rebands itself
2:41
to Newton Labs.
2:43
Magic's wallet customers will move
2:45
over to Payward services while
2:46
the remaining company pivots to
2:48
focus on its Newton protocol.
2:51
For context here, Magic Labs says it
2:52
created more than 60 million wallets
2:54
since 2018 and supports
2:56
over 200,000 developers.
2:58
So Kraken picks up serious wallet
3:00
infrastructure and Magic Labs gets
3:02
a fresh start under a
3:04
new name.
3:05
Then a big plumbing upgrade in
3:07
Ethereum staking, Lido
3:09
has started consolidating around a
3:10
fifth of all staked ETH,
3:12
more than 8 million ETH worth around
3:14
$16 billion onto
3:16
Ethereum's largest post-PECTRA
3:19
validators.
3:20
These upgraded validators can hold
3:22
up to 2,048 ETH
3:24
each versus the old 32
3:26
ETH cap.
3:27
Letting Lido's node operator secure
3:29
more stake with fewer validators.
3:31
The upshot here, it cuts Ethereum's
3:33
total validator count by roughly a
3:35
third, easing a load on the network
3:37
and pushes stake on the newest
3:39
validators past 50%.
3:41
And that's all part of Lido curated
3:43
module V2 rollout.
3:46
Then a story in the United States,
3:48
the NFL is stepping into the
3:50
prediction markets fight.
3:51
According to a July 27
3:53
letter reported by prediction
3:55
markets newsletter, the closing
3:56
line. The league is urging CFTC
3:59
chair Michael Selig to
4:01
tighten the agency's proposed rules
4:03
for sports events contract, arguing
4:05
they fall short on protecting the
4:07
integrity of its games and
4:09
consumers.
4:10
The NFL says its highest priority
4:12
is preserving game integrity and
4:14
that stable, orderly events
4:15
contracts depend on it.
4:17
It's a heavyweight voice weighing in
4:18
as prediction markets like Colshean
4:20
Polymarket push deep into
4:23
new territory.
4:24
Then some regulatory pushback out of
4:26
New York State Attorney General
4:27
Letitia James is urging
4:29
Congress to tighten the proposed
4:31
Clarity Act, warning it would
4:33
weaken states' ability to police
4:34
crypto.
4:35
In a written testimony to a Senate
4:37
subcommittee, James said the
4:39
bill seeks to interfere with and
4:41
preempt state investor protection
4:42
laws and would, quote, dilute the
4:44
ability to prosecute fraud.
4:46
She's calling for stronger
4:47
anti-money laundering ethics
4:49
safeguards and argues New York
4:51
has crypto regulation
4:53
for over a decade.
4:54
It's a notable state versus federal
4:56
flashpoint as market structure
4:58
legislation moves through
5:00
Washington.
5:01
Then a win for prediction markets.
5:03
A federal judge has blocked
5:04
Minnesota from enforcing its new law
5:06
banning them, letting Kalshi and
5:08
Polymarket keep operating in the
5:10
state while the case plays out.
5:12
Judge Catherine Menendez granted a
5:14
preliminary injunction on Monday,
5:16
siding with the CFTC Kalshian
5:18
Polymarket, Minnesota's law
5:20
set to take effect on August
5:22
1st. Would have prohibited
5:24
creating, operating, and advertising
5:26
prediction markets.
5:27
But the plaintiffs argued the state
5:29
was unlawfully trying to regulate
5:31
federally overseen derivatives.
5:33
For now, the ban is on hold
5:35
and the prediction market
5:37
boom rolls on.
5:39
If you're interested in reading more
5:40
about those stories, you can find
5:41
them on The Block's website.
5:43
That's theblock.co.
5:44
Go ahead and support our journalism.
5:46
We've got a fantastic newsroom full
5:47
of journalists working very hard
5:49
at their craft.
5:50
And of course making these block
5:52
briefing bulletins possible.
5:54
So big shout out to the team.
5:55
All right, that's it for the block
5:56
briefing. We're gonna move into
5:57
market polls and the hot seat
5:59
straight away.
6:01
So it's a great pleasure to welcome
6:03
on air, Alex Svanevec
6:05
from Nansen.
6:06
He's of course the co-founder and
6:08
CEO of Nansen and we
6:10
had the pleasure of meeting a few
6:11
years back, Alex,
6:13
at your offices in Singapore
6:15
in my first trip to token 2049.
6:17
It's been a crazy journey since
6:18
then. So.
6:19
First things first, how are you
6:20
doing?
6:21
Yeah, doing great.
6:22
I cannot wait to tell you
6:24
my founding story of Ethereum,
6:27
where I co-founded that blockchain
6:29
with 17 other
6:30
co-founders.
6:32
So this is going to be a great one.
6:33
I'm looking forward to it.
6:34
Okay, as long as you promise not
6:36
to shout at me on air
6:37
for asking the question.
6:39
I would never, I would not.
6:42
Look, I appreciate the words of
6:44
support and also just everyone out
6:45
there for saying so,
6:47
like just on the interview last
6:49
week, more than anything, I've
6:50
been doing this for a long time.
6:51
I worked on radio and TV in South
6:53
Africa as well.
6:54
I've done lots of interviews.
6:57
All I want to be is respected.
6:59
And I think that all
7:01
of the journalists and broadcasters
7:02
out there in this industry do do a
7:04
good job and we
7:06
all deserve a little bit of respect.
7:07
And if you don't want to onto
7:08
something.
7:09
Just saying, no comment, move on.
7:11
There's no need to verbally attack
7:12
people. I think that's the line in
7:14
the sand that needs to be drawn, and
7:15
it's not a lot to ask, you know?
7:18
I couldn't agree more.
7:19
Anyway, we
7:21
are
7:23
all on the same side, exactly.
7:24
It's great to have you on the show.
7:25
First things first, last month you
7:27
guys rolled out
7:29
hyper liquid trading on Nansen but
7:31
before we even get to that, I wanted
7:33
to talk to you a little bit about
7:34
the evolution of Nansen because when
7:35
we met in 2022, you know,
7:38
Nansen was kind of just focused on
7:40
like on-chain analytics and you know
7:42
giving people alpha to trade with
7:44
but it's completely evolved
7:45
since then and like what it is in 22
7:47
versus what it is.
7:49
Now is completely different.
7:50
Can you talk to us a bit about how
7:52
quickly that evolution happened and
7:54
why you went through those
7:55
iterations?
7:57
Yeah, absolutely.
7:58
I think a lot of people compared
8:00
Nansen to Bloomberg
8:02
back in the day.
8:03
They kind of thought of Nansen as
8:04
like an on-chain Bloomberg.
8:07
Uh, and to me, it
8:09
was always a very compelling idea,
8:11
uh, to think about what
8:13
if you could actually just trade
8:14
through the Bloomberg terminal,
8:16
right?
8:16
And so I always knew that
8:18
was going to be possible one
8:20
day once blockchains had
8:22
matured and we had the
8:24
infrastructure in place for
8:26
it. And, you
8:28
know.
8:29
Quite recently, we
8:31
kind of took that step into the
8:33
world of trading
8:36
and allowing our users to trade,
8:37
because at the end of the day,
8:40
most people who use our product,
8:42
they use it to trade and they use to
8:43
invest, right?
8:44
They come to Nunson to discover
8:46
which tokens they might want to
8:47
invest in.
8:48
And they come to perform due
8:50
diligence to figure out, should I,
8:52
should I buy this token?
8:53
Should I sell it? Should I just
8:54
ignore it? And we
8:57
thought the best way to
8:59
create, you know, what we
9:01
think is on the path to being the
9:02
best product in crypto
9:05
across any category within crypto
9:07
would be to just close that loop
9:09
and allow them to trade.
9:10
And there's another reason why I
9:12
think this is really important.
9:14
And that is, and maybe we'll get to
9:15
that later.
9:16
That's because when you start
9:18
looking at agentic trading,
9:20
you really want your trading agents
9:22
to be able to do everything full
9:23
stack. You want them to.
9:25
Look at what's happening on chain.
9:27
You want them to reason and do
9:28
research and you
9:30
probably want them to actually pull
9:31
the trigger, uh, at some point
9:33
as well, to, to
9:35
make trades on your behalf in some
9:37
capacity.
9:38
So that's kind of the evolution.
9:40
Like you could, you could think of
9:41
it as kind of executing
9:43
towards the same arc because
9:45
the mission has always been to
9:47
surface the signal and create
9:49
winners.
9:49
So now we are focusing more
9:51
on that on the second part of the
9:53
mission to create winners.
9:55
Um, and yeah, it's really
9:57
exciting. It's been, it has been
9:58
fun. We've, we've done a little over
9:59
half a billion dollars in trading
10:01
volume since we launched,
10:03
so it's good to see users,
10:05
uh, like it, but we still have a
10:07
long road ahead of us and a lot of
10:09
cool things coming out soon.
10:12
Yeah, look, I'm a big fan
10:14
of the pivot, obviously, and
10:16
just the evolution of the platform.
10:18
And like you say, if people were
10:19
coming there for alpha,
10:21
doesn't it just make it easier
10:23
if they could trade the alpha on the
10:25
platform as well?
10:26
So it's great to see this kind of
10:27
rollout.
10:29
Talk to me about the hyperliquid
10:30
integration, how long that took to
10:31
do, and how
10:33
interesting the last six weeks have
10:35
been. Because I can imagine it's
10:36
been crazy to see what people are
10:38
doing on platform on Nansen and then
10:40
just plugging that into hyperliquid.
10:42
Yeah, so our
10:44
first entry into the hyper liquid
10:46
ecosystem was actually to become a
10:47
validator, um, relatively
10:49
early on and, uh,
10:52
we've held the number one spot as
10:54
a hyper liquid validator from time
10:55
to time. I think we are in the
10:57
top three now besides the
10:58
foundation.
10:59
Uh, so that was actually the first
11:01
kind of contribution we
11:03
had to the hyperliquid ecosystem.
11:05
And then people, of course, wanted
11:06
to get hyperliquid analytics.
11:07
So we added our own analytics.
11:09
We added really great P&L tracking,
11:12
smart money tracking for hyperliquet
11:13
into the product.
11:14
And then of course adding
11:16
trading capabilities into the
11:18
product was kind of the, the final
11:20
piece of the puzzle.
11:22
Um, and
11:24
yeah, overall it's been, it's really
11:26
good. Actually people, people seem
11:27
to enjoy it.
11:28
I've personally been trading more
11:30
than I normally do through
11:32
our, through our product, uh, which
11:33
is fun.
11:34
Uh, to me,
11:36
what is probably gonna
11:38
have a bigger impact
11:41
is when we start rolling out
11:43
the autonomous trading agents
11:45
that we're working on.
11:46
We're doing a lot of R and D right
11:48
now on that front and
11:50
it's cool to
11:52
be able to trade agentically on
11:54
your mobile phone, people who
11:55
haven't seen that you should
11:57
download our mobile app and
11:58
literally just talk to our agent.
11:59
And you can sort of, you can trade
12:01
conversationally with an agent doing
12:03
research and.
12:04
Executing trades through the chat.
12:06
It's a pretty cool user experience,
12:07
but it's kind of a stepping
12:09
stone towards, uh,
12:11
autonomous trading agents, which,
12:13
which is kind of the full, full
12:15
self-driving analogy.
12:17
Um, and that's something I'm
12:19
extremely excited about, not just
12:20
for crypto, but the
12:22
ability, the ability to, uh,
12:25
be able to run agents that are,
12:27
you know, doing inference with
12:29
kind of, the frontier models or
12:31
open weights models and
12:33
then actually run.
12:35
Portfolio of sorts, I think is
12:37
a really compelling idea.
12:39
And, and I think a lot of us,
12:40
frankly, who are in crypto, we
12:42
have certain views
12:44
about the world and convictions
12:46
about markets, but we
12:48
also don't want to be glued to the
12:49
screen constantly.
12:51
So one of the things we, we
12:53
do at Nansen is we try to surface
12:55
the signal so that you don't have to
12:56
be glued the screen, constantly,
12:58
right? We want to make sure that
12:59
you're actually getting the alpha.
13:00
You're getting not just a bunch of
13:02
noise and a bunch of data, but
13:03
you're getting insights that you can
13:05
act on.
13:07
And I think once we get autonomous
13:08
trading agents out, that's
13:10
kind of taking that one step further
13:12
where you can express the strategic
13:14
intent.
13:15
And then you can kind of let the
13:17
trading agent go for
13:19
you without having to be glued on
13:21
the screen. Maybe you're out
13:22
bicycling or you're going
13:24
to the gym or you know,
13:26
out in the real world, which I think
13:28
would be quite nice.
13:29
So we think of all of this as like a
13:31
new way to trade.
13:32
We don't really wanna just recreate
13:34
the trading experience that you
13:36
could find in other products.
13:37
We're trying to create something
13:38
that's genuinely novel.
13:40
And we think our starting point of
13:41
having the best on-chain analytics
13:43
for traders gives us a really
13:45
nice edge.
13:48
Are there any platforms that are
13:50
currently doing this or doing it
13:51
better than what you've kind of got
13:53
like cooking behind the scenes?
13:55
Cause I mean, there's a lot of
13:55
centralized exchanges that have
13:57
touted on
14:00
like unplatform agentic trading
14:01
which might not be exactly what
14:03
you're describing.
14:05
No, I actually think every
14:06
centralized exchange and
14:08
every other trading venue in
14:10
the world is going to have to
14:12
grapple with agentic trading in
14:14
the months and years to come.
14:16
And my view has always
14:18
been that the centralized
14:20
exchanges, I think, are going to
14:22
struggle for a few different
14:23
reasons, although they are
14:25
definitely making investments into
14:26
this area.
14:27
The first reason is that they have
14:29
licensed businesses, which makes it
14:31
a little bit tricky to to
14:34
do.
14:35
Various things that you would want
14:37
to do when it comes to agentic
14:38
trading without getting into
14:40
too much detail on that front.
14:43
That's number one. Number two, they
14:44
have very lucrative legacy
14:46
businesses.
14:47
And I think agentic training is
14:49
quite disruptive in many ways
14:50
because it is fundamentally a new
14:52
way to trade.
14:54
So you see Robinhood is doing
14:56
agentic accounts, I think they call
14:58
them.
14:59
You have OKX, you have
15:01
Coinbase.
15:02
You have Kraken.
15:04
A lot of them are focused on kind of
15:06
agentic toolkits like CLIs,
15:08
MCPs, that kind of stuff.
15:11
The approach we're taking is that,
15:14
again, we use the
15:16
data and the analytics that we have,
15:18
which is unique.
15:18
We have more than 500 million
15:21
labeled addresses that we've
15:23
accumulated over the last six
15:25
plus years, which means that
15:27
the agents can literally see things
15:29
you wouldn't otherwise be able to
15:30
see on a blockchain.
15:31
Right. And that's a great starting
15:33
point. The other, the other part is
15:35
that we're doing a lot of R
15:36
and D in-house when it comes
15:38
to the performance of the trading
15:40
agents and I
15:42
tweeted, I think it was yesterday,
15:44
it was kind of tongue in cheek that
15:46
one of our trading agents, um,
15:48
made a $23 in profits,
15:51
but it spent $700 in inference
15:53
costs. And so that is kind
15:55
of, it's a, it' a funny, funny
15:57
situation to be in, but it shows
15:59
how like when you do R&D.
16:01
You know, sometimes funny
16:03
situations, um,
16:05
uh, occur, but of course the
16:07
idea is that you have trading agents
16:09
that give you a
16:11
better starting point than if you
16:12
were manually trading yourself
16:14
as a retail trader,
16:17
number one, number two,
16:19
obviously you have a runtime for the
16:21
agent that is incredibly reliable.
16:23
You know, a hundred percent uptime.
16:25
It basically just always works.
16:26
Number three.
16:28
We have scalable inference
16:30
economics behind the agents so
16:32
that we don't break the bank on
16:34
token spend, as
16:36
the example there shows.
16:38
And then number four, that we have a
16:40
delightful and entertaining user
16:42
experience around this, because I
16:44
don't think people in
16:46
trading and in crypto want to
16:48
just have an agent that they kind of
16:50
set and forget. I think people want
16:51
to be somewhat involved and maybe
16:53
see what it's trading, maybe get a
16:54
notification when it places a new
16:56
trade, like what is it doing?
16:58
So it's a really exciting
17:00
product area, maybe the most
17:02
exciting product area that we've
17:03
worked on. And I really feel like a
17:05
lot of our strengths kind of combine
17:07
together to create this unique
17:09
take on agentic trading.
17:12
Yeah, look, I mean, I've played
17:14
around a lot with Claude and like,
17:15
you know, building different things,
17:17
different workflows and stuff.
17:18
And, uh, it's very hands-on and
17:20
like kind of time consuming.
17:22
But once you get an agent to start
17:24
doing very specific things,
17:26
research, on-chain data analytics,
17:28
that kind of stuff, then it gets
17:29
exciting, right? Cause you can get
17:31
to the point where you're like, I
17:32
think there's an opportunity here,
17:34
help me figure out if there is, and
17:35
it might take a week or two for you
17:37
to figure this out, but eventually
17:38
it's like, yay or nay.
17:39
And if it's yay, then it can help
17:41
you actually figure out the workflow
17:42
to make it as automated as
17:44
possible, you know, because like you
17:45
say, no one wants to, no, one
17:47
wants be sitting in front of their
17:48
screen all day, if you think you've
17:50
got a great trading idea, and you
17:51
can actually make it work, if you
17:52
can automate it, well, that's
17:54
it, you know.
17:55
Exactly.
17:56
I think, I think people who've tried
17:58
doing that kind of stuff, for
17:59
example, you can, you can use our
18:01
MCP with Claude.
18:03
You can literally go to
18:05
app.natsun.ai slash
18:07
API, and you have like a one-click
18:09
installer for Claude desktop.
18:11
So you just plug in our Natsun API
18:12
key. And now you've given Claude,
18:15
you know, eyes on chain.
18:17
And people who have tried to do that
18:19
to your point.
18:20
I think they see that like, wait a
18:21
minute, I can kind of see where this
18:23
is going. Like right now, it's
18:25
not perfect. It messes something up.
18:26
I have to correct it. It doesn't
18:28
quite understand the significance
18:30
of this finding it, it
18:32
found on chain and so on and so
18:33
forth, but you can see where it's
18:35
headed. And so we're trying to skate
18:37
to where the puck is going and
18:39
the, the bet is that
18:41
we don't want to release like
18:43
slop trading agents.
18:45
We want to keep this in the kitchen
18:47
a little bit longer to make sure
18:48
that it's really high quality.
18:50
And then we want to give people sort
18:52
of a factory where you can
18:54
very rapidly back test
18:56
agents on
18:58
with our own actually pretty
19:00
incredible data platform that allows
19:02
you to place an agent back in time
19:04
and replay what the chain looked
19:06
like at that point in time.
19:08
And then once you've kind of
19:10
maybe tried 10 agents and then nine
19:12
of them failed the back test, you
19:13
take one and you put it to paper
19:15
trade. And then if it
19:17
paper trades well for a week or
19:19
something like that, you can
19:21
actually give it real funds and let
19:22
it trade. So that's kind of the
19:24
overall flow we're working towards.
19:26
And the idea is that you'll
19:28
be better off running this
19:30
through this kind of factory that
19:32
we are creating than just kind of
19:34
vibe coding some slop trading agent
19:36
that's gonna lose all your money.
19:39
I mean, my counterpoint
19:41
to that is like,
19:43
will there ever be a point of
19:44
diminishing returns when it comes to
19:46
agentic trading?
19:49
And basically, the beauty about
19:51
agents, right?
19:52
And like AI is that it
19:54
can have all the knowledge
19:56
of the internet and then
19:58
figure out, take the
20:00
knowledge and then apply it to
20:02
whatever market you're pointing it
20:03
at. So it can tell you of all the
20:05
different trading strategies out
20:06
there, we can probably execute this
20:07
and it could work.
20:08
If we all start doing the same
20:09
thing, surely at some point it's
20:12
like arbitrage trading.
20:14
Eventually the arbitrage takes the
20:16
trade away.
20:17
Yeah. So, so my counter take
20:19
to that is I think actually agents
20:21
will be more diverse than
20:24
human retail traders.
20:26
Uh, if you think about human retail,
20:28
traders, we are quite sheep-like,
20:30
you know, we'll see some meme
20:32
coin that Anson posts on X and
20:34
we'll all buy it.
20:36
Right. That, that is not a very
20:38
diversified or diverse,
20:40
uh, set of training strategies, uh
20:42
whereas agent.
20:43
Have a lot of different inputs,
20:46
right? They can consume more
20:48
different types of information.
20:50
There are a bunch of different
20:52
models. The models have different
20:53
settings, whether it's reasoning or
20:55
temperature or all these
20:57
types of things.
20:59
The agents have different harnesses.
21:01
There are different ways for agents
21:03
to use tools, whether it is MCPs
21:05
or CLIs.
21:06
And so, if you kind of add
21:07
up all of these different things,
21:09
and then in addition, you allow
21:10
users to express
21:12
some kind of strategy that they want
21:14
to run.
21:15
I actually think that trading agents
21:16
are going to be more diverse than
21:18
human traders.
21:20
So I'm not super concerned about
21:23
all of these agents doing the same
21:24
thing.
21:25
I think that might have been the
21:27
case when you think about trading
21:29
bots that don't do inference,
21:31
but they just execute an algorithm
21:33
or they
21:35
execute some have rules set,
21:37
but the trading ages
21:39
we see now.
21:40
Are run on inference, which means
21:42
that they are very flexible and very
21:44
powerful.
21:45
And by the way, there are many
21:46
pitfalls too, right?
21:47
Because you can basically
21:48
contaminate or poison the
21:50
inputs and kind of run different
21:52
attacks in that way,
21:54
which is something that we're gonna
21:56
have to
21:58
make sure, it's not an issue
22:00
for the users of this product we're
22:02
building.
22:04
But yeah, they're also, you know,
22:06
not going to be emotional in the
22:07
same way, but they will be easily
22:09
influenced, right?
22:11
So I think it's going to super
22:13
interesting.
22:14
And I would be very surprised
22:16
if we don't
22:18
have more
22:21
trading edges than humans, you know,
22:23
within two years.
22:25
Um, and, and of course it's gonna be
22:27
humans behind the trading edges.
22:28
But if you think of the analogy to
22:30
coding, Obviously
22:32
now most engineers are
22:34
running some kind of harness or
22:36
coding agent, you know, whether it's
22:38
through cursor or grok
22:40
or codex or cloud
22:42
code or,
22:44
you know, droid and so on and so
22:45
forth. I think the same thing is
22:47
going to happen with trading and
22:48
investing. It's just, it's
22:50
just going to give you such a big
22:51
advantage, but we
22:53
obviously need to get the tooling to
22:55
a place where it actually is a
22:57
genuine advantage.
22:58
Um, and we're not quite there
23:00
yet, but I think we're going to be
23:01
there within the next, you
23:03
know, months and
23:05
I think within the next two
23:07
years, we're gonna see a real
23:08
flippening in how people trade,
23:10
um, which is what we think is the
23:12
new way to trade.
23:15
Yeah, you posted online, I think
23:16
you said something like, you know,
23:18
open source AI is the new crypto.
23:20
And I wanted to pull on the thread a
23:21
little bit like the similarities
23:24
between the cryptocurrency industry,
23:26
probably between 2017
23:28
and 2020, when I think, you
23:30
know, I got into the industry around
23:31
then I think you've been here since
23:33
like 2015, right, something like
23:34
that. You were at an Ethereum
23:36
startup, you've seen it all.
23:38
But the evolution of
23:40
us figuring things out and then kind
23:42
of having to give it to people on
23:43
the platter.
23:44
Feels like it's happening again with
23:47
all the AI tools that we're
23:48
building, right? Because if you
23:49
really take the time and make the
23:50
effort, you can do a lot of things
23:52
on Claude, but what you're trying
23:53
to do now is again,
23:56
give to your users AI
23:58
agents on a platter and be like,
23:59
this is how it works, just talk to
24:01
it, and then it'll do what you
24:03
need it to do.
24:04
And for a long time, that's what we
24:05
had to figure out with crypto.
24:06
Eventually it was like, this is our
24:08
exchange and this is how you use it
24:10
to trade. It had to be very
24:12
ordered down for the plebs.
24:14
Yeah. So yeah, I tweeted that,
24:16
um, for a few different reasons.
24:18
The first reason is I see a lot of
24:20
the same people that
24:22
were active during DeFi
24:24
summer, now being active with open
24:26
source AI, uh, a good example
24:28
is Banteg for, for those of
24:30
you who don't follow him, you should
24:31
follow him. He's, he's awesome.
24:33
Um, and so I think it appeals,
24:36
I think open source appeals to
24:39
a specific type of person who,
24:41
you know, is intellectually curious,
24:42
likes to tinker.
24:44
Isn't afraid of some complexity
24:46
and kind of,
24:48
you know, seeing that this is
24:50
something that has great potential
24:51
and you might benefit
24:54
from being early, right?
24:55
I think that's kind of those are
24:57
some of the similarities.
25:00
And there's also kind of a
25:02
spiritual similarity
25:04
in that there is some degree of
25:06
decentralization, right?
25:07
People, obviously DeFi it's
25:09
in the name, it's decentralized.
25:11
Um, but even open source AI,
25:13
I think has a much more
25:15
decentralized approach, which I
25:17
think a lot of people like, uh,
25:20
and many
25:22
people have had concerns about AI
25:24
being a centralizing force, which I
25:26
think is true to
25:28
some extent.
25:29
But that's why it was so cool when
25:32
open claw in particular, uh
25:34
came out, but also, you know, you
25:35
have Hermes agent and a
25:38
bunch of other great open source.
25:40
Harnesses and tools, and then of
25:42
course, open weights models.
25:44
These are kind of almost
25:46
like giving power back to the people
25:48
or like back to people
25:50
who haven't raised tens
25:53
or hundreds of billions of dollars
25:54
to run some AI lab or something like
25:56
that.
25:57
And I think people really like that,
25:59
I think they spiritually connect
26:01
with that.
26:02
So that's probably another reason
26:04
why many of the
26:06
same people dabble with open source
26:08
AI.
26:09
Um, so, so yeah,
26:11
those, those are basically
26:13
the parallels I see.
26:14
And I just find it very exciting.
26:16
It's like, there's this saying, I
26:18
can't remember who said it might've
26:19
been Mark Andreessen, who says that
26:20
you should like invest in what the,
26:22
what the nerds are spending their
26:24
time on on the weekends and,
26:26
and, uh, it might have been Chris,
26:28
Chris Dixon or Mark Andreissen.
26:29
I can remember.
26:30
Um, but that's kind of what
26:31
open source AI is right now.
26:33
So yeah.
26:35
Is there a favorite open source?
26:37
I mean, I know Paolo at Tether
26:39
has been rolling out QVAC, but
26:40
obviously we had Ilya from Nia on
26:42
the show last week.
26:44
And I think they're
26:46
a protocol that's been pushing
26:48
really, really hard, you know, like
26:50
the envelope.
26:51
And if you look at Ilya, that's
26:53
kind of what he actually started.
26:54
You know, he was an AI engineer at
26:56
Google, you know?
26:57
And it's kind come full circle.
26:58
It's like what he needed, the
26:59
infrastructure he needed to build
27:00
was blockchain in order to...
27:02
Create the reality through which we
27:04
are now actually living.
27:06
Yeah, absolutely.
27:08
And that's super cool.
27:11
I mean, on our end, and
27:13
me personally, I use a lot
27:15
of these tools.
27:16
Some people are dunking on OpenClaw
27:18
now. They feel like, oh,
27:20
it's
27:25
I saw that.
27:26
I saw you in an interview someone
27:28
was asking you you give everyone an
27:30
open claw that joins Nansen now.
27:31
Is that true?
27:32
Yeah, that's right. That's right and
27:34
I would say probably
27:37
about a third or
27:39
maybe a little bit more 30 to 50% of
27:41
the company still uses their
27:43
open claw pretty much daily,
27:45
but not everyone uses like
27:47
some people don't really
27:49
like it for a bunch
27:50
of different reasons and
27:53
it doesn't solve all of your
27:55
AI needs but
27:57
it is really great.
27:58
It's like a sidekick that runs con
27:59
jobs for you that can like edit
28:01
documents and maybe
28:03
do a quick poll request, that kind
28:05
of stuff.
28:06
Um, so yeah, uh, we use a lot, we
28:08
also run open weights models in our
28:10
product. So if you, if you use our,
28:13
um, our agent either in
28:15
our web app or in the mobile app,
28:18
if, if use toggle to fast
28:20
mode, which is there's two different
28:22
modes, fast mode and expert mode
28:23
fast mode actually runs an open
28:25
weights model under the hood, um
28:27
currently is GLM 5.2, but
28:29
we've tested a bunch of different
28:31
models and we run evals on them
28:33
and kind of improve them for our
28:35
setup for trading and investing in
28:36
research.
28:38
And I wouldn't be surprised if
28:40
we end up switching out the
28:42
expert mode with an open weights
28:44
model not
28:46
long from now.
28:48
Right now, we're actually running
28:49
Opus. So we are
28:51
subsidizing Opus
28:53
tokens for our users.
28:55
But, uh, but.
28:56
But I think we might switch to an
28:58
open weights model.
28:59
And we've been doing like a lot of
29:01
fine tuning experiments and stuff
29:03
like that. So we're pretty deep
29:05
in the trenches on open source AI.
29:08
That's awesome.
29:09
Look, I think we've spoken about AI
29:11
at length. I wanted to chat to you
29:12
about a few different things.
29:13
And one of the threads I wanted to
29:14
pull on actually is like the staking
29:15
side of Nunson's business.
29:17
A couple of months ago, you said
29:19
that Nunson was the eighth,
29:21
number eight staking provider
29:22
globally across all chains.
29:25
How did that kind of work out?
29:26
I mean, why did you get into the
29:28
stake in business? And is it kind of
29:29
a core part of the
29:31
verticals at Nunson?
29:33
Yeah, there's a few different ways
29:35
to think about this.
29:36
The first one actually goes back to
29:38
the trading part, which is that we
29:40
know that most users of Dunson are
29:42
token holders, right?
29:44
They invest in tokens.
29:45
And so why not give them
29:48
the option to stay
29:49
with us? They already have tokens,
29:51
they're using our product.
29:52
So why don't you just stay with
29:53
Dunson? You trust us, use our
29:55
product?
29:56
That's number one.
29:57
Number two, we have
29:59
very good relationships with a lot
30:00
of great chains.
30:02
You know, Solana, Piper Liquid and
30:04
so on. We've integrated them into
30:05
our product.
30:06
We have, you know the best analytics
30:08
for a lot of these chains.
30:09
And so it was natural that we would
30:11
also contribute to their ecosystems
30:13
by running a validator for
30:15
their network.
30:16
So those were like the two main
30:18
reasons we did it.
30:19
And the way we did is we acquired
30:21
a company called Stake with us,
30:23
a Singaporean company that's
30:25
been around since 2019.
30:28
And they had built a really strong
30:30
team.
30:32
Very nimble, a great platform,
30:34
but they didn't really have that
30:35
much distribution.
30:37
So what we thought was, why don't we
30:38
fold this into nonsense?
30:40
We have great distribution, we have
30:42
a user base who can,
30:44
you know, stake with us.
30:46
And in nine months,
30:47
we basically 40x the
30:49
assets staked
30:52
with our staking
30:54
offering, which was awesome,
30:55
obviously.
30:57
We became the leading hyperliquid
30:59
validator after the foundation.
31:01
Uh, we've been climbing on Solana
31:03
and we actually just added a theory,
31:05
I'm quite late, but we've, we'd
31:06
added a theory mouse.
31:07
If you go to app.nansen.ai
31:09
slash stake,
31:11
you can stake all of these tokens
31:13
that I'm mentioning, uh, including
31:14
ETH now, which is a partnership
31:16
with Lido actually.
31:18
Um, so, so
31:20
yeah, that's kind of how it came
31:22
about and It's
31:24
a way for us to deepen the
31:25
relationship with our users and also
31:27
with the chains that we already have
31:29
very good connections with our
31:31
analytics offerings and now also
31:33
trading.
31:35
Yeah, exactly.
31:36
I want to pull on a few different
31:38
things here from tweets that you've
31:40
put out recently.
31:41
And one of the interesting things
31:43
that came out in recent weeks was
31:44
Balaji announcing that Network
31:46
States is going to move.
31:48
Yeah, well said.
31:48
I'm going to my late.
31:49
Yeah,
31:52
exactly. Well, I mean, I
31:54
think it was like a year or so ago
31:56
where you were also contemplating
31:59
pulling up your roots in Singapore.
32:00
Can you talk a little bit about the
32:02
environment there and what you think
32:03
about Pilaji moving and if it's
32:05
a sign of the times for that region?
32:08
I mean, it's a shame
32:10
is what I would say, first
32:12
of all, I think
32:13
at the end of the day, he's
32:18
a positive force for tech and,
32:21
you know, he would have been able to
32:22
create a lot of,
32:24
help create a lot of startups in
32:26
Malaysia.
32:26
So it is a shame in that sense.
32:28
I'm, I'm going to Malaysia literally
32:30
tomorrow morning, speaking there
32:32
at Malaysia blockchain week.
32:33
There are some incredible companies
32:35
in Malaysia, incredible founders.
32:38
Um, so I would just say that,
32:40
you know, I, I do think it's kind of
32:42
a lose-lose for both parties,
32:44
but, um, it looks like he's
32:46
got a pretty good setup now in
32:47
Kazakhstan.
32:48
Uh, I was checking flights to
32:50
Kazakhstan. There are no direct
32:51
flights from Singapore,
32:52
unfortunately.
32:53
Um, maybe he'll set that up for us.
32:55
Who knows?
32:56
Um, but yeah, so, I think at
32:58
the end of the day, it's a shame.
33:00
Um, for all parties to
33:02
be honest.
33:04
And I visited network school.
33:05
I thought it was super cool.
33:06
I was actually hoping to spend more
33:08
time there, but yeah, it is
33:10
what it is.
33:11
Yeah. Yeah, you you can't
33:13
change that.
33:14
One of the one of the other things
33:16
I wanted to pull on here as well was
33:18
your recent tweet about BitCub.
33:22
Oh, yeah.
33:22
And you know, the SEC
33:25
filing and the news that has come
33:26
out there. And this kind of also
33:28
just ties into, you know the whole
33:29
FTX debacle and what
33:31
Nansen used to be and all
33:33
of that. But can you talk a little
33:35
bit about the on chain stuff that
33:36
you see that has helped in
33:38
some ways with some of these big
33:40
criminal enforcements over the
33:41
years?
33:42
I think actually maybe the first
33:43
time we met was around the FTX
33:46
times in
33:48
2022.
33:50
And yeah, the kind of interesting
33:52
thing about our product is that you
33:54
can use it to see what's happening
33:56
in real time, right?
33:57
On a blockchain.
33:58
And we have more than 500
34:00
million addresses labeled, so
34:02
you can kind of know, oh, this is a
34:03
BitCub wallet.
34:05
This is FTX.
34:06
When FDX happened?
34:08
And Sam was saying that he
34:10
halted all withdrawals.
34:11
I could see on chain that he was
34:12
lying. I could literally see
34:14
that he was sending out funds
34:16
to other wallets.
34:17
And I was like, what's going on?
34:19
And it turned out that it was the
34:20
Bahamas KY seed
34:22
accounts that he was letting through
34:24
because he was afraid to go to jail
34:25
in the Bahamas.
34:26
So he just like said, send out
34:28
all the funds to the Bahama's, um,
34:30
the bit cup situation, I will
34:32
admit I have not studied it
34:34
now after the
34:36
news came out.
34:37
But back then, I did notice
34:39
that there were some flows.
34:41
Um, I can't recall if it was
34:43
in, I think it was out of the
34:45
hot, one of the main wallets of
34:47
bitcub and there
34:49
was a lot of interactions with
34:51
tornado cash.
34:52
And I just found that very curious
34:54
and I reached
34:56
out indirectly.
34:57
I don't want to kind of disclose who
34:59
I was talking to, but I reached out
35:00
indirectly, I didn't hear back.
35:03
And so in hindsight, maybe I should
35:04
have pulled that thread a little bit
35:06
more. There's one more case
35:08
I want to mention, which is a HODL
35:09
knot, which was kind of a
35:11
block fire here in Singapore.
35:13
And that's when I could see on chain
35:15
that they were doing some
35:16
shenanigans with the stable
35:18
coin deposits.
35:19
And I actually confronted the
35:21
founder. I WhatsApped him because I
35:23
was in a group with him and he gave
35:25
like very shady,
35:27
flaky answers.
35:30
And in hindsight, it turned out that
35:31
he was basically yoloing.
35:33
The user's funds into
35:35
anchor, I think on Terra and
35:37
lost hundreds of millions of
35:39
dollars.
35:40
I could literally see that on chain
35:41
and the difficulty
35:43
is like, I don't
35:45
always want to just call out
35:47
people and say, Hey, what the
35:49
hell is this? Because there might be
35:51
some reason for it.
35:52
That's kind of the tricky thing with
35:53
on chain. And it's not always
35:54
totally unambiguous.
35:57
But in that case, I did actually
35:59
call them out.
36:01
And with Bitcub, I reached out more
36:02
discreetly. With FTX, obviously, I
36:04
called them out, and so occasionally
36:06
you see some of the stuff on chain.
36:08
And I tried to be transparent about
36:10
it, but I also want to make
36:12
it clear that you don't
36:14
always know the full story just from
36:16
seeing what's happening on chain
36:18
Yeah. And I think that's like,
36:19
that's the game that Zach XPT plays,
36:21
you know.
36:23
Um, and I think he does a pretty
36:24
good job, but, uh, like three weeks
36:26
ago, we had Terrence Kwok from
36:27
Humanity Protocol.
36:28
And obviously after, you
36:30
know, that hack happened, there were
36:32
a lot of people on, on Excel.
36:34
We're kind of like, man, this is,
36:35
this is a straight up rag pool.
36:36
And, you know, and I asked him on
36:37
air, like, people are saying that
36:39
you, you stole the money here, what
36:41
happened? And, uh he, he kind of
36:42
pointed a finger at Zach XP T and
36:44
he was like, sometimes it's not
36:46
good when you pull the trigger and
36:48
cry wolf when you actually don't
36:49
have the full facts because on-chain
36:51
can tell you that funds are moving
36:53
but you don't have the whole picture
36:54
and sometimes
36:56
it is not quite fair to just
36:58
claim things that you don't I don't
36:59
know.
37:00
I think, I mean,
37:02
people will feel differently about
37:03
this. I think depending on the
37:05
situation, the right thing to do is
37:07
like, Hey, I see this on chain.
37:09
Can you explain it?
37:11
You know, and it's like, that's
37:12
fair, especially if it's a, if it is
37:14
a centralized exchange and you're
37:15
dealing with, you know, user
37:16
deposits.
37:17
Like, yeah, you should be able to
37:19
explain, Oh, we're moving these
37:20
funds to like another wallet.
37:22
In fact, Coinbase is
37:24
very tidy about this, they've
37:26
like published in the past
37:28
of these. I don't know if they still
37:29
do it. They published ahead of time
37:31
that we're going to move funds to
37:33
a new wallet architecture, so
37:34
don't panic if you see funds moving
37:36
on chain.
37:37
That's, that seems like a reasonable
37:39
thing to do.
37:40
So like, why not just accompany,
37:42
um, a movement like that
37:45
with an explanation either ahead of,
37:47
or roughly at the same
37:49
time, there was another thing with
37:50
multi-coin where they unstaked some
37:52
hype and they complained
37:54
about, you know, this is why we need
37:55
privacy, but it's like, I
37:57
mean, if you're not going to well.
37:59
Just say you're not going to sell
38:01
and you can do that ahead of time.
38:03
You know, it just looks a bit, I
38:05
don't know. It looks a little bit
38:06
odd.
38:07
Um, and, and by the way, like people
38:09
can do what they want, but then just
38:11
don't complain when people observe
38:13
things on chain and ask questions.
38:15
That's like totally reasonable.
38:18
I mean, when I was still at
38:19
Cointelegraph, the
38:21
social media team was often watching
38:23
stuff on chain and Vitalik's
38:25
wallets are very openly
38:27
labeled.
38:28
So whenever he was moving funds
38:29
around, it's like Vitalik is moving
38:31
another $10 million of
38:33
ETH. What's going to happen?
38:35
It's part of the beauty of the
38:37
transparent nature.
38:38
But again, the back and forth
38:40
between do we need more privacy now?
38:42
The Ethereum foundation as well
38:44
as like the, that's like a meme,
38:46
right? That's like Ethereum
38:47
foundation selling.
38:49
Yeah. But I mean-
38:50
they got very unhappy with us on
38:51
that as well.
38:53
Joseph Schweitzer would always
38:54
message me like bro you guys need
38:56
to calm down like at least ask
38:57
us for comment if you're going to
38:59
publish something about this
39:00
because.
39:02
I mean, come on,
39:04
like, do they not know that people
39:05
are tracking the wallets?
39:06
I mean come on.
39:08
They must know.
39:09
And so I think like, it's
39:12
possible to have two thoughts in
39:13
your head at the same time.
39:14
It's like, yes, we need privacy for
39:17
consumers, I think,
39:20
for a bunch of different reasons.
39:21
You're never going to use
39:22
blockchains to do, you know.
39:24
Shopping and that kind of stuff.
39:25
If everything is revealed on a
39:27
blockchain, everyone's monitoring
39:29
it. That seems kind of unlikely.
39:30
So you do need privacy, obviously.
39:33
But I think you also need
39:34
transparency.
39:35
If you go to a Bloomberg terminal,
39:37
you can see who the top holders
39:39
of shares are.
39:41
And it's not exactly the same
39:43
thing, obviously, there is a
39:45
difference. But I think transparency
39:47
is also good, especially for
39:50
retail investors.
39:52
It's, it's fair to
39:55
ask questions if one of the top
39:57
holders of the token is moving funds
39:59
and like, Hey, what's going on
40:01
here? Like, can we just, can you
40:02
explain, can you calm the markets?
40:04
Like, I think that's totally
40:05
reasonable.
40:06
And I do think we should
40:08
think of public blockchains
40:10
that are open and
40:12
transparent as having a feature and
40:14
like lean into that rather
40:16
than trying to make them
40:18
be something they're not where we're
40:19
like, Okay, should we just like
40:21
close our eyes when you make
40:23
a transaction?
40:24
Like, I mean, it's there, right?
40:26
We can see it.
40:27
So, so yeah, I think it's
40:29
possible to think like, yes, we need
40:30
privacy, but like, we also deserve
40:33
transparency to some extent as
40:35
retail investors, uh,
40:36
owning some of these assets that are
40:38
on public blockchains.
40:40
So yeah, that's, I, I it's
40:42
not like black or white.
40:43
I think people, um, most
40:45
people with common sense can see
40:47
that.
40:48
Yeah, for sure.
40:49
Look, we've had you for a while
40:50
Alex. I'm gonna I'm going to let you
40:51
go very short either one question I
40:53
did want to ask you. I remember
40:54
chatting to you last year when you
40:56
rolled out, you know Solana
40:58
analytics on on nonce and it's been
41:00
a long time since then when you
41:02
see meme coins like Ansem kind
41:04
of launching And then pumping what
41:06
kind of stuff do you see on on
41:08
your platform? And like what's your
41:09
general thoughts about this meme
41:11
coin trend happening again?
41:13
Do you think Ansem's kind of revived
41:15
the trenches trenches and
41:17
Might be brighter times ahead for
41:18
Solana. I mean.
41:20
This is not investment
41:22
advice or predictions, right?
41:24
But it is true,
41:26
empirically, that he has revived
41:28
the trenches to some extent.
41:31
And, you know, a lot of these meme
41:33
coins have like 15 minutes of fame.
41:35
It's like, I see it.
41:36
There's like a Holland meme coin
41:37
when he's playing a match and it's
41:39
like okay, it disappears.
41:41
Um, but with Ansem, it kept showing
41:44
up. You know, it was like every day
41:45
it was on the front page of Nelson
41:47
and it's, like, okay, that's
41:48
interesting.
41:49
And the approach he's taking to
41:51
distribute Ansem tokens
41:53
to people who are seemingly loyal,
41:55
I think is like an interesting way
41:57
to do it. So I'm all for
41:59
experimentation and
42:01
he seems to be doing it in a very
42:03
transparent way as well.
42:04
I mean, um,
42:07
you know, it's all, it' all on
42:08
chain, right? So you can see
42:09
everything he's doing and he's
42:10
talking about it a lot.
42:12
So I think it's interesting and
42:14
he, it, it does seem like he has
42:16
revived.
42:17
Some the trenches to some extent.
42:20
Obviously, we're not back to like a
42:21
full blown meme coin season.
42:24
At this point, I
42:26
think it's probably more likely that
42:28
people are going to be trading
42:28
stocks and,
42:30
you know, perps in stocks
42:32
in the form of perps.
42:34
In the next months and years.
42:37
But I think it's a it's good.
42:39
It's a good attempt.
42:40
And you can see on chain that
42:42
it's having some kind of effect way
42:44
more than any random meme
42:46
coin.
42:47
Yeah, for sure.
42:47
The last question you touched on it,
42:49
obviously, RWA tokenization
42:51
securities on chain.
42:52
It's massive now, like Robinhood
42:54
chain launched a few weeks ago and
42:56
quadrupled
42:59
Ethereum's base chain fees.
43:01
They're making a lot of money.
43:02
Do you ever think about your
43:03
platform now and the
43:05
possibility that in three or four or
43:07
five years time, the mainstay of
43:08
your users might be a load of TradFi
43:10
people that are looking for chain
43:12
alpha on, you know...
43:14
Mainstream stocks and shares that
43:16
are now tradable on chain 24-7.
43:18
Yeah.
43:19
First of all, I want to take this
43:21
chance to say that we actually just
43:23
added Robinhood to the platform
43:25
so you can get analytics on
43:27
Robinhood, you know, smart money,
43:28
all that stuff, uh,
43:30
both in the UI on the mobile app and
43:32
in the API.
43:34
And later this week, I think maybe
43:35
tomorrow or the day after we'll
43:37
switch on trading as well.
43:38
So you can trade on Robin to train
43:40
through Nelson.
43:41
Um, one interesting thing we
43:43
have seen, it's funny, you say like
43:45
in four or five years, uh.
43:47
Because when we look at what people
43:49
are trading, on Anson in
43:50
terms of perps,
43:52
if I look at the top 15 perps
43:55
that are traded on Anston,
43:57
like 10 of them are non crypto.
44:01
So the majority are non-crypto.
44:03
I don't know this, it's a little bit
44:04
sad almost, right?
44:05
Because we are in crypto, but
44:07
people are trading, you know, SpaceX
44:09
perps S and P 500
44:11
gold, silver,
44:14
crude oil, Brent oil.
44:15
So,
44:17
interestingly...
44:18
It kind of seems like we're already
44:20
here.
44:21
Like it's already happening.
44:22
It's already happening.
44:24
Like people are, I mean, our,
44:26
our vision for the future is that
44:27
every asset is going to be on chain
44:29
and it actually feels like
44:31
we might be at the tipping point of
44:33
that now when we will just have
44:35
a deluge of assets
44:37
coming on chain, either tokenized
44:39
or through perps.
44:40
So I think this is super exciting
44:41
and it's going to take on
44:43
chain, you know, a hundred thousand
44:45
X.
44:46
Make it 1000x bigger than
44:48
what it has been.
44:49
So I'm super bullish on it.
44:51
I'm super bullish on it as well and
44:52
I'm surprised that more of the like
44:54
exchange founders that I've spoken
44:55
to haven't doubled down on the,
44:57
we will be the everything exchanges
44:59
and we will eat TradFi because I
45:00
think that it's, I actually think
45:02
it's inevitable.
45:03
I mean, the product vision we
45:05
have with Nansen is that you're
45:07
going to be able to trade everything
45:09
on chain with agents.
45:10
It's very, very simple.
45:12
Trade everything on chain,
45:14
with agents, and that
45:16
means we have to make a shift from
45:17
analytics only to
45:19
trading as well.
45:21
We have to make a change from crypto
45:22
only to every type of
45:24
asset. And we have to make it shift
45:26
from humans picking the
45:28
tokens to agents actually taking
45:30
the tokens.
45:31
And when you put all that together
45:32
you get this very simple phrase,
45:34
trade everything on chain with
45:36
agents. And that's what we're
45:37
building towards.
45:39
Beautiful.
45:40
Next time I have you on that'll be
45:41
the title for the show.
45:42
Alex, thanks so much for coming on
45:44
the show and chatting to us.
45:45
Luckily, we got through about an
45:46
hour with not fighting with each
45:48
other. So one day forward.
45:50
I appreciate it, man.
45:52
It's good to catch up.
45:53
My pleasure. Likewise.
45:54
And thank you very much out there
45:56
for joining us on the starting
45:57
block. We will be back again
45:58
tomorrow. We've got Sergey Kunz
46:00
from Oneinch to join us.
46:02
They actually released some news
46:03
today, so we'll be unpacking that.
46:05
And the rest of the week is also
46:07
looking pretty good.
46:08
I've been working hard to try and
46:10
get Jack Mallers on the show to
46:12
chat about his departure from 21,
46:14
and hopefully we can get him on in
46:15
the next couple of days as well.
46:17
But keep it locked on our socials to
46:18
hear who's coming up next.
46:20
We will be back tomorrow.
46:21
Until then, look after yourself,
46:23
and we'll see you then.