Subscribe
Share
Share
Embed
The Starting Block is The Block's new flagship daily live show — the breaking news, the context behind it, and the people shaping where crypto goes next. Each episode is a 45-minute live broadcast hosted by award-winning journalist and broadcaster Gareth Jenkinson and powered by The Block's newsroom and research desk.
Welcome back to The Starting Block.
Happy Monday to you wherever you are
around the world.
It's week number five, I think we're
number six. No, I we're week number
six, so we're a
good six weeks into the start of the
show. It's been an absolute pleasure
broadcasting every day.
This is The Block's new daily show,
and if you're joining us for the
first time, my name is Gareth
Jenkinson. I'm your host and head of
multimedia here at The Block.
If you wanna catch up on some of the
content we've been putting out over
the last few weeks, you can go and
check out The Block YouTube channel.
Give us a subscribe
if you're there and watching us
today. And if you watching us live
on X this afternoon or this
morning, if you in the United
States, please like and read
broadcast so we can get as many
people joining us today for what
promises to be a jam-packed
show. As you can see by the name of
the show, we are gonna be dedicating
a lot of time today to talk about
the cold card exploit.
And if don't know what that is, I
will give you a full rundown once
we have our guests on the show
a little bit later today.
If you are here for the first time,
the show is very simple.
We start off with a news bulletin
that's called the Block Briefing.
We'll run down six or
seven of the top stories
in the news over the last 24 hours.
Obviously it's a Monday, so we'll
look at some of the stuff from the
weekend and then we'll jump into
the hot seat today.
We won't have a market pulse because
we're gonna be deep diving into that
cold exploit and what it means for
Bitcoin self-custody going
forward. Really, really excited to
have Zach Herbert from Foundation
and Jameson Lopp.
From Casa to talk
about Bitcoin self-custody and
the cold card exploit.
But before we do any of that,
I'm gonna give you a rundown of the
top stories you need to know this
Monday.
Starting off with the big headline,
and this news dropped less
than 15 minutes ago,
Michael Saylor's strategy continues
to sell Bitcoin.
A fresh SEC filing shows the
firm offloaded 1,638
Bitcoin last week.
For around $105 million at
an average price near $64,000.
That trims its total stack to
roughly 842,000
Bitcoin, still around 4%
of all the Bitcoin that will ever
exist, but now carrying close
to an $11 billion paper
loss at current prices.
My producer Jordan was teasing
us all fair a little bit earlier
once he saw this news come out, but
just to put this in perspective,
this is about 0.
1% of their total Bitcoin
stack that they've sold and
strategy says the proceeds went
towards preferred preferred
stock Distributions and buying
back its STR shares a
clear shift away from sailors long
running the buy only Playbook
and if you go and look at X
and strategies announcement, there's
plenty of negative comments there
But I am seeing a few Bitcoin
is that I follow quite closely
giving a dose of perspective here
and kind of calling
out strategies,
building up of their credit standing
as they build up a much
bigger US dollar reserve
to give themselves more than
a two year runway to meet
dividend obligations.
Moving on to our next story,
a big on-chain move raised eyebrows
while it's tied to Trump Media, the
company behind True Social shifted
2,628 Bitcoin
with around $165 million
to crypto.com over the weekend.
Spread across two transactions,
but the company says don't read
too much into it.
A Trump media spokesperson told the
block the Bitcoin was transferred
and not sold, the same explanation
they gave back in May.
The tag wallet still hold roughly
4,261 Bitcoin,
about $260 million worth
closely matching the collateral
backing its convertible notes as
of the first quarter filing.
Over in Washington, a regulatory
turf war is heating up.
The SEC has hit pause on
NASDAQ's PHLX's
plans to list cash settled Bitcoin
index options after CME
group petitioned the commissioners
to overturn a staff approval.
CME's argument here, these
contracts are commodity option swaps
that belong squarely under the
CFTC's oversight,
not the SEC.
The commission granted CMEs petition
for review and kept the
approval state.
Written statements are due on the
24th of August, later this month.
It's a jurisdictional
tug of war that could help shape
who regulates the next wave of
crypto derivative products in
the United States.
Then a sobering story out of Hong
Kong. Police say a local insurance
agent lost more than $3.3 million
after being lured into
a fake cryptocurrency investment
scheme through a romance scam.
It wasn't isolated.
It was the single largest case among
25 romance scams reported
in just one week, from July 24th
to the 30th, adding up to nearly
$9 million in total losses.
It's a stark reminder as crypto goes
mainstream, so do the scams
targeting everyday people.
Then prediction markets continue on
a tear.
Combined trading volume across
Polymarket, Polymarket U.S.
And Kalshi hit a record.
$50 billion in volume
was up 7.7%.
From June, according to the block's
data dashboard.
Continued momentum around the World
Cup helped drive that, but there's
a twist under the hood.
Polymarket US's volume surged
54 percent, while the main
polymarket platform's volume
actually fell 26 percent
compared to June, a sign that
regulated US-facing prediction
markets are pulling in more of
the action as the sector keeps
gaining steam.
And finally, South Korea,
but the country's second largest
crypto exchange.
Has officially confirmed it's
targeting a public listing in
2028.
In a notice posted on its
website, the exchange laid out a
multi-year preparation plan,
including a request for a
preliminary listing review in 2027
before pushing for the IPO the
following year.
It's a notable step for a major
Asian exchange moving towards public
markets, and another sign that
South Korea's crypto industry is
maturing fast.
We'll be watching how that roadmap
unfolds.
Over the next couple of
years. And that's it for the block
briefing.
I think we are gonna take a very
short break before we bring Zach
Herbert on and Jameson Locke a
little bit later on the show.
Stay with us as we talk about the
cold card exploit and everything you
need to know about what it means
for Bitcoin self custody going
forward. Don't go away.
We'll be right back.
Welcome back to the starting block.
We're jumping into the hot seat and
we're gonna dedicate two interviews
today to the hotseat segment.
It's a great pleasure to have Zach
Herbert, who's the co-founder and
CEO of Foundation, joining
us today to talk about the Cold Card
exploit.
And if you're joining us for the
first time and you don't know about
the cold card exploit, Cold Card is
of course a Bitcoin-only
hardware manufacturer that
was hit last week with an
exploit that...
As far as we would know right now,
could impact all of their devices.
And we've seen more
than a thousand Bitcoin drained
from certain models
of cold card wallets
in the past 36
hours or more.
It's been a concerted effort across
the Bitcoin ecosystem to,
first of all, make sure that
everyone knows that this is
happening, especially for cold card
users that have not been online the
last few days.
They might not even know, and I've
seen some harrowing stories as well
that. Some people went away on
holiday for a couple of days, saw
the news, they weren't able to get
to their devices. And when they got
home to their device or got someone
to actually open
up the devices and check their
balances, Bitcoin had
been drained. So a really dire
situation.
And we're gonna be talking about
that at length today with
Zach and Jameson Lopp, who's joining
us a little bit later on the show.
Zach, first things first, thank you
so much for taking the time to chat
to us today.
It's the first time we've spoken in
person, virtually.
So it's good to meet you.
I can imagine it's been a rough few
days, how are you doing?
I'm doing okay a
little tired and some of our
team is a little tire as well
But I mean compared to those folks
who lost money right and it's just
yeah and comparable and it has been
a really difficult day for
Bitcoin
Bitcoiners and Most hardcore
Bitcoiners right that I've been told
self-custody and that have helped
other self- custody So it's
definitely everyone's reeling from
the events of the last few days.
I don't think it's gonna
uh, stop either anytime soon,
we're going to be feeling the
effects of this potentially for
months, maybe even years.
And so, um, yeah, thank you for
having me on to, uh, talk about it.
Yeah, sure. And I think just for
our viewers out there that actually
might not be too familiar
with the story, maybe you can give
us your kind of TLDR
of what happened here,
a very high level overview before
perhaps we get into some of the
specifics.
For sure. And I'll try to keep it
brief so that you can ask me
questions, you know, because I don't
know how to in the weeds, right?
We want to get here.
But essentially a cold card,
right, is a, is a popular Bitcoin
only hardware wallet made by a
company called coin kite.
That's MDK or Rodolfo Novak's
company. You know, he's a pretty
huge name in the, in the
Bitcoin space.
And on Thursday afternoon reports
start to come out that funds were
being stolen.
It seemed like from folks
with cold card devices.
And so you can imagine the
entire industry scrambled to
figure out what was happening, what
was going on.
And the root cause that was
first disclosed, I think probably
blocked it, the best job, getting
on top of this and disclosing it
fast and writing up a report.
So many of us were kind of figuring
out in real time and
obviously checking that for all the
other hardware wallet companies,
checking that our customers were
also not affected.
But there was a bug introduced many
years ago into the cold card code
base in early 2021
that effectively reduced
by a significant factor
the amount of randomness
or we call entropy that goes into
the seed generation process.
And so what that meant is that
cold card seeds starting from early
2021 and even basically
through today or I should say maybe
through over the weekend, because
they released an update a
day or two ago,
which you could install and would
fix that ongoing on new devices.
But from 2021, basically,
until now,
all seeds generated on cold card
devices are insecure.
Now the level of
the security for the seed varies by
model, right?
So in this case, the mark three
model, which was the model that was
introduced around that time
a little bit earlier is the one
that's most affected.
And the thought is that
most of the funds have already
been taken from
those wallets and devices.
When we say taken, we don't
actually mean that you need any kind
of physical access to the hardware
wallet, right. And so that's what
makes this really difficult,
really acute.
It means that an attacker can
sit down at the computer, they can
write a script that can run on,
very often it could run on your
laptop, it could run on Nvidia GPUs,
they could be renting, you know,
compute somewhere, and they're just
cracking seeds.
And so they're going through seed by
seed, they're generating potential
seeds, and then they have scripts to
check, is there is
there money at this Bitcoin address,
right, that's that they're able to
derive that based on the seed.
And so.
Our best guess now is that there's
multiple attackers doing this right
this copycat once you hear about it
there's other bad actors that want
to jump on And um,
I know the counts the counts of our
said over a thousand bitcoin
Uh stolen I believe it's
now something like over 1700 from
what I can tell Um, it's very
hard to know exactly though.
So we're only going to figure it out
in the coming weeks and months
And even though it's not a huge
number of Bitcoin compared to some
of these mega exchange collapses
So I'll stop there, yeah.
Yeah, sorry, someone's ringing my
doorbell. I'm sure you can hear that
in the background.
Someone will get that downstairs.
Yeah, so I think let's take it from
there. Obviously,
once news of this had come out,
there was a concerted effort across
the industry to try and figure out
what this happened.
One of the questions I had
posed to me online, just in
discourse with other people was,
how much do you think AI tools
played in the eventual.
Figuring out of this
bug.
Do you think these frontier models
that we've seen come out in recent
months might have been the
key that has allowed some attackers
to figure out this
exploitable attack vector and go
through it and then obviously now go
through the process of generating
seeds and then running them through
all these addresses to see if they
can steal this Bitcoin basically?
My quick answer is yes.
Absolutely.
Some folks think that
maybe the reason that this happened
on Thursday was because
the latest Kimi model had just come
out some number of days before.
I think it was just a few days
before and these
models that are open weight models
mean that
there's no cyber security safeguards
and so anyone is
allowed to use them and they
will do a pretty good job
finding bugs and code.
We as foundation have access
to like GPT's cyber
program, right?
And so we're also able to run the
frontier models on our own code
or on other code, but.
As you know, new models are coming
out every month or every two months.
And in our experience,
up until recently, the Frontier
models have not been
the best at embedded engineering
work, meaning like firmware for
physical devices.
We've been running Frontier Models
that are our own code for months
now, and there's been so many
false positives.
If there's, a misunderstanding
of how our code bases work,
because typically you have
this firmware and it's calling on
other libraries.
And then sometimes the bugs exist in
the interaction between the hardware
and the software on the phone
or computer.
And AI still has a really hard time
holding all of that in context.
So I think we may have crossed the
threshold in the last few weeks
with some of these new models,
because I pointed GPT
5.6 sole, which is
the latest model from OpenAI.
At the cold card code base
that afternoon on Thursday.
And it was able to find the bug
without any issue.
Well, there was some
inflection point that took place
and there's a few lessons learned
but I do not think the takeaway
is that open source is bad.
My takeaway is actually that free
and open source software is
necessary because
If this had been a different
project in the space, like let's say
there was a common library
that means like some block of code
that multiple companies are
using and building off of.
And there's a lot of examples in
Bitcoin, like Block has Bitcoin
Development Kit, EDK.
There's like Rust Bitcoin libraries.
There's all these Bitcoin libraries
that are used by multiple companies.
They're free and open source
software.
I think that the likelihood of
someone catching that, if it's in a
commonly used library, is much
higher because of all these real
companies that are using the code
and sharing on it and building on
it.
In this case, this was a cold card
library that was
a set of code that was
publicly available.
So we say source available,
but not fully open source.
And my understanding, though I can't
say with certainty, but my
understanding is they were the only
company that was
using this specific library.
Now, obviously, this was just a bug
that was introduced back then.
There was no, it was not
intentional, right?
It's, there's a lot of
bugs that exist in any code base.
Every time we turn a frontier model
out of our code, it finds bugs.
The question is, like, what level of
criticality, right, is that bug?
Is it like a critical bug?
Is it a high, medium, low, and so
on? I do think AI is
part of the reason why this
happened. I also think it's part of
a solution.
And so it's just like any new
technology, right.
You have the good effects, you have
the bad effects.
If we try to create a thought
experiment about what happened if
the cold card code base was closed
source and this exact same thing
happened, which is totally possible,
by the way, his attackers, like they
reverse engineer binaries, they hook
stuff up in the lab.
You could have like a bad actor
within the company or at a different
company that has or even at
Microsoft who has access to all the
GitHub private repos.
So like, let's just say is the
thought experiment. If this was
closed sourced, we would not
have seen this immediate response
from the Bitcoin community.
Figuring out exactly what happened
and helping thousands of people get
to safety, we probably would have
seen like a fog of war for days or
weeks because no one would know what
happened or why it happened or what
the bug was.
So I'm still a huge open source
advocate, but this was definitely,
I think,
definitely had something to do with
the latest AI models.
Obviously, Foundation does
hardware wallets as well,
and we've seen lots of reactions
from different self-custodial
hardware wallet service providers
coming out into port,
but I can imagine internally, there
was a huge fear initially that,
like, holy shit, what happens if
this is something that is affecting
all devices?
How quick was it, or how hard was it
to rule it out and figure out that
this only the
specific bug that that was
impacting. Mainly the MK3s.
And how important is
it for people in the industry and
wider crypto natives
to understand that this is not
impacting all hardware wallets?
Because I've been seeing a lot of
people saying online that
self-custody is dead.
And I fundamentally don't believe
that, and I wanna push back against
that narrative myself.
Yeah, I'll answer that.
But one thing you just said, I want
to make really clear,
um, MK threes have been,
are, are most affected, but
all MK foreign leaders are still at
risk. They still have some better
entropy. So just like, if you have a
cold card, if it's MK three
or newer, I would, I will get your
funds out of there as fast as
possible. So that's, that's the
first thing I want to address, um
with self custody.
It's really, it's really dangerous
to say self custody is dead because
of this. I mean, I, I completely
disagree.
However, it's scary.
Obviously, every hardware
wallet maker, probably every wallet
maker in the last few days across
the entire industry has been, I
hope, turning AI loose
on their code base and reviewing
some of the most critical parts of
the code.
There's been some amazing Bitcoiners
that have been helping with that.
Like Rob Hamilton at Anchor Watch
has been using, I believe
Kimmy to submit bug
reports to virtually every hardware
wallets, software wallet.
We've been talking with him.
He submitted stuff to us,
there's one that scared us because
Kimi thought that we had a potential
bug. It was nothing like as
bad as the cold card bug,
but we had
to really dive into it with human
review and realize that it was
wrong. And so there's another aspect
here where there's
going to be a lot of panic and
uncertainty in the short term
because customers are thinking, is
my hardware wallet safe?
They're running AI models now
against their hardware wallets that
they're using, and sometimes AI
models are saying, oh, there's some
critical bug, but the bug doesn't
actually exist because the AI models
are still not at that level
and probably will never be at that
levels of perfection, right?
Nor are humans.
And so there's a lot of complexity
happening. So yeah, I mean,
obviously, the first thing I did
Thursday evening was
run the
Frontier models against our code
base, and our CTO did the same.
We breathed, you know, a sigh of
relief, right? That there was no
entropy bug. And I'm sure that exact
same thing happened at all the other
wallets.
Um, yeah.
Look, I think the
next kind of question that comes
into my mind now is the
notion of trust in Bitcoin, because
the old saying goes
very hard, don't trust, verify, and
a lot of people trusted here.
And I think that's the great tragedy
of this, is a lot people trusted
CoinKite, the
company, and NVK, and
I'm not here to bash them, I'm just
not that person. And no one could
have known, and it was very hard to
find this bike.
And no one found it for a very long
period of time, right?
But there is that element of
trust that most people don't even
realize is still there when
they're using a hardware wallet like
this. And it does make
me kind of wonder like how does the
industry address that?
Like this element of trusts because
it was supposed to be trustless,
right.
Like here's a hardware wallets,
now you can be the master
of your own coins.
And in this instance, that failed.
How do you feel about that?
Yeah, I think it's funny.
I'm running a blog post right now
that has to do with don't trust
verify to talk about all this
because I think what
happened is that um
The whole Bitcoin media and
podcast. I don't want to say the
whole but a majority of the Bitcoin
media podcasting community
has
Trusted
You know, I don't want to stay
blindly trusted this company I think
I think you know, they have good
intentions I do not think that
this bug was intentional.
MBK has been a Bitcoiner for a long
time, but
this don't trust verify is really
important.
And I feel like the Bitcoin
community should be leaning more on,
if you're going to trust a wide
array of technical, you
know, folks in the Bitcoin
Community, it shouldn't be
based around who
sponsors what podcast,
who's friends with whom, it
should be much more about
trusting technical leaders within
the community.
And I do think that the
media has to be careful, both within
Bitcoin and within all of crypto,
with recommending specific hardware
wallets or software wallets, which
is tough, right, because you want to
be able to take sponsorship money.
I mean, we've sponsored stuff before
and sometimes still do, but
it creates a really difficult
cascading effect.
And I know so many amazing.
Know, Bitcoiners who have large
audiences who are just torn up about
this completely, right, because
you recommend to so
many people a
specific device, you tell them to
self-custody, not
your keys, not your coins, don't
trust verify, you help so many
people, thousands of people, tons of
thousands of people on board to self
custody, and then this happens.
And so I do think that
there needs to be some introspection
within the Bitcoin community.
I don't know if that's a good enough
answer.
Like I wish I had a better answer to
your question.
I think the principles are still
there.
I just think that we as a community
need to really figure out what that
means. One thing I would say is that
like, what
are we looking for in terms of
verification?
Things like security audits are
really important, engaging with
the security research community.
Like we'll get bug reports all the
time. It's really important when you
engage with security researchers
that you do so with a great amount
of respect.
It can sometimes be an antagonistic
relationship, right?
Someone reaches out to us, they say
they found a critical bug.
Sometimes your first reaction as a
wallet company is to get upset about
that. And I think it needs to be the
opposite.
And then also just free and open
source software, collaborating with
each other, building on shared
libraries.
All those things I think are really
helpful. And when we say don't
trust, verify.
I think that's the kind of
verification that the space should
be moving towards, you know, less
about marketing
a specific company, more
about giving
the technical leaders and the
security experts a voice as much as
possible, especially in the coming
weeks.
Yeah Zack we've got
a lot backstage.
So I'll ask my final question to
you I mean just just in terms of
the the hardware wallet space, you
know We've seen kind of newcomers
come in in recent years I think you
know block streams Jade bit
key from block a few
newcomers that have definitely added
More hardware wallets to the
industry. I'm kind of curious.
Do you think that this situation
does call for for more?
Options in the industry and will
that kind of help drive things
forward? And from a usability
perspective, it does kind of beg the
question. It's like, is it still too
complicated? Does it need to be
easier than it is?
Because a lot of cold card users
were told like,
this will be safe.
You're fine.
But if you really want to go the
extra step, you need to do the
the coin toss dice
roll.
Option to go and generate a seed,
that's hugely complicated for the
layman who's looking to self-custody
their Bitcoin.
So A, do we need more
variety of hardware wallet
manufacturers in the industry?
And do we have a user experience
problem?
I think that more is not
necessarily the answer, though I'm
conflicted, obviously.
I think more free and open source
software and hardware options
is part of the
answer.
Some of those newcomers you
mentioned are open source,
BitKey is open source.
It was not fully open source at
first.
Many fiercely criticized that,
myself included, and they have now
open sourced everything.
Jade is open source, there's
many that are open source.
And so I think the open
source aspect is very important
here. And I think only
newcomers to the hardware wallet
world should be taken seriously if
their stuff is open source, is a
big takeaway that I want to
push as part of this moment
in conversation.
With regard to the UX, absolutely a
problem.
I mean, 100%.
And I've actually been trying to
think about how to react to that,
because I've been in a lot of these
Twitter spaces.
Trying to help some folks over the
last few days.
Um, there's been some incredible
Twitter spaces, uh,
one, one guy who goes by wicked
has been hosting one basically for
the last like four days straight is
extraordinary that he's been able to
do that, but everyone's coming
up on stage asking about dice rolls
and coin flips, and I don't think
that should take away here.
I think it makes sense.
If you're a cold card customer and
you, and you can't switch to another
device and you're like, how do I
generate a seed securely now?
What could I have done differently?
But for every wallet that we know
of, every mainstream hardware wallet
besides cold card, like the
C generation code has been
okay, right?
It's been okay.
And I'm sure Jameson who's gonna
be up next will talk about
multi-vendor, multi-SIG, which I
think is a solution.
There's some good solutions
that offer a good user experience as
part of that.
From my perspective though,
I think it really does come down to
making sure that you're
code is bulletproof for the C
generation as a hardware wallet
vendor, which is not that much good
to, you know, inspect and
trying to create a really great user
experience, both for onboarding and
backups.
You know, we use like NFC key
cards that we call them, where you
can tap to the device as an option
right now.
BitKey does some really interesting
stuff.
We're going to be, I think, further
improving our onboarding and
backup process because you want to
make it as easy as possible for
folks. And so...
I don't think the takeaway should
be do not trust all
random number generation after this.
I think that's the takeaway I've
been seeing on, you
know, these spaces and everything
because people are scared.
But I do think that this was more
about a horrible bug
with a specific vendor.
But if you are advanced enough and
you do want to generate your own
seed words, You can pick seabirds
out of a hat, for example.
Uh, you can enter them into the
device, a device like passport can
auto fill the last word for you,
like the 11th, the 12th or the 24th
word, where there's a bunch of other
ways that folks have been covering,
like you should do that.
Right. Cause it's empowering to be
able to do that yourself and not
rely on anyone else.
So I think that would be my major
takeaways, but if I get just the one
big takeaway, right.
Free and open source software I
think is, is our, uh,
longterm solution.
Zach, thanks so much for taking the
time to chat to us today.
And for anyone watching out there,
you can follow Zach on X and go and
check out Foundation as well if
you're looking for alternative
options. It's at Z Herbert on
X and at Foundation
HQ I believe is the X handle.
Zach, thank you very much.
We're gonna take a very short break
just to check Lop's
audio and visuals.
We'll be right back. Don't go away.
Welcome back to the starting block.
We are continuing our hot seat
conversation. We've got Jameson
Lopp, who's of course a CTO at
Casa, very well-known name in the
Bitcoin industry.
Jameson, it's great to have you on
the show.
First things first, I'd love to hear
your kind of recollection of events
last week, when you first saw
the cold card exploit
news and what your immediate
thoughts were and what you did in
the hours after that.
Well, actually and
oddly enough, I believe I
was one of the first people to
have a hint of it.
And that's because I had someone
email me privately through my
website in
the wee early hours of
the morning, because they were
actually over in Australia in
that area.
And they detailed that
they were a cold card user and they
had seen a partial loss
of funds and they described their
entire setup and their best
practices to me.
I was pretty baffled for several
reasons.
First of all, because it was only a
partial loss of fund and usually you
see full loss of funds when a seed
is compromised.
And at the moment,
I had just gotten back to them
and said, hey, we
need more information.
Need to better understand if
you have done something that could
have caused this.
And then of course, as the day went
on, we started seeing more and more
reports and the
only real commonality was that
these were all cold card users.
And so that's when the
security community really
woke up and said, okay, well,
we still don't know what it is,
but... Signs are pointing to
cold card and that's when a lot of
people started going and inspecting
the firmware more closely and
I think it was by the end of the day
that it was determined that
there was essentially
an entropy problem
you know with the random number
generator.
Once that was discovered then
really the goal was to
understand what the severity of it
was and you know
how Basically,
how long... People had before
they really could expect to get
popped, depending upon a number of
different variables of how they had
generated their seed phrase.
I mean, it's pretty crazy
how much Bitcoin has been stolen
since, but it's
kind of still ongoing, right?
And I saw a few anecdotes over the
weekend of people being
on holiday away from the cold card
devices, knowing that this has
played out and they can't do
anything, they can get to the
devices and move their funds
off.
Yeah, I guess my first kind
of question to you
is when you kind of saw that
the issue was with the random number
generator and the entropy involved
in mainly the MK3 devices,
but perhaps more of the cold card
devices.
You said to me, Orfea, that this is
something that one of the
sort of value propositions of
something like CASA is
ruling out this kind of
single point of failure.
I mean, I don't want to say, do you
think that this was kind of long
time coming that we would see an
issue like this with a hardware
wallet manufacturer?
But- Is it quite in your
mind rare that something like this
happened?
And do you think it sets a stake in
the ground where hopefully we don't
have this issue ever again with
any hardware wallet manufacturers?
It has happened, I want to say,
a dozen times before, and I
fully expect it to happen again
now.
The good part
about having critical
security failures like
this is that
the result is that
there is a sort of shared
learning.
And that it will result in
increased protections and
better practices taken,
but not by everyone.
And so, you know, the.
Various classes of security
failures tend to happen over
and over and over again because
even though our collective
understanding and consciousness
absorbs them and learns from
them, there will always be
new entrants in the future who come
in and don't learn the
hard lessons of history and end up
repeating them.
So, there's trade-offs
to all of this, good and bad.
And
I'm somewhat hopeful
that at least the
ecosystem-wide impact will
be less severe than what we've
seen from centralized exchange
providers that have gone bust
for various reasons.
I think
one nice
thing about this is that it is
somewhat distributed.
I don't know exactly the
size of Cold Cards
business and how many people bought
from them, but I think that
they are a lot smaller than
like Trezor, Ledger, whatever.
So on orders of magnitude impact,
it would have been far worse.
I think if this had occurred in
Trezr or Ledger.
But, and this is something that
we'll probably end up talking about
a lot in a variety of
ways, but
the size of the
organization and the project is
also somewhat
correlated to the security
of the project.
And so I think there's a strong
argument to be made that one
of the reasons why this
happened and went undetected for
so long was because
CoinKite is one of the smaller
hardware vendors.
And so, you know, didn't have the
same level of.
Eyeballs and other resources
focused on
auditing the code.
Now, we
know that people have looked
at the code over the years and
this particular bug was
fairly obscure.
It was not obvious I think just from
looking at the line
of code that was calling the random
number generator. And that's why it
was missed for so long.
And yes, I think AI
advancements have led
to this being found when
it was found.
Could it have been found earlier?
Sure.
You could probably play
coulda, shoulda, woulda
all day long,
but ultimately there is the
fact that advancements
in large language models are
drastically changing the security
landscape of everything.
This is something that
has been concerning.
To those of us in the security
community, really for the past
year, because we've been
seeing obscure
hidden vulnerabilities
in a variety of different popular
code bases, getting found,
some of which had been there for
five, 10,
even 15 or 20 years,
simply because...
You know, being open source is
not panacea.
It is not a guarantee that all
bugs will be found.
It's a combination
of being open-source
plus the level of scrutiny
and resources that are applied
upon that code.
And it's just that the LLM stuff
is it's
reducing the resource requirements
of applying level of screening
by orders of magnitude.
And, you know,
the the models that are
coming out on a monthly basis are
improving so
much that
it's
hard for the defenders to keep up
with the attackers because the
tool sets are improving for
both of them, but I think the
attackers are more motivated to
take that first mover advantage.
Yeah, I mean, it's been interesting
to see like cold kites,
I think, in one of the earlier sort
of announcements suggested
that it might be, you know,
AI related, obviously, we've
got these frontier models and
they're so powerful now that they
can go and search hard for these
bugs and then suggest ways to
exploit them.
But no one has really come out and
said like, this is probably how
it happened and this is who's doing
it. Our industry has obviously
been plagued by DPRK.
Hackers in the last few years,
would you expect sometime in the
next few months for us to actually
have a decent post-mortem that does
point a finger at someone with
the on-chain analysis that
we have today.
It's complicated because
obviously even though there was one
actor who was the first,
there's probably at least half a
dozen and like the number of
people who will be following
their footsteps and trying to
iteratively improve upon the
sophistication of the,
we'll call it search for
these keys,
that's ongoing.
And it
does sound like
the first attacker was
probably pretty naive
and didn't
fully understand the
intricacies of the protocol.
And we're probably doing all of
this by just like chaining LLM
prompts together.
And so that was actually, as
I mentioned at the very beginning,
one of the things that was baffling
to me was that the first person who
contacted me only had a partial
loss.
And it appears that the
reason for that is not because
the attacker
had only compromised some of the
keys, but rather the attacker was
unsophisticated in actually
doing their wallet
derivation and address
and funds search on Shane.
So they probably naively
had some LLM do
a fairly shallow search
for funds related to these
keys, and they were hitting various.
Roadblocks that were causing them to
stop searching the key
space of a given wallet, even after
all of the keys had theoretically
been compromised, you still have to
go generate
and do a search basically
in the UTXO set to figure out which
ones are available.
So yeah, I mean, that first one
or possibly few may have been more
naive and left enough of
a bread come trail that they might
get caught, but I expect that.
The later attackers are
just going to be more and more
sophisticated.
Lop, obviously, your website's a
wealth of information that I often
go to for reading and
recommendations on different things,
and obviously, you've got a whole
section dedicated to wallet
basics,
different hardware, manufacturers
that you would recommend,
et cetera.
The industry has always said that
self-custody is how you should do
Bitcoin, right? You know, you're
supposed to be banking yourself.
After this incident has happened,
you've been around a long time, so
you've seen worse.
I mean, I would say you've see worse
than this, but nothing like this has
ever really happened up until now.
How do you feel about, you know,
like self custody as it stands right
now, having seen this play out?
And do you think that we're getting
to a point where self custody really
isn't for anyone?
I don't wanna put any words in your
mouth. I just wanna get a sense of
where your head is at, just
considering the severity of this
exploits and what it might mean for
self custody going forward.
So although I don't think we've
had random number generator
vulnerabilities in hardware devices,
there have been many vulnerabilities
found in hardware and devices over
the years.
And there have many software
wallets and other,
I would say even more convenient
self custody
products that have
had random number generators
failures.
Self-custody is
for anyone
who is willing to
take on the responsibility that
comes with it.
So, you know, with great
power comes great responsibility.
This is true for self custody.
It greatly empowers the
end user.
And I think that
it's...
It can be a bit tricky,
like it may be too
easy for people to get set up
in self custody.
I know a lot of people are saying,
oh, self custody is hard.
Well, in order to
have a robust self custody with
no single points of failure,
yes, that does take some work.
And that's what I've spent over
a decade doing myself
is trying to make that easier for
people.
But...
It's deceptively simple to just
download a Bitcoin wallet,
a software wallet, and run it and
send money to it, and really have no
idea of all of the implications
behind that.
Um,
I think when one thing
that might.
Need further conversation
is, and
you were saying it earlier and I
think it's gonna be coming up a lot
in these discussions,
is the concept of don't trust,
verify.
I mean, obviously I think that
that's a good mantra and
that one should do
that whenever one can,
but you have to understand that
verification of
complex software and hardware
is simply not feasible
for 99.9%
of the population.
And I've had this idea floating
around in my head for a few years, I
actually have like a
draft blog post I've been working on
regarding it where,
ultimately what happens is because
it's just not feasible to
verify a lot of these things,
what we end up doing is
we end trusting.
Someone that that we suspect
has verified it, right?
And so you do end up with
various layers of trust,
which is understandable.
This is also in
a way how human civilization
works, like if you think of like
specialization of tasks, right.
So, you know, I think there
are a lot of people out there who
trust the various security experts
in the space.
Or just influencers in the space,
many of whom are gonna be getting a
lot of backlash for recommending
products that
resulted in people losing money.
But.
Perhaps it will be
easier if we just stick to
the don't trust aspect
and understand that you probably
can't verify.
And once again, and you know, this
is obviously shilling my own
product, but like the entire point
of CASA is to not trust
any one thing.
Not trust any one hardware vendor,
not trust anyone piece of software,
not even trust CASA ourselves.
And the main point that
I have here is that...
I understand there's
going to be a lot of hate and
vitriol and sadness
and other emotions and stuff over
this mistake.
I don't believe that
there's any evidence to show that it
was intentional.
I do believe that this
was an honest mistake.
And there will be good points
of like, well, this was their one
job and they failed at their one job
and that's completely fair.
But from an
end user perspective,
I believe it's
only fair to understand
that everyone is fallible.
Every company out there
is fallable.
Yes, like CoinKite has
been a smaller company with less
resources and so maybe more prone
to fallibility than others,
but ultimately there is
no such thing as
perfect software.
Perfect hardware,
perfect project.
Even Bitcoin Core, which is arguably
like one of the
most well engineered,
strongest, robust
open source projects
out there has had
critical vulnerabilities.
There have been critical
vulnerabilities over
the past year found in
low-level libraries like SSL
and SSH, like
code that runs on
billions of devices around
the world, and you could argue
is responsible for
the security of
many, many, or orders of
magnitude more time than
the entire Bitcoin ecosystem.
Open source is not panacea.
Like there, and there is no such
thing as perfection.
And so.
This is why I believe
that instead of
saying, OK, we're going to
trust any one given piece of
software or hardware, because we're
going to just verify the hell out of
it, whenever possible,
you should instead be
distributing your trust
across multiple different
diverse things.
Diversity is strength is one
of our mantras at CASA.
Because we expect that everything
can fail.
And so you want to put yourself in a
position where you
expect failure and you
are able to
recover from failure.
And so there's a million different
paths we could go down here.
And I was hearing and you were
talking about dice rolling a bit
earlier.
And yes, it's interesting that in
this particular case the
people who followed the
super high paranoid roll
your dice stuff were actually safe.
And, and I think it's,
it's somewhat ironic and unfortunate
that, um,
you know, I myself have
recommended rolling dice.
I have a blog post from three or
four years ago that I
entitled like how to generate
a secure seed phrase and,
and this is meant to be normie
friendly. And I said, basically
you want...
You want real-world entropy,
such as via playing
cards or throwing dice or whatever.
And in my guide,
I specifically say that the
easiest way to get a
super high security,
high entropy seed
phrase is to buy
casino-grade dice and buy a cold
card and use the built-in dice
rolling functionality.
And, as
far as we know, that claim
still ...
Holds true.
But I would note that,
you know, even that is
not a guarantee.
It's not panacea.
The reason for that is that
it would
have been and still is
entirely possible that there
is a bug in the
dice rolling algorithm
in the cold card.
Now, we...
As far as I know, that's not the
case. And people have actually been
looking at that over the past few
days.
And that code seems to be fine.
My point is, once
again, there is no guarantee that
that code is 100%.
There's no guarantee that any code
is a 100% perfect.
It just so happens that in this
case, that's the way to go.
Now,
going forward,
like if you wanted, for example,
to take kind of my perspective
of how you diversify your
trust since you can't really verify
any of the stuff, I
think the best way
to go forward would be to have
multiple different
means of
rolling the dice and
inputting that input and
and getting output from completely
different software and hardware.
That's how you verify that
it's unlikely to have bugs.
If you're doing the same thing
against multiple different software
and hardware, the
likelihood that they both
implemented the exact same bug
is very, very low.
And that's the same reason why
we are
highly supportive of
this idea of multi-vendor,
multi-SIG.
It's just because you expect
that anything can and will have
bugs, But.
As you diversify
your trust across many
different organizations,
then it's just a
sort of hedging game.
Yeah.
Talking about the hedging game,
obviously you've been at CUSA
a long time, you've been in the
security game a long, do you
think there needs to be a message
out there now considering everything
that has happened about A, best
practice for self custody, does
it need to be multi-seg,
where's your head at with regards to
that?
I mean, it's gonna be the same
thing that I've been preaching for
a decade.
It's very difficult to talk about
self-custody generally
because there are so many variables.
You know, I'm not gonna tell people
everybody needs to put
their Bitcoin into a
geographically distributed
multi-vendor multisig.
I believe that only makes
sense when you get to the point that
you have life
saving. Things.
When you have enough
Bitcoin that losing it
would be a sufficiently
large financial catastrophe
that you would never recover from
it, that's when you
should be getting to
the level of
paranoia where you're putting in
enough resources to ensure
you don't have single points
of failure, where if one thing goes
wrong, you have catastrophic loss.
But, you know, if you're just
playing around with Bitcoin, if you
only have a few hundred dollars
worth of Bitcoin, then yeah, leave
it on an exchange, put it
in a hot wallet.
If that blows up, then
it's not going to, you know,
negatively impact the rest of your
life.
If you have,
I generally say, if you have
thousands to tens of thousands
of dollars, That's
when it starts making sense to
invest a hundred or a few hundred
dollars into a hardware device
to get
your keys off the internet.
That protects you from
99% of
common loss
vectors.
But it's when you get beyond
that is when you should start
investing more.
I mean, generally just think of
it as.
And this is true for all types
of security.
The amount of
time, effort, and resources that
you invest into securing
a given asset should
be correlated to
the value of that asset.
So whether that's,
you're like insuring a house
or a car or
protecting the
physical security of your...
Family and loved ones,
you know, you should be spending
an amount on that that
is relative to how much you
care about it, how valuable it is,
and how irreplaceable it is.
I'm gonna quickly finish
off with a final question to you.
I mean, I know I think a lot of
people out there will naturally
gravitate towards some sort of
a custodian
or a custodeal solution.
CASA is still self-custodial,
right? Do you ever get people
knocking on your door and they're
like, I don't wanna use CASA because
I wanna be in full control and maybe
they don't understand that is
what it is.
And yeah, I don't know
if you want to just comment on the
other hardware wallets out there.
And I guess the coming together
of the industry, because it's been,
the last few weeks in Bitcoin has
been hugely divisive around, you
know, Bip 10, Bering in the
background. And this kind of seems
like everyone suddenly forgot about
that and got together and tried
to support each other as best as
possible as Bitcoin is first
and foremost.
Yeah. So, you know, the
way that I've framed
it is that,
um,
you
Diversity of
security properties and that
collectively added all together
uses this
really magic of
additive security properties that
puts you in a much more secure
position than you could ever be if
you only have one key of
any type.
But we're not reinventing
the wheel.
We're not doing anything novel from
a protocol point of view.
So it has always and will
always be possible for someone to
create their own multi-vendor
multi-SIG using nothing but
free open source software
that there's no subscription or
anything like that.
I think the reason you would
want to pay for
CASA Is
your your You're paying for
consulting, you're paying for
being given a
really a guided
software experience that
we
expressly do not support a lot
of advanced things that Bitcoin
supports because we think those are
foot guns that are too
easy to backfire.
So really, we're guiding
you through what we consider to be
the best practices.
And saving people a lot of time
from having to spend days,
weeks, months learning about all
of these best practices and
employing them.
And you're also, you're paying for
the CASA recovery key
that we safeguard and
build additional authorization
parameters around.
And there's plenty of other
functionality as well that
theoretically you could build
yourself, but we've been.
Building for eight years.
And so we've probably been able to
put a lot more resources into it
than any one person.
So optionality is great.
Just be aware that when
you are, if you're architecting
your self custody on your own,
it's problematic for
the same reason that these
other vulnerabilities have been
found, which like, as I
said, open source is not Patasia.
And it really comes down, I
think, to the principle of various
and numerous, which is strength
and numbers.
And when you're using CASA,
you're a system that a lot
of people are using that
we've had a lot
of engineering talent look at that
we get regular
security audits on.
And actually at this point, we have
our own AI pin test that
we're running all of the time.
So...
You're basically paying for
the expertise of a
security focused organization.
Whereas if you're doing it yourself,
it's highly likely that simply
because you're the only one looking
at it, you're just going to overlook
some potential vulnerabilities
and something over a long enough
time frame could go wrong.
Yeah, my final question to
you, I find this funny that,
you know, there was some influences
and stuff online being like, you
know, Bitcoin self custody is dead.
You're better off just, you know,
buying the Bitcoin ETFs.
Like my general take on that is
like, I fundamentally believe
Bitcoin is money and it will be
money that we all use
in the future.
And I want to own it and I want to
have total control of it.
I fundamentally believed that.
And I pushed back, you know, on
those assertions online, but Do
you think it's funny that people are
saying that online and is that just
completely misguided and we should
really still take Bitcoin
self-custody as the first prize and
really just be in control of
our own keys?
I mean the freedom of the protocol
is that people
can put
themselves into what security model
they want. They can choose whether
they prefer convenience
or security. That's really what this
all comes down to is that
if you want to
be empowered, if want
to actually
take full advantage of.
What I would say Bitcoin's
fundamental premise of not trusting
third parties is,
you gotta put in some work to do so.
If all you want is financial
exposure to Bitcoin as a
volatile asset, and
you don't
care too much about
the risks of
custodial failure, despite the
history of that happening in
the space, then you're
free to do that as well.
The the only thing that worries me
and the reason why I've spent the
past decade focused on self-custody
is That I do believe that
if we don't continue to iterate
on it and make it
easier for people make
people Believe
that they are capable
of taking on this Responsibility
then it will end
up being a niche thing that
only like super nerds
and paranoid folks are
going to end up bothering to do
and ultimately that is
arguably not great for
the fragility
of the overall ecosystem because
it would result in massive
amounts of Bitcoin being centralized
into a few trusted third parties and
then we're just recreating the old
system all over again.
Yep.
Completely agree with you.
I don't think I could have said it
better than that.
Jamison, thank you so much for
taking the time to chat to us.
I know you're probably busy on this
Monday with the amount of people
that are just looking to migrate
funds or cold card,
but salute you and every other
Bitcoin out there that's doing what
they can to help people in these
pretty tough times.
So Jamison we'll say goodbye to you.
Feel free to drop off the stream.
Thank you so for watching the
starting block today.
We broadcast live every day,
8.30 a.m. Eastern Time, 2.30
p.m., Central European Time, if
you're in Europe like I
am.
Please come back tomorrow.
We have Ophelia Snyder, who is
one of the co-founders of 21 shares.
She exited the company last year.
We're going to be chatting about the
institutionalization of Bitcoin and
many other cryptocurrencies.
So we'd love to have you join us
on the show tomorrow,
wherever you are around the world.
Look after yourself.
Look after your Bitcoin.
We'll be back tomorrow Until then,
bye-bye.