All right.
First a little housekeeping.
Is anyone in this room an attorney or a process server who is working on behalf of Craig Wright?
We might have one in here.
Well in that case, Jameson Lopp was unable to be here today.
My name is Mike Hearn and I'm going to be standing in on behalf of Jameson Lopp.
So Jameson put together 65 slides but was only given 15 minutes to get through them
all.
So I'm going to do my roasbeef impression.
But really what we're doing here today is presenting the results of some stress tests
on metal seed storage devices.
So I know the Bill Fodl guys are out there and we'll see how well they did.
They were a little unhappy with my stress test last year so I went back and redid some
of them.
What is our threat model here?
Specifically we are not going to be talking about the physical theft possibilities that
might happen.
That is kind of a separate talk.
What I'm going to be talking about is the potential for natural disaster loss due to
fire, flood, some sort of massive building collapse, what have you.
And so we're going to be testing heat, corrosion, and crushing.
And your particular use case or things that you might be worried about of course will
vary depending on your own living situation and the various hazards that you are exposed
to in your day-to-day life or in whatever situation you decide to place your metal seed
storage devices.
When we're talking about fire resistance, this is probably the most common thing that
people need to worry about, house fire, really any building that might burn down.
And we can see that the melting points of various metals are fairly well known.
And the vast majority of the devices that are on the market are made out of various
types of steel which usually melts around 2500 Fahrenheit, 1371 Celsius.
And the average house fire I think is only around 1100 Fahrenheit.
When we're looking at corrosion resistance, this is the galvanic scale which is really
more for rust resistance.
But there really wasn't enough time to wait around and see how many years it takes for
things to rust.
So my corrosion tests were actually very strong muriatic acid which may or may not be something
that you actually have to worry about but was nevertheless some interesting results
we got from that.
Finally, crush resistance, the hardness of metals is also fairly well known though even
within the steel alloys, there's a number of different types of steel and so the hardness
there can vary.
But let's get started with the contenders.
There were about 13 different devices that got tested this time around, more than double
of what was tested last year.
And most of them were probably around $30 to $50 in price though the priciest one was
the crypto tag there which was close to $300 US dollars but a very high end kit.
So we're testing the block plate which is just a very simple steel plate which you use
a center punch to put holes in to denote your letters and their positions.
We're testing the block stream metal of course, why not?
The cold bit pass phrase which is just a hexagonal piece of steel.
The cold bit steel which is a companion piece to the cold bit pass phrase but also really
just two steel plates.
Cold storage coins which are more of a decorative type thing.
They have your private key laser etched onto them.
Crypto tag as I said, a very high end kit where it comes with its own stamping template
which in my opinion is probably one of the most user friendly stamping devices out there.
Crypto vault which I believe is titanium but you can both etch and stamp into most of these
things.
One thing worth noting which hopefully you can tell, I tried a few different varieties
of etching and stamping where on the left the lighter stuff is a free hand where I was
just taking a Dremel and trying to write it whereas the better formed darker ones that
are large are using a template which I found allowed you to get a lot more pressure and
thus engrave a much deeper cut into the metal and this made a difference as you'll see later
in the test and then of course there's the regular stamping which is just taking a steel
bit hitting it with a hammer.
The hodl wallet is also quite similar to block stream metal which is the rebranded Bill Fodl
which is based upon an open source design called Crypto Steel.
The key walk capsule is another one of these more artistic devices.
It's a couple of steel rings that are encased in glass and other metals and they also have
the public and private key preset on them by the creator.
Seed Steel is really just two thin sheets of steel sandwiched between two thick sheets
of steel.
Interestingly enough, this company seems to have already gone out of business which is
unfortunate because it actually did reasonably well in the test.
Simbit is another style of tile based but it is not the same as the Bill Fodl.
It is not the open source Crypto Steel design.
It's rather using a much smaller set of tiles which are very challenging to actually get
into the device.
Steel disk is yet another one which it seems to be the same shape as the Crypto Steel.
At least the tiles and the rails are the same but it's a different configuration and then
the steel wallet is basically the same as the block plate, just very simple piece of
steel that you use a center punch to put your dots into.
So we start with the trial by fire.
One of the differences between the 2018 tests and these tests are that I got some feedback
from manufacturers saying that I was using a very small butane torch which had probably
only about half a centimeter width diameter on the flame and they argued that I was unevenly
heating the devices and causing warping that would not happen during a normal house fire.
I agreed with that assessment.
So this time around I used a propane torch that had about a 5 or 6 inch diameter, 2,000
degree based upon my thermometer, 2,000 degree Fahrenheit output on it.
And so these were able to heat up the entire devices quite well.
And the end results from the fire testing, as you can see, block plate, no real damage,
just a bit of discoloration and that's probably even like the oil from my hands that got melted
and screwed up.
The block stream metal also did very well, just a bit of discoloration but still quite
readable.
The cold bit pass phrase also pretty well.
You can see that the freehand written on the left, the freehand etching did not hold up
quite as well as the template etching in the middle and then the stamping on the right.
The cold bit steel got a lot of discoloration but was still quite readable.
The cold storage coins unfortunately did not do so well.
Even after I tried my hardest scrubbing away all of the detritus that was left on them,
I was unable to scan the QR code.
So have to consider that to be a complete loss of the private key.
Crypto tag, kind of hard to tell but it got a lot darker but you could still read all
the stamping.
Crypto vault, very similar, very dark but still readable.
HODL wallet, as you would expect, same as block stream metal, still readable, no damage.
The key walk capsule, unsurprisingly all of the artistic glass and other metal pieces
around it melted quite quickly but the important bits that the private keys were on did quite
well and were still readable.
Seed steel had a lot of discoloration but was still quite readable as well.
The SIM bit, though it's kind of hard to tell from the photos, is also still readable, had
no warping.
Unfortunately the steel disk had a lot of warping in the rails and as you can hopefully
see there, a number of the tiles fell out.
So we'll have to consider that a complete loss as well.
And of course the steel wallet had no real problems other than discoloration.
So next up, I got a big tub of muriatic acid which is generally used for etching concrete.
You can buy this at really any home improvement store.
Threw it in there, came back 12 hours later after most of the bubbling had stopped.
And of course the bubbling indicated that something was happening.
So what really happened?
Well, as far as we can tell, not much happened to the block plate, it was still readable.
Block stream metal unfortunately appears to have something going on with the rivets and
the rivets got dissolved, therefore the pieces of the rails separated and all of the tiles
fell out.
So it would be very difficult to recover your seed phrase after that happened.
The cold bit pass phrase, as we can see, the freehand etched side on the left got completely
wiped away but the other deeper engravings and stampings were still quite legible.
So perhaps if this was left in the acid long enough then the other ones would also disappear
but didn't have quite enough time to test all of that.
The cold bit steel, I believe the freehand etching had a bit of data loss on that but
the template etching and the stamping were still quite readable.
Just for fun, threw the cold storage coin in there, it didn't get any better.
Still not readable on the QR code.
Crypto tag did quite well and could read pretty much all of the stamps that were still on
there other than maybe one or two light hammer taps from the beginning.
The crypto vault, it definitely had data loss on the freehand etched stuff.
The template etched stuff was still legible as was the stamping.
And as expected, since these are same configurations, the hodl wallet seemed to have the same problem
with the rivets where the rivets dissolved in the acid and the entire device basically
fell apart.
Kiwa capsule appeared to have no problems whatsoever.
Seed steel also did quite well, was still legible on all the three different types of
imprinting.
And interestingly, the SIM bit seems to be the one tile-based device that survived it.
I guess whatever screws they're using are the same material that is also not susceptible.
It's very hard to tell from this photo but at the right angle you could still read the
letters that were on the tiles.
And of course steel disk got even worse as a result of having rivets dissolve.
Steel wallet itself, no problems.
So the only thing left to do was of course to get a 20-ton hydraulic press and see what
kind of deformations could be done to these steel devices and see whether or not it even
really mattered.
Now, is this something that the average person needs to worry about?
Probably not unless you plan on having a boating accident while you're above the Mariana Trench
or maybe if you keep your devices in some sort of abandoned mine shaft that collapses
or maybe if you're keeping it in like a vault under a very tall building that collapses
or something.
But these are very extreme edge case scenarios.
So the block plate, still quite readable.
I think you could probably bend this thing six ways from Sunday and you'd still be able
to read those center punch dots on them.
RIP of course to the block stream metal.
From my previous testings last year though with the crypto steel based tile designs,
they tend not to hold up to deformation because the rails separate and the tiles fall out.
Cold bit pass phrase looked pretty much the same except for a slight bend to it.
Little bit steel, very similar, still able to read the same things we wore after the
acid test.
Just for fun, the cold storage coin I felt would probably deform a bit and in fact it
left imprints on the press plate and even got stuck to the hydraulic press.
I had to kind of knock it off there.
Crypto tag also did quite well and I was able to read all of the data off of it.
Same with the crypto vault which of course had I believe lost some of the legibility
from the freehand stuff but it lost no additional data from deformation.
Rest in peace to the hodl wallet as well.
The Kiwa capsule, it really did not deform at all so whatever they're using for those
rings is highly resistant to pretty much everything and thankfully the rest of the stuff that
is decorative around it is not important to the integrity of the data.
Seed steel also did quite well.
Sim bit was actually one of the most surprising.
This is the only tile based device that I've ever been able to deform to this level where
the tiles still manage to stay in it and the explanation for that is probably just due
to how small the tiles are and the fact that they are more malleable than the crypto steel
type tiles.
So as you're deforming the device, all the bits of it are already so compacted tightly
together that the whole thing will just deform at the same time and you don't end up having
tiles popping out and flying away from the rails.
So I was quite surprised that there was no data loss to the sim bit.
Good job sim bit.
Steel disk of course we lost early on.
Steel wallet same as block plate did not lose any data as a result of deformation.
So these are my completely unscientific ratings that I came up with here.
For most people you probably only really care about the heat results and really pretty much
everything did well under the heat results but if you want to go the next step, if you
want to be incredibly paranoid and safe against the most extreme disaster scenarios that might
happen to your devices, then you probably want to go with one of these devices at the
top that shone through all of the stress tests with flying colors.
So what can we really take away from these results?
Well the KISS principle is definitely in play.
By keeping it simple, by having no moving parts, there's nothing really that can fail
other than the metal itself.
And metal is very well known as we showed at the very beginning, the properties of various
metals are very well known.
Basically the problems you start running into are when you're creating more complex pieces
of metal and putting them together in ways that then create potential failure points.
From a usability standpoint, etching and stamping is a real pain.
It takes a bit of learning to start to get it right and I definitely recommend using
a template for etching.
And for stamping, the best experience I had was with the Crypto Tag because they actually
come with a special stamping template where you place the Crypto Tag in there, then you
place another template to hold the actual bits with the letters on them, and then of
course you take that awesome hammer and just slam it down.
And you don't have to worry about crushing your fingers with a hammer, you don't have
to worry about hitting the steel bit with the hammer and then having the steel bit fly
across the room somewhere, which was happening to me quite often on all of the other metal
devices that I was trying.
And also from a standpoint of the block, plate, and steel wallet, that center punch just putting
a single dot in there takes basically no learning, it's very hard to screw that up.
And so if I were to choose an optimal setup for these devices, I would recommend using
something simple like a block, plate, or steel wallet to actually hold the seed, but then
because I would worry about physical theft, you would want that to be a BIP39 passphrase
protected seed, so then you have to BIP38.
So then you want to have some other data with the passphrase, and where are you going to
put that passphrase?
Well, that's when I think the cold bit passphrase really comes in handy, and so you probably
want to have multiple of these different devices for redundancy in multiple different places.
An alternative solution of course is go completely seedless, go digital, that puts you in a very
different type of security scenario, and you have to have special software to manage that.
But I think I managed to get through all 65 slides in 15 minutes.
So I hope you learned something, it was fun, and we'll do it again.