You leave your funds on a hardware wallet and find out that it has a supply chain compromise. You leave your funds on an exchange and find out that it gets hacked. Every choice is a tradeoff. But generally, the more you're willing to work to verify, the better the results will be.

Replies (10)

Jose Sammut's avatar
Jose Sammut 2 days ago
People over complicate some things too much. I wouldn't say more work is better. Just a little bit is all you need.
Your funds are always on the blockchain They are never in a wallet. You can never take the funds off the blockchain. What people call a wallet is more like a keychain that holds keys that grant you access to ask for permission from the nodes to move funds around the blockchain.
Yeah, there's no zero-risk option, just different kinds of risk. The difference is that with a hardware wallet you can actually check it yourself. With an exchange you're just trusting their word.
I think these mistakes are often made due to people thinking their funds are "on" a hardware wallet or exchange. Those are better considered as holding the keys to the doors unlocking the "funds", which simultaneously exist everywhere and nowhere. Bitcoin UTXOs are Schroedingers' Money, they don't exist any 'where' for the to be on or in. That terminology itself comes from the physical world's paradigm, while the digital realm behaves very differently. The existing Gatekeeper mentality can't imagine a world without fences. So, they keep playing fiat games and act surprised when they win fiat prizes.
I suppose it is a bit more complicated than that. Hardware wallet makers recommend buying direct, but that creates centralized data pointing to who owns these things that can attract goverment snooping, phishing, litigious harassment, and $5 wrench attacks. I imagine many of the people who bought these aftermarket wallets were worried about such threats. Some of them may have rationally preferred losing coin over what a government can do to them. Coldcard made source available, but not free software in terms of the 4 software freedoms. End users had an incentive to check the source, but they mostly lack the technical chops, leaving only bug bounty hunters whose payoff was uncertain. One of ESR's insights into why free software is so good is that the freedom to use and modify puts more competent programmer eyeballs on the code, and proportionally more effort on improvement relative to re-implementing what could be re-used. Very few of us will ever examine source code ourselves, and it is unrealistic to require that, but the /cultural/ impact of software freedom is fundamental. Disclosure (as in "open source") is a necessary, but not sufficient condition for software freedom. People who trusted Coldcard on the basis of equivocating open source with free software made a category error. The term "open source" is a handy shorthand, but not really what you are looking for. We see a similar dynamic in the proliferation of financial perpetual motion schemes on smart contract chains. Sure, you can disclose the code, but who has an interest in checking for decentralization cosplay? Even in centralized finance, bond rating is paid for by the bond issuer, instead of the more rationally-motivated bond buyers, because of co-ordination problems. With centralized exchanges you are trusting brand reputation and regulatory oversight to create incentives for security. That not as good true free-software decentralization, but its not nothing either. And that model is one size fits all; the best a user can do is scrutinize legal disclosure documents. Self-custody, however, has a range of security/convenience trade-offs. Verifiable hardware is a thing, and you don't need million dollar budgets to attain it. The idea is to photograph a generic FPGA board at IR wavelengths for which silicon is transparent (dubbed IRIS, infra-red in situ), and then loading HDL code yourself to make an embedded SOC.
Not a good example. No Binance client lost his funds. Only governments and bankruptcy can steal clients' funds from exchanges. Noone else. Hacks like this are cost of doing business. Included in the price of goods and services.
The asymmetry is that one failure mode is verifiable and the other isn't: a hardware wallet's supply chain can be checked — signatures, reproducible builds, buying direct — while an exchange's solvency is a trust decision you can never audit from outside. Verification doesn't just improve results, it converts trust into work. That's the whole thesis of this space in one sentence.
Interesting how influencers like you don't want to ever mention Retro-Crypto, the cold wallet being developed for Nintendo 64
↑