The Adoption Dilemma

Imagine a person claiming to be your brother texts you from an unknown (overseas) number asking for money. He informs you his phone and wallet were stolen while traveling, and he’s borrowing a stranger’s phone to text. Let’s say you know your brother was in fact traveling, and the text sounds like him — he even calls you by a nickname most people don’t know. What do you do?

You might ask him a few questions to which only he would know the answer, like what football team does he root for, and who was the player on that team who cemented that rooting interest? (If it were my actual brother, the answers would be Dolphins, Dan Marino, even though he grew up in NY, and now that I’ve written this, I’ll have to ask him a different question should this ever happen.)

After a couple such questions, you could be virtually certain it were actually your brother and not some scammer (Of course, he could be relaying this information under duress, but if we’re talking about a few hundred dollars, better you send it in that case anyway.)

What’s actually taken place here? Your brother, via his unique, individual knowledge was able to ID himself and induce you to send him money. A scammer, without this particular and unforgeable knowledge, would not be able to trick you into doing so.

. . .

One of the debates in Bitcoin circles is the tradeoff between scaling (growing the number of users) and self-custody (people holding their private keys rather than leaving their bitcoin on an exchange.)

Most people are used to having bank accounts wherein you have a login and password, and you access your funds via a third-party-controlled interface. If you forget your password or username, it’s not a big deal, you can just call customer service and have them send you a re-set email. And if someone gets hold of your credit card information and makes a fraudulent charge, you can have it reversed.

As such, most people prefer to leave their bitcoin on an exchange, wherein there’s a company that can help them with their login and manages to protect their bitcoin from theft. The idea of holding their own keys, wherein if they lose their seed phrase or expose it to thieves, it’s irretrievably lost forever is understandably terrifying, especially if they were to have significant savings in it.

But having your bitcoin on an exchange isn’t that different from having your dollars in a bank. If the government decides you’re spreading “disinformation”, for example, it could order the exchange to confiscate your coins. In all likelihood the exchange will comply — its principals don’t want to be shut down or jailed to protect you. One of the key qualities of bitcoin is that it’s decentralized and not subject to confiscation by the government*.

Moreover, as we saw recently with the FTX scandal, the exchange itself could steal your coins, or misrepresent its financial solvency. Unlike a bank there’s no FDIC insurance, nominal as that might eventually turn out to be in a true collapse.

*(They could obviously come to your house and physically arrest you, but it’s far easier to target a few exchanges remotely and have them do the dirty work, than go after millions of individuals privately holding coins. Moreover, if you said you lost your keys, they could never prove otherwise or access the coins.)

So there is a tradeoff between rapid mass-adoption (likely via exchange) wherein people are not really adopting bitcoin, but IOUs of bitcoin controlled and owned by a centralized depository with many of the same risks and chokepoints of the legacy system, and slower adoption via self custody that requires individuals to accept the responsibility of holding their own wealth.

In summary, self-custody requires a certain amount of understanding, responsibility and effort that are a non-starter for most people, while exchange-custody doesn’t solve the essential problems the technology was designed to solve: censorship resistance and counterparty risk. How then can bitcoin scale without losing its soul?

. . .

I’m not a tech person, and I don’t understand some the specifications of bitcoin, but it’s my understanding that at one point private keys were solely accessible as long strings of hexadecimals (numbers and letters.) That is, you had to know a 64-character string of numbers and letters to access your coins.

Eventually, bitcoin developers (via Bitcoin Improvement Protocol 39) figured out a way to convert this unwieldy 64-character string into seed words — the 12 or 24 words you can memorize (and write down) to generate your private key and have access to your funds. Most people find it a lot easier to memorize 12 (or even 24) words than a long-string of hexadecimals.

This was an improvement in UI (user interface), but even so the average person might not want to depend on recalling random words (in the exact order and without making an error) to access his life savings. Moreover, the more one makes sure the words are accessible and unlikely to be lost or destroyed, the more likely it is they are seen by someone else who could steal his coins. In other words, the more you secure your private seed words from loss, the more you expose them to theft, and the more you secure them from theft, the more you expose them to loss.

This is another difficult tradeoff you have to make even after you’ve signed onto the heavy responsibility of custody over your savings.

. . .

It occurred to me that maybe there’s a simple solution to both tradeoffs. Bitcoin is quite literally information money. If you have the information (private key), you have control of the coins. The information is the money, so to speak.

The problem is the information must be sufficiently complex that it can’t be guessed or accessed by someone who shouldn’t be privy to it. That makes it daunting for the average person to use in the way it was designed.

The problem is similar to the scenario posed in the opening section: how do I digitally (over text) verify that a person is who he says he is. How do I prove I am who I say I am? How do I access my funds while at the same time preventing others from doing the same? I need specific, complex information only known by me.

The 24-word seed is a version of this, but the UI is still challenging enough to prevent the system from scaling barring what would seem like a herculean effort on the part of the public. But what if we could have the equivalent complexity of a 24-word seed, a 64-character hexadecimal, a 256-bit string of 1s and 0s, but with almost no effort whatsoever?

My brother and I don’t have to memorize that his favorite team is the Dolphins and his favorite player Dan Marino. We know that already, and we won’t forget it — or at least are less likely to forget it than we would random seed words. A person who grew up in large city doesn’t need a map to navigate its streets and neighborhoods, even decades later. It’s not information one needs to memorize, but simply acquires in the circumstances of living life.

What if there were a way to convert the seed words into known, private facts about one’s life? Of course, just as you want to ensure the seed words are truly random, you’d want to ensure the known facts are not ascertainable via public records or social media posts, etc. Imagine a super-intelligent AI absorbing all the publicly available information about you, and pick something it would not be able to find. It’s also an incentive not to make the particulars of your private life public — ideas yes, personal details no.

. . .

I wish I could offer the precise implementation of this idea in this space. I cannot. I do not know the best way to create a “personal seed creation kit” or whatever you want to call it. But I imagine it could be done. A offline electronic device you’d use to generate the relevant information that serves as your private key.

An old T-shirt, a childhood pet, your childhood phone number, the friend who lived across the street when you were six, your dad’s favorite football player, etc. It’s not easy because things like your favorite TV show or your favorite restaurant could be crossed checked by AI with your age and home address. But there could be enough questions/options that no one would know even which parameters you were using.

I realize some financial institutions already have this kind of thing for when you lose your password, but the parameters are bad. They ask you questions for which you don’t have an answer! “What is your favorite sport?” I like a lot of sports and might choose a different one on a different day.

But a well-designed kit would be able to generate private key information that’s not ambiguous like my brother’s favorite team and player. (In truth, sports teams would probably be off limits as the AI would be quick to guess those along with their most prominent players.) But a good kit would reject weak seed words the way websites reject passwords like "12345”.

. . .

The idea that information itself is money is actually not a new one. What makes Warren Buffett (someone who rejects bitcoin as a viable money, incidentally) richer than me? Information. He knows his accounts and his holdings, and he can verify that he’s the person entitled to them.

The main difference between the legacy system and bitcoin in this respect is in the former, the information you have must be sufficient to convince a third-party to turn the money over to you, while in the latter, the information is itself the money. The middleman has been eliminated.

Everyone given his unique life situation has information and knowledge that no one else in the world possesses. No one else knows exactly how your bookshelf is arranged, which books on it you have and have not read. The idea is to take this unique information and convert it into a private key, so that the average person can use it as digital money that’s durable, unforgeable, of limited supply, censorship resistant, hard to confiscate or steal and without counterparty risk.

. . .

One last thought — I had the “big idea” that TREE(3) was like the human mind, while Graham’s Number was like AI. I also compared countable infinity to AI and uncountable infinity to the human mind in a subsequent piece. In last week’s podcast, while I was explaining this to co-host Ted Bell, he asked what all of this was for, what the point of these huge numbers really was. It occurred to me that this is one application: While the AI can brute-force search through ungodly amounts of data to try and guess your private key, you could pick 12-24 examples from your own life that you won’t forget that it would never guess.

That’s because you’d be choosing your seed based on objects that had meaning to you, something the AI can never truly grasp. You would remember it because it’s encoded emotionally, not just as cold data. That’s the problem with the seed words and the hexadecimal code, why it will be hard to generate mass adoption with that machine-like UI. Most people prefer a story, a narrative, something that makes them feel something. That’s why parables are often more effective than commands, and literature or film more life-changing than philosophy texts or self-help.

The human mind is a higher-order of complexity than an AI mind in part because the emotionally-encoded memories relate to a dimension inaccessible to it. Life and consciousness run deeper. It would be ironic if in the end, one’s wealth were untouchable by the machinery of state because it were secured by information relating to meaning, to the things a person loved.

It occurred to me maybe that’s what Jesus meant when he cryptically replied to questioners trying to bait him into illegally saying Jews shouldn’t pay taxes to the Roman regime:

Render therefore unto Caesar the things which are Caesar's, and unto God the things that are God's.