“All models are wrong, some are useful.”
This phrase was coined by the statistician George E. P. Box to describe probabilistic models, but it also perfectly encapsulates all the mental models we use to make sense of the world around us.
Human time and attention are scarce, and the universe is extraordinarily complex. As a result, we are forced to operate under imperfect mental models, also known in psychology as “heuristics.” Regardless of our level of understanding of any given subject, these models and ideas are necessarily erroneous or incomplete. The deeper one dives into any one subject, the more obvious George Box’s aphorism becomes.
Perhaps nowhere is this more readily evident than programming, where one of the most foundational principles is that of abstraction. To the visitor of a website, no knowledge of code is required to click links and input information, just as one doesn’t need to understand combustion engines to drive a car. We might have an approximate mental model of how they work but not an accurate one.
Similarly, web developers themselves need not understand the intimate workings of TCP/IP and the other core protocols on which the internet is built in order to build applications on them. We regularly use and incorporate software written by others in our own applications without ever knowing how they actually work. Software development, and technological advancement more generally, can thus be thought of as building on top of a series of these nested “black boxes,” with each box containing an even more abstracted-away mystery.
To those who haven’t invested the time to truly master the innerworkings of a particular technology, it might as well work by magic. The deeper you dive, however, the more the magic falls away.
This is what Jimmy Song did for me and the various other students of his workshop, Programming Blockchain: strip away the magic.
As a contributor to the Bitcoin Core repository and former vice president of engineering for the early bitcoin wallet software Armory, Jimmy Song is well known in the cryptocurrency space.
Through his regular written and video content, he’s established himself as a vocal figure in crypto, one who is passionate about improving bitcoin.
He also isn’t shy about sharing his opinions on what’s needed to do that:
“Training more developers is the biggest bottleneck in the ecosystem.”
Enter Programming Blockchain, Song’s flagship effort to give interested developers a deep crash course into the fundamentals of how the magic behind bitcoin and the blockchain actually work. Finite fields, elliptic curve cryptography, transaction parsing and validating proof of work are just some of the topics covered.
“It’s like a water hose of information for two straight days,” explained Song.
As a web developer fascinated with the wider implications of cryptocurrency for the last several years, I couldn’t resist.
Since blockchain is a global technology and phenomenon, it’s fitting that such an ambitious endeavor to demystify it is itself global in scope. The locations for Programming Blockchain vary widely, having been held and scheduled for areas as disparate as China, California, North Carolina and Israel.
“If the idea is to make more developers, I want to do this in as many jurisdictions as possible.” explained Song. “By doing this in different areas of the world, I am hoping developers in different areas of the world create more things. Having more businesses start in different jurisdictions reduces risk for bitcoin.”
The latest iteration of the workshop took place in Tampa, Florida. While not the most internationally recognized city, Tampa is home to a vibrant cryptocurrency community and the newly opened BlockSpaces, a co-working space dedicated to blockchain projects which played host to Programming Blockchain.
Choosing Tampa as a location paid off. This latest iteration of Song’s in-person instruction was his largest yet with 30 students. While some of these developers naturally hailed from the Sunshine State like myself, others had flown in from various locations including Washington D.C., California, and Brazil.
Blockchain is the hottest buzzword in tech, one that’s being thrown at everything. Surveying the ICO and blockchain landscape, you can find a project or startup for every use case from health data to banana tracking. No matter your problem, blockchain is the solution to your ills.
But what actually is it, how does it work, and what makes it so special?
It’s common to hear that blockchain is “the technology behind bitcoin,” a distributed and tamper-proof database which could be leveraged in many other applications. It’s also common to hear that like AOL or MySpace, bitcoin could quickly be overtaken by competitors who better leverage this technology.
But blockchain is so new and inherently different that all analogies aimed at simplifying it or the crypto ecosystem quickly fall apart in their usefulness.
Blockchain’s uniqueness makes it exceptionally difficult to understand because try as we might, we possess no preexisting conceptual pigeonhole to fit it into. By extension, it is exceptionally easy and tempting to project upon it a panacea for every problem without any clear idea of how it will help.
We take descriptions of the blockchain’s emergent properties such as “immutability” and “decentralization,” and often seem to conclude these are magical passive properties of blockchain which can be dragged and dropped onto any application. But there is no such thing as magic, and even the most seemingly benign assumptions made when thinking about cryptocurrencies and blockchain can be surprisingly off.
Take even the very concept of a bitcoin, which is itself nothing more than an abstraction. The bitcoin protocol tracks units of value only in satoshis, not in bitcoins. What many know as the “smallest” unit is actually the only unit in the protocol.
It was simply an arbitrary decision on the part of Satoshi to make a “bitcoin” equivalent to 100 million of these units, which subsequently became standard notation for all wallet software built on top of the protocol. But even the concept of some kind of “coin” or “token” itself is a total abstraction. The structure of bitcoin transactions has a surprising detail brought to our attention by Song that showed this to be the case.
When it comes to monetary transfers one thinks of X unit of value being sent to the address or account of a recipient. In a raw bitcoin transaction, however, nowhere is the amount of satoshis being “transferred” specified. There is simply a reference to the unspent transaction output, or UTXO, with which the transaction is being funded. A UTXO can be thought of as debit entry on the blockchain ledger. The total amount of bitcoin displayed on a wallet is the aggregate of all the UTXO it controls rather than a single account which holds funds.
Additionally, if the value represented by a single UTXO is less than that which a user attempts to spend, multiple UTXOs must be included in the transaction to provide the liquidity. However, a UTXO must also be spent completely, meaning that by spending an amount smaller than that represented by a single UTXO, your wallet software must actually generate a “change” address to send itself the difference.
As Jimmy Song demonstrated to us, there are no tokens being sent back and forth, even digitally. Rather it’s a conceptual metaphor. All there is is simply a quirky accounting ledger, the particulars of which are of course abstracted away completely by basic wallet software.
“Once you understand these raw transactions, it’s like reading the Matrix,” Jimmy said.
The pitfalls of abstraction
Many abstractions, like easily understood currency denominations, are obviously useful. They are necessary for operating in a vastly complex world, yet they can still introduce intellectual pitfalls.
Take unit bias, which is when a cryptocurrency seems like a better buy relative to a more “expensive” coin, despite the fact that the unit price of a coin is irrelevant in this context.
If two cryptocurrencies possess the exact same market cap, but their supply and denomination is such that you are capable of purchasing a “whole” cryptocurrency A over a “fraction” of cryptocurrency B, we are predisposed to own a whole of something rather than a part. Yet the denominations of these cryptocurrencies are, necessarily, totally arbitrary.
Unit bias is a fairly benign mental error. When it comes to simplifying details for the sake of explanation, however, other pitfalls can be much more dangerous.
For example, bitcoin’s so-called “immutability” isn’t the result of some special line of code which can simply be copied and pasted into any application. It is the result of the ongoing interplay of incredibly intricate mathematics and economic incentives. The structure of the blockchain is rooted in a type of computation known as hashing. It is easy for a computer to verify if the answer to a hash is correct but difficult for it to find the answer itself from scratch, though far from impossible.
Miners, however, create a hashing arms race, where reproducing their total and ongoing sum of computations in order to make changes to the blockchain is exceedingly expensive, rendering it all but impractical the more that time passes. This is only possible because the miners have a powerful profit motive: the reward of bitcoins themselves.
Thus it’s not even accurate to think of the bitcoin blockchain as perfectly immutable. It most certainly could be tampered with, under certain conditions like 51% attacks. But neither is it possible for any blockchain to promise even practical immutability without a native and valued token with which to reward those who secure it.
“Bitcoin is the technology that powers blockchain, not the other way around,” summarized fellow student Nick Baldwin.
A sense of perspective
The more deeply you delve into blockchain, the more the magic falls away. You realize that like all things, there are no true mysteries. Only that which we haven’t taken sufficient time to understand.
As our simplistic and flawed models are replaced by more sophisticated ones, there are interesting ramifications. You may think your sense of wonder fades along with the magic. Sometimes it does. You become acutely aware of how little you actually know and how much there is left to solve and build. A sense of disillusionment can be the natural reaction.
But by pressing on you earn something much more valuable than naive wonder: a sense of perspective. The work left to do is immense, but the work that has already been done by those who have come before us is just as terrifyingly intimidating.
It stands testament to the fact that we already stand on the shoulders of giants, and all the challenges ahead of us can be conquered, just as those before us were.
With this knowledge and shift in perspective comes a sense of focus. All we can do is solve the next problem. Take the next step. All else is noise.
As Song imparted to us as our impactful workshop came to an end:
“Wisdom is cutting things out of your life, not adding more to it”.
Astrological image via Shutterstock