The Role of Bitcoin in the Era of Generative AI
The History of Scarcity
The history of life is a story of scarcity, but the history of humanity is a story of triumph over scarcity. Millions of years ago, hominids began to differentiate themselves from other animals by mastering fire. While some other animals can rudimentarily use tools, only humans gained control over fire and more complex tools. With that we started building shelters and weapons. With the invention of agriculture and animal husbandry, we gained greater control over our own sustenance, achieving a crucial victory over our primary risks: security and food. Previously, food and security were extremely scarce, requiring hunting and foraging for sustenance and defending against threats.
In primitive societies, ideas were scarce. Before the development of language, one had access primarily to their own thoughts. Some philosophers even argue that thought is intimately linked to language, implying a scarcity of complex thoughts. With language, people began exchanging information and developing a sense of history, passing down some information verbally through generations, but still limited to what could be memorized, necessitating the creation of memorable stories that rhymed. With writing, we gained access to the precise ideas of countless people from the past, many of whom had ideas that would have been lost to time if not for their manuscripts recovered years later. However, even with a way to reduce the scarcity of food and security, a large part of the population still lived in need with limited access to education and to these manuscripts, which were fragile and carefully guarded in monasteries. Ideas were still scarce, limited to a few individuals, and the recorded manuscripts had to be painstakingly copied by hand and carefully managed.
With the printing press, and then the industrial revolution, manual book copying became a thing of the past, and written ideas could finally reach millions of people worldwide. “Writer” became a profession rather than a hobby. Capitalism also began to alleviate global misery, and the goal of literacy was universalized. Now we had access to the ideas of professional writers, people with the time and money to be hobbyist writers, and the ideas shared by our relatives, friends, and close colleagues. However, writers were often at the mercy of publishers and newspaper owners who chose what was worth publishing.
With the advancement of capitalism and the internet, social media, blogs, YouTube, podcasts, ebooks, and self-publishing systems emerged. If there was once a problem of scarcity of ideas and strict curation, we now find ourselves facing an abundance of ideas and the challenge of navigating through them.
The History of Work
Something scarce is not necessarily impossible to create more of; it has a limited quantity based on the amount of labor applied. Thus, we can say that gold is scarce, even when we have increasing supply, as it is difficult to increase the world’s stock by more than 2% annually. Scarcity is a spectrum and is always related to labor.
The scarcity of food millions of years ago existed solely because of the need to expend labor to hunt and gather it. Something that does not require labor, i.e., can be acquired in large quantities without effort, is not scarce. Surplus labor played such a fundamental role for humanity, both for surviving harsh winters and for supporting family and friends (or the tribe) in times of need, that various ancient socialist doctrines valued labor highly and considered forms of income without work, especially interest, rent, profit, and inheritance, as immoral and in need of prohibition. The labor theory of value, which posited that the value of something derived from the labor expended, led Marx to declare that all profit was necessarily exploitation, and all capital (anything that could generate income, such as machinery, businesses, and productive land) should be common property. In the Soviet Union, a worker who skipped work was seen as harming their nation and could be arrested. Christianity in various countries also prohibited the charging of interest under usury laws. Usury, in the sense of charging any interest, was widely denounced as immoral or a sin by various religious leaders and philosophers throughout history. Sloth, or a lack of effort, was classified as one of the seven deadly sins. Even idleness, when a person was capable of working, was once a crime under various vagrancy laws. Even in Sir Thomas More’s book “Utopia,” his ideal society still had mandatory work, despite the presence of slaves. “Arbeit macht Frei” (Work Sets You Free) was already a famous phrase, but it became widely known in the Nazi forced labor camps, which exterminated Jews, a people historically criticized for making money by lending at interest.
On the other hand, in ancient Greece, Aristotle regarded work as something that prevented us from improving our minds, souls, and moral aspects, things that could only be done during leisure. However, in order to achieve this, work was performed by slaves. In Greek mythology, there were already automatons, metal beings similar to what we now call robots, that sometimes performed tasks for the gods. They were “self-willed” beings, meaning they did not require a human operator, and existed in various forms: humanoids, assisting the god Hephaestus in his forge, and animal forms like dogs or bulls, created to guard temples and treasures, among others. But it was probably in Karel Čapek’s play “R.U.R.” (1920) that the modern notion of robots (from the Czech word “robota,” meaning forced labor) popularized the idea of autonomous servants that exist to perform work for humans and create a utopian society without work. In 1930, Keynes envisioned that, with the growth of national productivity, people in the future would only work 15 hours a week. This is possibly the first feasible idea of how to achieve a post-work society without slavery. With the advancement of generative AI technology, this discussion is more relevant than ever.
Generative AI
Artificial Intelligence is a controversial term coined by John McCarthy in the mid-20th century. The very definition of intelligence is still the subject of much debate, so I won’t delve into the details of distinguishing between ML, AI, and AGI. I’ll use the popular term AI informally and generically. What matters is that AI algorithms have changed the way we solve problems.
Instead of programming a computer program with rules on how to solve a problem, it’s now possible to feed it examples and let it discover the rules on its own. In other words, rather than programming that a bird is an animal with feathers, beak, two legs, wings, and the ability to fly, etc., and all the complicated forms in which it can appear in a photo, I can simply provide millions of photos of birds, and the AI will “learn” to recognize birds, not only those provided but also others not in the training set. In essence, AI is a universal approximator. But if it learns the patterns of a bird, can it create a bird that has never been seen before?
Today, we know that it can, and using this property of AI algorithms, it’s possible to create images, photos, videos, texts, and sounds with what we call generative AI. With the extremely rapid advancement in the ability to create flawless texts, artwork, and computer code, some people worry about job destruction, while others celebrate the increasingly imminent post-work society.
We are in a unique moment in human history where ideas are quickly approaching perhaps the pinnacle of abundance. Not only do we have access to this abundance of ideas on the internet, but the process of thinking an idea through to its final shareable format requires less and less work, depending on its simplicity, it’s now a matter of seconds.
Problems arising with Generative AI
Generative AI can not only create previously unseen photos but also create a model of anyone in the world and generate convincing photos of that person doing anything, clone their voice, and make “them” say anything in an increasingly indistinguishable manner. All of this requires only a few photos and seconds of audio. While such capabilities were not always impossible in the past (we know Stalin was the precursor of Photoshop), they required specialized work and had more limited results. Today, it takes just a few minutes on a computer and almost no technical knowledge. Soon, it will take seconds and no technical knowledge at all.
One of the early consequences is the difficulty in detecting credit card fraud. Companies increasingly need more Know Your Customer (KYC) information, and soon, it will be easy to impersonate anyone, real or fictional, generate a credit card number, and use that information for mass fraud in the banking system. But of the problems, I consider this one the least concerning.
A more serious problem, in my view, is the possibility for almost anyone to take any publicly available photo of another person and generate convincing videos incriminating them, with photos, videos, audio, all well-crafted and virtually impossible to distinguish from reality, basically making a virtual clone of them. It’s surprising that, to date, we haven’t seen reports of political use of generative AI, only as jokes, like the famous photo of the pope.
Probably the worst-case scenario in a world of generative AI is a totalitarian country using its power to, instead of preventing its citizens from accessing the internet, Wikipedia, and sources of truth, allows all its citizens access to an edited version of these sites. With generative AI, it doesn’t take long for a totalitarian country to rewrite the entire Wikipedia and give its citizens the impression of having more internet freedom, even though they are reading lies invented by them.
Another category of minor problems arises with autonomous agents that may emerge and have self-ownership. Today, our entire financial system relies on a complex dance between various entities to use your money. A bank needs to recognize you as a person and, for that, uses your government identity and also a KYC company that verifies information about you. Every time you transfer money to another person and another bank, all these data need to match among various entities. An autonomous AI agent cannot have these identifications or a bank account. But it can have bitcoin.
Modern Problems Require Modern Solutions
Unfortunately, I don’t have all the answers to these problems, but we can speculate on ways to solve some of them, at least in part. The simple use of bitcoin as a currency instantly solves the two simpler problems. An autonomous AI agent will soon be able to hold its own bitcoins, make transactions, receive bitcoins, and pay for its own maintenance. The L402 protocol is currently being developed for micropayments between AI agents.
The issue of credit card fraud is only a problem because settlement takes more than four weeks to occur and the bank takes credit risk during that period. If money were not credit-based, which is due to our fiat financial system, and had instant settlement, like bitcoin, payment institutions would not need to undergo this credit risk or hire expensive KYC companies to identify customers. Customers would also have a better experience, as they wouldn’t need to struggle with their phone cameras all the time to prove their identity.
When I think about the problem of evidence forgery related to identity, we probably don’t have final solutions yet, only some palliative ones. One of them is public communication systems with personas. Since forging false images, videos, and audio to use someone’s information against them is almost effortless, I believe that in the near future most people will need to use personas. Since nonverbal characteristics of our identity, as ethnicity, fashion choices, accent, among others, play a part in making communication natural, it’s possible that software with filters will emerge to transform you into another person and alter your voice on the phone and over the internet. These filters would allow you to choose characteristics close enough to your own for rapport but distant enough to prevent virtual cloning. Protocols like NOSTR and others using DID will help navigate this hostile world. However, none of this prevents you from being virtually cloned from a physical world recording.
In this post-work world where almost anything is abundant, we have only one unforgeable source of work, which will be the source of truth: the timechain (commonly known as the blockchain) of bitcoin. Bitcoin’s proof-of-work (PoW) algorithm is currently the only form to have an unforgeable record, and its role as a decentralized temporal public ledger will become increasingly important in providing proof of existence. While the rest of the world will rely in a trusted subjective social source of truth, the timechain can provide a trustless objective unforgeable source of truth.
In the specific case of a totalitarian country shutting down the internet and creating a new edited Wikipedia, it’s entirely possible that they would do the same with a cryptocurrency that uses Proof of Stake (PoS). Since the PoS algorithm relies entirely on social factors, when they reconnect the internet and there are five identical cryptocurrencies, it will be impossible to distinguish which one is the true one, unlike bitcoin, where it is highly unlikely that the country can replicate all the proof of work spent since the beginning of the protocol, which will indicate the true timechain.
Today, this kind of notary service is already being done mainly with the OpenTimeStamps tool. In 2017, Peter Todd, one of the creators of this tool, used it to record the entire Internet Archive in a single bitcoin block. More recently, the company Simple Proof used this type of record to prevent fraud in the Guatemalan election. This type of record requires that the original data be preserved, which can be lost. I believe that in more important cases or when people want to signal higher status, some will pay very high amounts to use the scarce space of the block to register directly with Inscriptions.
In conclusion, this will continue to be a highly debated topic, and I believe that various new solutions will emerge. While a world of leisure and abundance is highly desirable, work is still and will always be a fundamental aspect of human nature.
