10 Dec 2024 Reading time: 5 minutes This is Part 5 in a multipart series on nationalizing AI:
Superiority
Throughout history, moments of intense competition have driven humanity to achieve extraordinary feats, often under the shadow of immense pressure and high stakes. These races—whether against time, an adversary, or nature itself (often all three)—are marked by innovation, sacrifice, and moral complexity.
They reveal the lengths to which nations and individuals will go to gain an advantage or secure survival, even when the cost is extraordinary. Here are three such defining moments that changed the course of history.
The Space Race
On October 4, 1957, a crisp autumn evening, the Soviet Union sent shockwaves through the world, and through history, by launching Sputnik, the first artificial satellite. As its beeping signal echoed from orbit, the Space Race was born, a battle between superpowers to dominate the cosmos.
Four years later, in 1961, the Soviet Union’s Yuri Gagarin became the first human to leave Earth’s bounds, orbiting the planet aboard Vostok 1. His journey, lasting just 108 minutes, was a triumph for the Soviet Union and a humbling reminder to the United States of the stakes in this cosmic competition.
In 1962, President Kennedy inspired a nation and the world with his bold challenge: “We choose to go to the Moon.” This visionary declaration transformed the Space Race into a daring quest to push the boundaries of human achievement.
The race culminated on July 20, 1969, when Apollo 11 successfully landed on the Moon. As Neil Armstrong descended the ladder and declared, “That’s one small step for man, one giant leap for mankind,” the moment symbolized more than an American victory—it was a triumph for humanity.
The Space Race forever changed our understanding of exploration, turning the impossible into a reality and setting the stage for future journeys into the unknown.
Oppenheimer
On July 16, 1945, in the remote arid deserts of New Mexico, the world changed forever. Under the guidance of J. Robert Oppenheimer, the Manhattan Project successfully detonated the first atomic bomb in the Trinity test. As the fiery blast lit up the sky, Oppenheimer reflected on a line from Hindu scripture: “Now I am become Death, the destroyer of worlds.”
A theoretical physicist with a keen intellect, Oppenheimer led a team of brilliant scientists in a race against time during World War II. Their mission: to harness the atom’s power before Nazi Germany could. The result was a weapon of unimaginable destruction, marking both a triumph of human ingenuity and the dawn of a terrifying new era.
When atomic bombs were dropped on Hiroshima and Nagasaki, ending the war, Oppenheimer became a symbol of scientific achievement—and its moral complexity.
Enigma
During World War II, the Allies faced a deadly challenge: Nazi Germany’s Enigma machine, a cipher believed to be unbreakable. Decoding it would mean turning the tide of war, but the clock was ticking.
In 1940, Alan Turing and his team at Bletchley Park rose to the task. Through ingenuity and relentless effort, they built a revolutionary machine that unraveled Enigma’s secrets. Their success delivered critical intelligence, saving countless lives and shortening the war.
The race to break Enigma was a triumph of brilliance and determination, fought in silence and secrecy. It stands as a testament to the unseen heroes who shaped the course of history.
The Race to End All Races
All of the aforementioned events were born from intense rivalry, propelled by urgency, and achieved through groundbreaking innovation. They highlight humanity’s capacity to push boundaries under pressure, often at great cost, leaving legacies that reshaped the course of history. Notably, these were not privately run or owned efforts—they were fully centralized, publicly funded, and operated under governmental control to ensure their scope and impact could be effectively managed.
The AI superiority race is larger than all three of these, combined, times a trillion. This is it.
Such an entity would dictate the direction of humanity, able to directly and quickly reshape every aspect of life on Earth for all living beings.
This is the last race; there will be no others. Superintelligent AI will be humanity’s final innovation, and its control will determine our survival.
It being so critical to be first to achieve superintelligent AI, resource centralization is imperative. Notably, centralization was also a defining feature of the projects mentioned earlier. It’s obvious we are already investing a lot, federally, in AI already, but the vast majority of the talent and resources are currently in the private sector.
While no entity is perfect, the best choice is one accountable to the people through voting power.
I believe the case is clear - nationalization is necessary for the above reasons.
In the final post of mine, I will briefly recap my arguments and conclude this series on nationalizing AI: Nationalize AI - Part 6 - Conclusion
09 Dec 2024 Reading time: 3 minutes This is Part 4 in a multipart series on nationalizing AI:
Inefficient
Overall, the details here on how AI systems work are hand-wavy and explicitly high-level. In reality, it’s a minor point, but worth keeping in mind as I make generalizations here.
Markets, generally, are an excellent mechanism for allocating resources and driving innovation through competition. However, in this article, I argue that the unique characteristics and risks of AI make it an exception where market forces fall short of ensuring efficiency, equity, and safety.
The Current Landscape & Its Inefficiencies
There are really not many companies competing in building these very large systems, and much of their effort is duplicated. The models primarily use the same or similar datasets, techniques, and algorithms.
While duplication of effort is common across industries, in AI, it poses unique challenges:
Resource Intensity
Developing state-of-the-art AI models requires enormous computational resources, energy, and capital. This duplication wastes resources that could be used elsewhere. Not only could these resources be used in novel AI development, but they could also benefit other industries—or their conservation could help maintain reasonable pricing for others.
Not only this, the amount of resources required for each firm with an AI system is going up and accelerating very fast. This is especially true of energy (and therefore money), tomorrow’s AI systems will require massive upgrades to energy generation and likely the grid infrastructure. This is to say, while the duplicated resources are already incredibly intense - it will only get vastly worse.
Limited Competition
The high cost of developing these systems creates significant barriers for smaller players, consolidating control among a few major corporations.
Additionally, with similar datasets and techniques, most AI systems do not offer substantially different capabilities, meaning the duplication does little to advance the field.
Missed Opportunities for Collaboration
Instead of pooling efforts to address broader societal challenges, companies compete to slightly outpace each other with similar technologies.
You can see the rat race by simply looking up the AI benchmarks leader boards. Each company essentially strives only for marginal percentage improvements over the others in these benchmarks, without attempting to solve real world issues.
Why Nationalization Solves For This
Due to the unprecedented nature of AI—especially the high costs of developing these systems—and the lackluster competition in the industry, I contend that the current state of AI development is clearly failing to optimize AI for the public good. This represents a misallocation of resources that requires some form of government intervention.
A clearly defined and transparent public AI program removes a significant amount of the unnecessary massive duplication of resources. It also ensures that the developments and research is pointed in directions that are more aligned with solving massive societal issues that AI can and therefore should be helping to fix.
I’ve made the following point on each of the articles in the series so far - because I believe it to be true on each, and this one is no different; I believe these issues alone demonstrate the necessity of nationalization.
In the next article in the series, we will talk about why it’s important to be the first to control super intelligent AI: Nationalize AI - Part 5 - Superiority
08 Dec 2024 Reading time: 11 minutes This is Part 3 in a multipart series on nationalizing AI:
Corporate Power
Current Consolidation
One of the largest current issues globally is the massive consolidation and increase of corporate power.
There’s a lot of reasons for this, but one of the main ones has been the increasing number of mergers:
(Number of acquisitions per year from 1985 - 2023)
Why is Corporate Power an Issue Already?
At a high level, it is pretty easy to understand the literal consolidation - regarding # of companies that control certain industries - as was just covered. The problem does not stop there, however. With more and more power, and less competition; corporations are not using their power to better society, they are using it irresponsibly to abuse people.
Let’s walk through general corporate abuse, big tech’s existing abuse, and the abuses of current AI technology.
General Corporate Abuse
Corporations do anything for a dollar. They do not care about anyone except their C-suite and the shareholders. Corporations and the rich are more than happy with harming you, killing you, bankrupting you, killing your job, stealing from you - just for their next buck. I do plan on writing about a specific part of this at some point in its own post - but let’s briefly go through these:
- Turing Pharmaceuticals’ Price Hike
- In 2015, Turing Pharmaceuticals, led by CEO Martin Shkreli, increased the price of the life-saving drug Daraprim by 5,000%, making it unaffordable for many patients.
- Tobacco companies
- Despite clear scientific evidence linking smoking to cancer and other diseases, tobacco companies publicly denied these risks, lying to everyone about the dangers of their products.
- Targeted Marketing: They aggressively marketed cigarettes to vulnerable populations, including children and adolescents, to cultivate lifelong customers, often using themes of glamour and rebellion to appeal to youth.
- “The tobacco industry continues to amplify misinformation in the media, including the recent egregious attacks on tobacco control organizations[1]. The tobacco industry is the only one that stands to benefit by undermining tobacco control organizations.” Source
- These actions, over decades (and they continue), are the definition of killing people for money. By the way, tobacco stocks pay some of the highest dividends to shareholders in the stock market.
- Lead in Gasoline
- Many years ago, lead was added to gasoline despite the fact the companies knew that it was a known poison.
- For years, these corporations created fake studies, spent tens of millions on legal fights, created propaganda and marketing campaigns – so they could profit from costing a massive amount of harm. Source
- “A century of leaded gasoline has taken millions of lives and to this day leaves the soil in many cities from New Orleans to London toxic.” Source
- Long-Term Effects on Survivors: Millions of people exposed to lead during childhood suffer lasting impacts, including reduced IQ, diminished cognitive function, lower socioeconomic status compared to their parents, and an increased risk of mental health issues. These effects hinder their quality of life, career achievements, and overall well-being. Source 1 Source 2
These are not even the biggest examples, or all of them - obviously - here’s a short list of some of the other biggest instances: Purdue Pharma knowingly fueled the opioid crisis, BP caused the Deepwater Horizon spill by reducing safety measures, Nestlé exploited water resources, the asbestos industry knowingly sold deadly products, Volkswagen cheated emissions tests, Enron committed massive fraud, Lehman Brothers caused the 2008 crash, Wells Fargo created fake accounts, Ford sold cars with fatal defects, Equifax exposed personal data, Turing Pharma exploited drug prices, Amazon monitored and exploited workers, Coca-Cola drained water in drought zones, and Shell suppressed climate research, etc.
Big Tech’s Existing Abuse
Beyond abuse in many non-tech industries - another even larger problem is with the big technology companies controlling technologies that are also more powerful than any technology that has come before.
In particular - the technology companies have been horrific in this regard - and there’s been a very very slow regulatory response or really any accountability for some massive wrongdoings, including (but not limited to):
- App Store rent seeking
- The fees on the App Store(s) are the definition of rent seeking.
- The top social media sites are heavily consolidated and owned by the most powerful entities in the world, as shown in this chart:
Meta owns the top two and 3 of the top 9 social media sites. Not only this - Microsoft owns 1 of them, and the richest person in the world (Musk) owns 1 as well. Why have we all simply accepted that essentially all of our consumed social is owned by the biggest corporations and people in the world? By the way, this does not even count Threads (owned by Meta) because its a chart from 2023. How has the government not forced, at least, Meta to divest Facebook and Instagram already??
This power has been wielded for extremely nefarious purposes - such as election interference, general misinformation, massive foreign interference in our society causing unrest, selling of data, substantial increase in depression of adolescents, etc. Social media is the most powerful technology, that you’re allowed to privately own, ever.
Existing AI Abuse
By the way, the abuse of AI for profit is already happening. Let’s walk through a few examples, check out the links on each list item for more information on one specific instance:
- Deepfakes in Revenge Porn
- Many cases have emerged where AI-generated deepfake technology was used to create fake explicit videos of individuals, often women, and shared without consent.
- Voice Cloning in Fraud
- In 2021, cybercriminals used AI-based voice cloning to impersonate a company director in the United Arab Emirates, deceiving a bank manager into transferring $35 million. This sophisticated heist involved cloning the director’s voice to authorize fraudulent transactions.
- AI-Generated Harassment
- Some individuals use chatbots, such as GPT-based systems, to generate hundreds of abusive messages or harass others online, overwhelming victims with toxic content.
- Disinformation Campaigns
- Elections in many countries have been influenced to the point of changing the result, AI is being used already to subvert democracy.
- AI-Powered Stalking
- Tools like Clearview AI, which scrapes social media and other online platforms, have been misused by private individuals for stalking, identifying, and tracking people without their consent.
- AI in Child Exploitation
- Offenders are utilizing AI tools to create hyper-realistic, sexually explicit images of children. Law enforcement agencies are intensifying efforts to combat this misuse, as AI-generated child sexual abuse materials complicate the identification of real victims and pose significant challenges in prosecuting offenders.
- AI being used to automatically rejecting healthcare claims
- UnitedHealthcare is charged with using an AI system with a 90% error rate to deny elderly patients medically necessary post-acute care, forcing early discharges and financial burdens; or worse.
- Many instances where AI is convincing people to do extremely bad things.
And these are just some of the abuses.
While nationalization won’t fix all of these things, nationally owned AI can help fight against bad actors instead of being used by the bad actors (to profit the richest corporations and people on the planet). A lot (the majority?) of these abuses use AI developed and deployed right here in the USA.
The Power of AI
What is particularly interesting about AI at the scale of some of the systems we have seen recently is that they have an extremely large barrier to entry for any potential competitors. Could any non-massive company/individual afford to buy thousands of GPUs and pay the energy to train a model that is even remotely competitive with the existing models? Of course not. There is some nuance here due to the fact that some components created by some of the corporations are open source, but that fact largely doesn’t change much (although I do think this presents issues that I will cover in the next article in the series).
The current power of AI and the inevitable power of AI is extreme. Even beyond the power of labor automation (covered in Part 2) - the applications in weaponry, impersonation, mass social manipulation, etc make AI a large enough risk to not leave it in the hands of the tech companies that have already let us down and hurt society. For example, any AI weapon will not be a normal weapon - they will be almost assuredly unbeatable, it’ll be like the Manhattan Project all over, but accelerating a lot faster and far more dangerous. More on weapons a little later.
For context, remember, no private company owns or has ever owned even a singular nuclear weapon, and they cannot be deployed by a corporation.
It will not be long before whoever controls the most powerful AI systems will have the ability to control every other institution, organization, and individual. And in the meantime, the increasing power the corporations control can be used to prevent competition, sway public opinion, prevent backlash & regulation, censor citizens & journalists, etc.
As powerful as owning social media is, AI is exponentially more powerful. By the way, the companies who are currently leading, and are far ahead, in AI - also own social media networks. That combo is even more terrifying.
The leeway ceded with social media must not be repeated with something exponentially riskier. For all of the above reasons, I believe I have demonstrated the risks of corporately owned AI.
Why Nationalization Solves For This
Even ignoring the points made in Part 2 (Labor) - the points here alone, I believe, demonstrate the risks of maintaining private ownership of AI and therefore the need for nationalization.
In the USA, corporations are legally much harder to hold accountable for abuses than the federal government. The Bill of Rights directly limits government power and ensures protections for individuals, such as freedom of speech, due process, and protection against unreasonable searches and seizures. These constitutional guarantees provide a legal framework for holding the government accountable in ways that simply don’t apply to corporations, which are primarily governed by statutory laws and regulations rather than constitutional mandates.
This fundamental difference underscores why nationalization can provide greater oversight and public accountability. This is not to say the government is held to account enough, by the way - this is not the case but here are some reasons why they are held to account far more than corporations:
- Fiduciary Duty: Corporations prioritize shareholders, not public welfare, limiting accountability.
- Limited Liability: Corporate structures shield individuals from personal accountability.
- Lobbying Power: Corporations influence regulations, often weakening enforcement.
- Opacity: Corporations operate privately, unlike the government, which is subject to transparency laws like FOIA.
- Legal Resources: Corporations use vast resources to delay or avoid accountability.
- Public Oversight: The federal government is subject to checks, balances, and public scrutiny that corporations avoid.
- Examples: Cases like Enron, Purdue Pharma, and the widespread and extreme corporate misconduct during the financial crisis show how corporations and those who run them evade meaningful penalties, even when their actions cause massive and unprecedented harm.
AI’s immense and accelerating power should be controlled by an entity that the people collectively own and can oversee with transparency. In the USA, that entity is the federal government; lest the entities who control it, control us.
In the next part of the series, I will continue making the nationalization case by talking about how the market is not the most efficient way to allocate resources or drive innovation in AI, especially when public welfare and safety are at stake: Nationalize AI - Part 4 - Inefficient
07 Dec 2024 Reading time: 6 minutes This is Part 2 in a multipart series on nationalizing AI:
Labor
Overall, my thesis regarding labor and AI, for the purposes of this article, is broken down between two major points - ethics and employment. There’s plenty of interchange and correlation between the two, however specifically breaking it up here will allow me to be more pointed, especially about a specific ethical problem.
Ethics of Stealing
A critical foundation of the recent developments in AI is the data that is used for training. The specifics of how it is used and it’s part of the process here is less important than the fact that this data is largely publicly available data generated by billions of humans.
Ethically, how is it acceptable to do the following:
- Use data generated by billions of humans
- A lot of which is copyrighted, by the way!
- Create an AI system that uses that data
- Profit by charging businesses and other people for use of the system
- In reality, currently, a lot of the money is from venture capital, but the same ethical issue persists (or perhaps is even worse?).
Granted, a lot of the data is not generated necessarily by citizens of the USA - which is an interesting counterpoint; although my response to that is that should be globally owned instead of by any one nation. But for now, that seems far more impractical. Nationalization would likely be a critical first step towards globalized ownership, anyway.
While this is already a problem - profiting by using the free labor of others (who generated this data) at this scale - this only gets worse as the systems improve.
Billions of people contributed, and the CEO of OpenAI is driving a car around that costs 5 million dollars.
Unethical.
Employment
We do know, that these AI systems are useful for various tasks. I’ve used ChatGPT, for example, a lot of people have. It is useful! The usefulness is already resulting in increased productivity in a lot of industries.
Overall, as a society, we do want productivity to increase, so what is the problem?
Let’s take a step back and talk about, largely, non-AI related productivity gains. Other technologies, process improvements, etc. A critical issue we’ve been facing - which has led to historically bad inequality - is that productivity gains are not being met with the same % of an increase in wages, and it’s getting worse:
By the way - this is not strictly a partisan issue. Control of the presidency and congress have been both parties during this time frame (1979 - 2024), although it really started accelerating around when Reagan’s policies started (which Democrats & Republicans have refused to address / fix because of their corporate donors - but that’s a topic for a different day).
There’s many different angles to this, and it involves multiple groups in the upper class (shareholders, as well, for example) - but here is one example that is particularly stunning.
The two lines are realized vs granted compensation for CEOs represented as multipliers of the typical worker in their respective industries. The difference between realized and granted (the two lines) are not especially important for the context of this section. The main point here is that before 1980 - CEOs made “only” ~31 times the typical worker and now it’s over ~344 times as much. Note, the footnote in the image says this but this data is from the top 350 companies in the country specifically.
This is not a post primarily about income inequality - but it is critical to my thesis to show that productivity gains are not going to the average worker but to the members of the upper class.
Critically, I am not claiming nationalizing AI will solve this preexisting problem; but that it will prevent it from becoming worse due to the actions of private AI companies.
Effects on Employment
Short and Medium term
The effects on the labor market have already started. Increased productivity is common in many industries, as I mentioned before. As we know - the outcome of productivity gains are not resulting in rewards for workers, but rather increasing compensation for CEOs and shareholders. Nationalization of AI would ensure this is not the case.
While it is true that some segment of the affected labor force can shift, for now, this will not be the case forever.
Long Term
Despite my focus here on the current/short term effects on the labor market - it is my opinion that over some longer time frame, that AI will certainly be able to do every single job. This is a marked difference of AI versus new technologies in the past (printing press, internet, etc) which generated or shifted jobs in new or other sectors. An infinitely duplicatable, more-than-human-capable AI does not leave room for any job for any human.
The Long Term Is Likely Relatively Short
OpenAI is already talking about how AI could replace all / subsections of the labor force:
- “He’s going to be right there at the beginning of it, maybe even as things like AGI, we get there,” she (OpenAI Chief Financial Officer Sarah Friar) said on Tuesday, referring to autonomous systems that surpass humans in most economically valuable tasks. Source
- The time frame being referred to here is within the next 4 years (!!!!!!)
- “I want the door open to everything,” Friar said in an interview, when asked about a recent report that the company has discussed a $2,000 monthly subscription for its AI products. “If it’s helping me move about the world with literally a Ph.D.-level assistant for anything that I’m doing, there are certainly cases where that would make all the sense in the world.” Source
- “OpenAI CEO Sam Altman and Chief Technology Officer (CTO) Mira Murati said last fall that AGI will be reached within the next 10 years.” Source
- As previously mentioned, this is the point where most/all economically valuable tasks could be performed by AI.
This is coming soon.
Nationalization
To fully round out the earlier point on ethics for a second, it is wholly unacceptable that people and companies are getting rich by stealing from everyone and then benefiting from workers’ productivity gains and/or obsoleting their jobs.
In summary - the socialization of AI’s benefits, which in my view requires nationalization - is necessitated for these reasons:
- AI is already driving significant productivity gains in the labor market. However, as demonstrated earlier, the fruits of productivity gains, of course including those predating AI, have not been shared equitably with workers.
- Over a longer period of time, it is clear that AI will make humans obsolete for labor.
- These systems are built upon the stolen labor outputs of everyone.
This is only part two of the series, but in my estimation - already - these issues alone prove my thesis on the necessity of nationalization.
In the next part of this series, we will explore how corporate consolidation of power further necessitates the nationalization of AI: Nationalize AI - Part 3 - Corporate Power
