Universal Basic Compute
The Promise and the Pitfalls
The concept of Universal Basic Income or UBI, goes back centuries. First introduced by Saint Thomas More in his 1516 book Utopia, More describes a fictional island society where every citizen receives a guaranteed income, ensuring their basic needs are met. The citizens of Utopia live in a state of shared wealth and collective ownership of property. The kind of society that might make Karl Marx crack a smile.
Today, Universal Basic Income (UBI) has gained renewed attention and is championed by OpenAI Founder Sam Altman. As AI begins to progress toward AGI, fears are growing that millions of jobs across all sectors, from manual labor to finance and creative roles, could be at risk of displacement. UBI offers a safety net, providing everyone with a guaranteed income to meet basic needs.
Recently, Altman introduced a compelling variation on this idea.
Speaking on the All-In podcast in May, he suggested offering access to compute power, rather than money, could be a viable solution: Universal Basic Compute.
“I wonder if there are better alternatives to the traditional concept of UBI. Perhaps the future looks more like Universal Basic Compute than Universal Basic Income, where everyone gets a slice of GPT-7s compute power. They can use it, resell it, or even donate it for cancer research. Instead of receiving dollars or money, you own a piece of the productivity.” - Sam Altman
In this vision, instead of receiving capital, people would receive a share of the compute running GPT-7. People could use these resources however they like—whether for personal use, selling, or donating. It’s a compelling idea that reimagines how value is created and shared.
However, while well-intentioned, relying on OpenAI alone to deliver access to compute—and by extension, intelligence—isn’t democratization; it’s dependency. Sam Altman isn’t proposing Universal Access to Compute; he’s proposing universal access to his product.
I want to redefine Universal Basic Compute. True access and agency come from solutions that empower individuals directly—not from a compute nanny state with Sam as the gatekeeper. Instead, compute should be made affordable and accessible to everyone through free markets.
The goal is to make compute as cheap and available as possible so everyone can afford it. This approach empowers people to help maximize their agency—whether it’s earning income, enhancing their AI tutor, or supporting their child’s education.
Monopoly on Compute
Sam Altman envisions a world where everyone has the ‘right’ to use OpenAI. However, this dependency enables OpenAI to collect the data of millions, determine who gets access, and cohesively shape the minds of children.
When a small number of companies control most of the access to compute, they also influence which models are trained, which ethical guidelines are followed, and which innovations see the light of day. This risks creating a future where AI reflects the values and priorities of a select few companies rather than the diverse needs of society. This setup doesn’t truly empower individuals; instead, it creates a controlled ecosystem that dictates access and usage, making people dependent on one gatekeeper for opportunities and growth.
Universal Basic Compute Redefined
True Universal Basic Compute should mean unrestricted, independent access to computational resources, fostering competition. This approach empowers individuals to use compute power as they choose—whether for personal growth, innovation, or community impact.
A young Steve Jobs often noted he was envious of the fact that Alexander the Great’s personal tutor was Aristotle. Rather than just reading his words, Alexander was able to ask Aristotle questions directly. Jobs imagined a future where computers could eventually provide everyone with their own personal ‘Aristotle,’ allowing students to interact with great thinkers in ways that would redefine education. They could ask questions, get answers, and engage directly, fundamentally changing the learning experience.
Universal Basic Compute could turn this vision into reality, providing every child with the compute power needed to learn, earn, and engage deeply with AI-driven experiences. Imagine democratizing access to conversations with virtual figures like Aristotle, Jesus, or the Buddha, transforming how we learn and think globally.
Redefining Universal Basic Compute means moving away from dependency on a single provider toward a model where compute power is an open resource that enhances individual agency.
A Practical Perspective on Decentralized Compute
To help realize this vision, decentralized compute networks have emerged as a possible solution, aiming to break the monopolies of major providers like AWS, Google Cloud, and Microsoft Azure. However, while this vision of a more open and decentralized infrastructure is compelling, turning it into reality is complex and faces significant challenges.
Decentralized compute providers currently struggle to deliver the reliability and performance of established cloud providers. Competing with companies that have invested billions in infrastructure is no easy task. As a friend at NVIDIA noted, for AI engineers, the priority isn’t whether the infrastructure is decentralized—it’s whether it works. Engineers value consistency, low latency, and reliable performance over lower costs. Challenges such as network latency, managing distributed clusters, and security make it hard for decentralized platforms to compete effectively.
These challenges highlight a broader dynamic within Web3: a problem of misaligned incentives, where financial success can come from token speculation rather than building functional products. Unlike previous tech eras, where companies had to create valuable products that people used and took years to go public, crypto can enable wealth generation through tokens alone. As Mert from Helius outlined in his LightSpeed episode, what are the negative repercussions of giving teams millions of dollars before they even have a working product? Is it any wonder why we don’t have that many great apps in crypto? Balancing token incentives with genuine product development will be a critical challenge for the industry in the years ahead.
One thing is certain: we have trillions of dollars of dormant compute resources locked up globally, and figuring out how to utilize these resources effectively in an efficient manner could unlock incredible value. The success of decentralized compute will depend on moving beyond idealistic visions to tangible improvements in performance, user experience, and reliability. Decentralization must go hand-in-hand with practical, effective systems. By aligning openness with the practical needs of users, we can build a future where compute is not just accessible, but also efficient, reliable, and independent of big cloud monopolies.
Moving Forward
Sam Altman’s Universal Basic Compute is intriguing, but it falls short of true democratization. Real UBC will come from independent access to computational resources—not through dependency on a few providers. To redefine Universal Basic Compute, we must shift towards open solutions that put control in the hands of individuals, enabling them to harness compute power for personal growth, innovation, and societal impact.
However, the path to this future is not without challenges. Compute providers must overcome performance, reliability, and user experience issues to compete with established giants. Success will require a balance between idealism and practicality, focusing on tangible improvements and prioritizing real user needs over hype. By addressing these challenges and fostering open, efficient, and reliable systems, we can build a future where compute power is a universally accessible resource, driving innovation and elevating human potential for all.
