Transformative technologies are the ones that disrupt the balance of power. They arrive as tainted vessels of progress, carrying both destruction and revitalization. From splitting the atom to modern machine learning, we keep arriving at the same crossroads. Each era's breakthrough is weaponized first, then settles into equilibrium once access spreads. AI will follow the same arc, and the West is approaching it all wrong.
The Atom
When scientists first unlocked the atom's power, they opened the door to unprecedented energy. Before anyone imagined using it to boil water, we used it for destruction at Hiroshima and Nagasaki. That move ended the Second World War, and within a few years its aftermath had rewritten the world order. Transformative technologies wielded first as weapons are really the exercise of new strategies made possible by asymmetric information. As nations raced for nuclear power through the Cold War, the enrichment process largely determined whether a state would operate with some guarantee of autonomy or be subjugated by the actors who held that asymmetry.
After witnessing the bomb's devastation, many physicists grappled with regret and urged global restraint. Restraint was arguably achieved, but you can't help wondering what a menace the United States might have become had the Russians not matched its strategic abilities. An advantage that large could easily have inflated American ambition into something more overtly bent on global dominance.
Biology's Turn
Stem cell breakthroughs ignite the same hope and spark the same ethical dilemmas, particularly around embryonic sources. Just as we faced the atom's duality, we now confront the responsibility of wielding biology's power. The difference this time is access. Research groups and wealthy private citizens reached the more dubiously sourced methods first. One can read in the Epstein emails discussions on how to produce compatible cells for repair and rejuvenation from fetuses unethical sources. Their methods were monsterous to civilized people, but their wealth gave them an advantage in accessing better health outcomes. Stem cell research has since progressed, Muse cells among the results. Over time gentler techniques emerged, including stem cells grown from our own tissue. Today we inherit that legacy without the same ethical cost. As science matures, so do its methods, and innovation and ethics start to align. As access to the knowledge at the edge of our understanding widens, the strategies within reach grow less destructive.
CRISPR is a biological asymmetry that was killed in its cradle. Discovered in 1987, it has yet to materialize as a truly destructive weapon for asserting some new strategy. You would assume gene editing is as powerful as a nuclear weapon, yet the accident of its discovery as an editing platform made the secret impossible to protect.
Asymmetry Is the Problem
Technology becomes dangerous when asymmetry emerges. In game theory, asymmetry is often what determines advantage, and technological asymmetry is just one form of informational asymmetry. So de-escalating weaponization means reducing asymmetry. When a technology is broadly accessible, the public can manage it. They can audit it, regulate it, and understand it.
That is why I would propose three principles for steering Artificial Intelligence onto a less destructive path.
AI must remain open source. Full stop. The Web 2.0 model will not work here. The competitive advantages organizations gain through trade secrets, IP, and insular monetary velocity will act as destructive strategies in Web 4.0. A race toward self-improving AI without wide distribution of the technology all but guarantees the dystopian outcomes Max Tegmark lays out in Life 3.0. The West is willing to sacrifice everything it has built for AI supremacy, and that will be its ruin.
AI systems that influence public life must be transparent. We need to understand how information flows through them, or the benefits of individualism are lost entirely. This is critical for democracies and important for the future of humanity more broadly. A monoculture under a one-world government might survive terrestrially, but it will shatter within a light-minute.
Centralized control without public access has to be off the table. We must innovate with symmetry, so that the future is shaped by all of us. Closed-source models should be illegal, treated as threats equal to or greater than nuclear weapons. This technology can unite humanity and its pursuits, or it can fracture our civilization so badly that we may not heal for millennia.