The biggest misconception about the Manhattan Project is that it was a race to design a bomb. The truth is, nuclear weapons are simple, even crude in design. The scientists at Los Alamos quickly devised two ideas. First, a “gun-type” weapon, wherein one large piece of uranium is shot into a smaller one, setting off a chain reaction. Second, there was an implosion device, where a much smaller plutonium core would be blasted evenly from all sides by lensed high-explosives. The latter design wasn’t proven it would work until Trinity, and the former was so rudimentary that they dropped it over Hiroshima without ever testing it. After that, the U.S. didn’t have enough enriched uranium stored for another gun-type, and only enough plutonium for another implosion over Nagasaki.
The real race, which the Germans lost once the Manhattan scientists learned that their Nazi counterparts were only experimenting with enrichment through heavy water reactors, was that of explosive material. This created a battle within the Atomic Energy Commission at the start of the Cold War, especially once it was found that the Soviets had also completed a weapons project. Washington and the Pentagon mistakenly believed that the secret to nuclear domination lay in weapons development, when any physicist at the time nuclear fission was discovered could’ve come up with rudimentary designs for a bomb. At best, the Rosenbergs got the Soviets on the right track for refining, but the paranoiacs killed them as if they really handed over some skeleton key to the burgeoning American military industrial complex.
The key to the Manhattan Project was central planning. While the production line was decentralized—from Berkely to Chicago, Hanford to Oak Ridge—everything flowed back to Los Alamos under Oppenheimer. This lies in stark contrast to the German bomb program, which had no real head, and ultimately consisted of many scientists working on pieces of what would become a bomb independently of each other, with many of their respective efforts scaled further and further back as the Nazi’s war against Europe became unwinnable. Maybe when Niels Bohr revealed to American scientists in 1939 that the Germans had proven fission, and that it was replicable, it seemed historically inevitable that Nazis would have a bomb, but by the time the Americans were coming close, that was no longer the case.
People in Silicon Valley spheres like to make equivocations, with their grandest at the moment that AI is the next industrial revolution. Some have gone as far to say that the current race for AGI (“artificial general intelligence”) is like that of the Manhattan project, and whoever gets there first will be in control of the new nuclear weapons of the future. This has led to a rapid escalation between some of tech’s top power brokers—the CEOs of OpenAI, Meta, Anthropic, Google, and xAI—to rush to build more and more datacenters in hopes that their models will be the first to become (what they consider to be) sentient.
Trillions of dollars are being spent on these data centers, and their current existence is already stretching power grids and water systems to a breaking point, and it will be environmentally catastrophic if they continue to exist at all, let alone at their current rate. All of its promises of a more efficient future are marketing ploys, like everything out of the current American tech sector, to receive more venture capital money and massive evaluations towards unprofitable products. Even the idea that AI is cheaper than human labor has been proven false. The only thing the tech giants of today are good at is market dominance, squeezing out old industries by undercutting them, before eventually gouging their own prices once they’re the only option.
AI is stretching the limits of their salesmanship, too, because the product that they are pushing is one that they don’t even understand, and all their proposals about what it could become are purely speculative. Unlike the Manhattan Project, where the scientists had a very clear idea about the weapon they were building and specifically what it would do, everything presented about AGI are aspirations about how “sentient” or "conscious" a computer could be. The tech sector might as well be early modern philosophers creating theodicies—these are people that the public considers ostensibly data driven logicians performing in the realm of conjecture.
Moreover, the competition between the main companies vying for AGI, with each new company made up of people who defected from another, is like if the Manhattan Project really did split—not on ideological lines, but ones of people simply wanting to dominate the others. And unlike the Manhattan Project, this is all happening before the bomb is even made. In the meantime, people are already feeling its fallout through water shortages and insane energy bills. In the case of AI, you don’t need to have the world-ending weapon to kill the planet, you just have to pursue it.