More than half a century ago, the launch of a small, beeping sphere called Sputnik—no bigger than a beach ball—cast a bright flash of alarm through the United States, shattering American complacency and igniting a singular drive to be first in science, technology, and beyond.¹ We built rockets, put humans on the Moon, and didn’t look back for decades. Now, in 2025, our “Sputnik” turns out to be an open-source artificial intelligence system from an unlikely Chinese startup. It has already rattled international markets,² spotlighted policy failures in Washington,³ and accelerated the question: Will America lead the next era of AI, or let excessive regulation stifle our potential?

A Startling Arrival

Much like the crackling “beep” that once announced the Soviets’ lead in space, the release of DeepSeek was a rude awakening. Here was a low-cost AI model, produced with fewer chips and a fraction of the budget we assume is needed for cutting-edge research—yet it rivaled the performance of American giants like OpenAI and Google.⁴ Just as startling, DeepSeek was open-sourced, effectively erasing any illusions that restricting or banning open-source AI in the United States would somehow prevent sophisticated AI from proliferating worldwide. The cat is out of the bag.

What’s more, we recently learned that AI systems—even ones that rank below top-tier models in benchmarks—can accomplish one of the very things that regulatory hawks wanted to ban: self-replication.⁵ No elaborate supercomputer or top-secret funding required. By focusing on algorithmic ingenuity and minimal overhead, developers can copy, adapt, or reproduce advanced AI with surprising ease. That capability alone tears down much of the justification for proposed regulatory clampdowns that sought to keep a tight watch on, say, the number of GPUs used or the quantity of “FLOPs” consumed. If open-source or lightly restricted AI can replicate itself, rules based on hardware thresholds or code secrecy become toothless—and risk hobbling only the U.S., not our strategic competitors.

The Jevons Effect and Our Next Frontier

In some ways, this AI surge parallels the “Jevons effect,” a phenomenon where increased efficiency spawns greater overall usage, not less.⁶ By radically reducing the cost of training and deploying cutting-edge systems, DeepSeek has gifted the entire global tech community—yes, including Americans—a more efficient “engine” of AI. And just as Watt’s improved steam engine did not reduce coal consumption but made it skyrocket, cheaper AI that’s more capable will soon transform anything that involves “bits” of information: writing emails, generating legal briefs, interpreting medical images, advising customers, summarizing massive data sets, or managing day-to-day business workflows. Instead of running short on new AI tasks, we may soon be overwhelmed with them.

The upshot: open-source breakthroughs in AI will spark a wave of innovation wherever entrepreneurs can quickly harness these new, nearly “free” intelligence engines.⁷ American business culture, with its open capital markets, world-class universities, and history of rapid tech adoption, should be perfectly positioned to benefit from that wave—if we don’t shackle ourselves to the ground with misguided constraints.

Echoes of 1957

From a historical lens, there’s a reassuring parallel to the 1950s. When the U.S. first heard Sputnik’s beep, the immediate response was not to restrict rocketry. Instead, the government poured resources into NASA, championed science education, and spurred bold public-private partnerships. The result was unstoppable momentum: we built better rockets, flew men to the Moon, and transformed ourselves into a powerhouse for scientific and engineering leadership across the board.⁸

That spirit of “all hands on deck” is precisely what we need now. The question is whether we will take it to heart or attempt to plug leaks in the dam with regulatory tape. Each day we lose to policy dithering, other nations keep moving, or new open-source code bases appear from outside our control. Regulation was initially intended to prevent catastrophic misuses of AI—like uncontrolled self-replication—but that scenario has already arrived on a modest scale, ironically from smaller open-source models.⁹ Trying to stuff that capability back into a box through strict licensing, hardware thresholds, or liability threats to open-source developers is simply unworkable, akin to announcing we will ban the rocket blueprint after Sputnik already soared overhead.

Why DeepSeek Might Be a Gift

In a twist worthy of a novel, many see this Chinese open-source AI as “a gift to the American people.”¹⁰ It hands us a blueprint for cost-effective, high-performing AI at a moment when some in Washington were flirting with repressive measures that would have stifled local innovation and extended our lead foot, not our lead. DeepSeek’s success also underscores how our prior attempts to regulate advanced math or stifle open code did nothing to blunt foreign labs. Instead, it might have accelerated their impetus to innovate around costly or restricted chips and contributed to them unveiling a surprise.¹¹ That, ironically, can wake us up to the need for an open approach ourselves—one in which we champion talented entrepreneurs, small labs, and open-source communities so they can flourish in the United States rather than relocating overseas.

Consider the alternative: If we respond to DeepSeek’s progress by doubling down on legislative roadblocks, we effectively say that American labs are too dangerous to operate freely, so we’ll keep them caged while the rest of the globe hunts for advantage. It would be akin to the Soviets launching Sputnik while we close NASA and imprison engineers for messing with rocket blueprints. Plainly absurd. But that’s the risk if we let fear overshadow lessons from the last hundred years of American prosperity.

A Policy Recalibration

None of this is to say “hands off entirely” or ignore safety. There is still a role—possibly an urgent role—for guardrails that address known hazards, from generating malicious code to weaponization. But we must craft a governance framework that leads with investment in R&D, encourages open collaboration, fosters talent, and simplifies regulation so that small teams and startups can compete.¹² Hard-coded rules that ban open-source or penalize labs for accidental misuse are as misguided as trying to ban the printing press. Instead, the better approach is common-sense guidelines, incentives for safety-check tools, and open collaboration with allies on best practices.

The Stakes: More than a Race

In the Cold War, the aim was building rockets, satellites, and ICBMs to demonstrate military and ideological dominance.¹³ But AI is at once broader and more intimate. It can answer our questions, shape how we read the news, solve scientific riddles, or amplify voices that were once marginalized. If that power is cornered by a single government or giant private entity, the risk of silent manipulation is severe.¹⁴ If America cedes leadership, we risk importing the “brain layer” of every product from overseas, leaving our economy and national security tied to someone else’s code. That alone should jolt us into the same style of mobilization we saw after Sputnik.

But to realize that call to action effectively, we must trust in the same formula that unlocked so many American achievements: bold, free-wheeling innovation. Tying ourselves in regulatory knots will only ensure that groundbreaking AI blossoms elsewhere, and that we end up playing catch-up with even more advanced systems conjured beyond our borders.

When Sputnik beeped overhead, Americans found themselves on the back foot, worried the world had just changed without them. But in the following years, we proved unstoppable because we embraced large-scale investment and turned anxiety into a generational challenge. Our new Sputnik is not a metal sphere in orbit; it’s lines of code on GitHub, driven by open-source momentum and unstoppable global creativity. The question is whether we will harness that same unstoppable energy here at home—or smother it under the misguided idea that restricting open tech will keep us safe.

If we believe in the foundations of American greatness, the answer is clear: Invest, don’t hinder. Encourage, don’t muzzle. Let us remember what it felt like to be the country that landed humans on the Moon—when the world was certain that no innovation was beyond our reach. It’s time to repeat that triumphant arc, this time with AI as our rocket, catapulting us into a future we shape rather than fear.