Ask anyone to name the most dangerous invention in history and most people will say the atom bomb without hesitating. The mushroom cloud is the universal symbol of human self-destruction. Yet there is a serious argument that a far humbler invention — a design tweak from 2006 that removed the "next page" button from websites — has already done more cumulative damage to human minds than nuclear weapons ever have.

This is not a rhetorical trick. It is a question about how we measure danger: by potential catastrophe, or by actual, ongoing harm. Let's take both inventions seriously and look at the evidence.

The atom bomb: maximum potential, minimal use

The destructive record of nuclear weapons is horrifying but, historically speaking, short. On 6 August 1945, a single bomb dropped on Hiroshima killed roughly 66,000 people according to a 1945 US government report, with another 69,000 injured; the Nagasaki bomb three days later killed an estimated 39,000. Those two bombings remain the only times nuclear weapons have ever been used in war.

That fact is remarkable. According to the Federation of American Scientists, nine countries possessed roughly 12,187 nuclear warheads at the start of 2026, with the United States and Russia together holding about 86 percent of the global inventory. This is dramatically down from the Cold War peak of around 70,000 warheads in 1986, but the trend is now reversing: the usable military stockpile is growing again as China, Russia, India, Pakistan and North Korea all expand their arsenals, and the arms-control architecture that restrained the Cold War — the START treaties, the INF Treaty — is mostly broken or expiring.

The danger is real and arguably rising. A single modern warhead detonated over a major city could kill hundreds of thousands of people; ICAN estimates a single detonation over New York would cause more than 580,000 fatalities. A full exchange between major powers could plausibly end organized civilization.

And yet, for eighty years, the very visibility of that horror has acted as a brake. The bomb produced deterrence doctrine, the Non-Proliferation Treaty, hotlines between capitals, and a global taboo so strong that not even the most reckless nuclear-armed states have crossed the line since 1945. The atom bomb is the rare invention whose danger is so obvious that humanity built institutions specifically to contain it.

The infinite scroll: minimal drama, maximum exposure

Now consider the other invention. In 2006, designer Aza Raskin popularized the infinite scroll — the feature that loads new content automatically as you reach the bottom of a page, so you never have to click "next." (Microsoft engineers had filed related patent work around the same period, but Raskin became its most prominent evangelist and, later, its most prominent critic.) His motive was pure usability: don't force the user to make a choice they don't care about.

What he removed, though, was the stopping cue — the natural pause where a person decides whether to continue. Combined with algorithmic recommendation feeds, autoplay, and notification loops, the result was a machine optimized to capture attention. Raskin has since said publicly that he regrets it. In a 2019 interview he admitted, "I regret that I didn't think more about how this thing would be used," explaining that a designer who removes the stopping cue can steer user behavior almost at will. He has estimated that infinite scrolling wastes time equivalent to hundreds of thousands of human lifetimes every day. He now works at the Center for Humane Technology, an organization dedicated to undoing the damage of the attention economy — and has testified against Meta in court.

He is not alone among the regretful architects. Sean Parker, Facebook's founding president, described the platform's like-button feedback loop as "exploiting a vulnerability in human psychology." The people who built the attention economy have been unusually candid, on the record, that engagement-maximizing design works on the same variable-reward mechanics as slot machines.

The scale of exposure

Here is where the comparison becomes stark. Nuclear weapons have directly harmed hundreds of thousands of people. The scrolling feed touches billions, daily.

According to DataReportal's 2026 figures, the average person worldwide spends about 2 hours and 21 minutes per day on social media alone, and roughly 6 hours 40 minutes on screens overall. Gen Z users average around 4 hours of daily social media time, and in countries like the Philippines usage approaches 4 hours a day across the whole population. Run the arithmetic across a lifetime and the average person will spend well over 15 years of waking life looking at screens — a meaningful fraction of it in passive, algorithm-directed consumption.

The mental health evidence, while still debated in its causal details, is accumulating:

The US Surgeon General's 2023 advisory reported that up to 95% of American teens use social media, with one third using it "almost constantly." It found that 46% of teens said social media made them feel worse about their body image, and more than a third of girls aged 11–15 described themselves as feeling addicted to a platform. Research cited in a 2026 statistics review found that teens spending 4+ hours daily on screens report anxiety symptoms at around 27% and depression symptoms at around 26%, with 6–7 hours of daily use roughly doubling the risk of an anxiety or depression diagnosis.

Perhaps most telling: teenagers themselves have turned against the product. Pew Research Center found in 2025 that 48% of US teens now say social media has a mostly negative effect on people their age — up sharply from 32% in 2022. Some 45% admit they spend too much time on it, roughly the same share say it hurts their sleep, and 40% say it hurts their productivity. The users of the machine can feel what it is doing to them.

The reckoning has begun

For decades, the difference between the two inventions was that one was regulated by treaties and the other by nothing at all. That is finally changing — and the change itself is evidence of how seriously governments now take the harm.

In December 2025, Australia became the first country in the world to ban social media accounts for children under 16, covering platforms including Instagram, TikTok, YouTube, Facebook and Snapchat. More than five million under-16 accounts were deactivated in the first months, and by mid-2026 the government moved to double maximum penalties for non-compliant platforms to A$99 million after concluding that tech companies were still not doing enough.

In the United States, the fight is happening in courtrooms. More than a thousand school districts have sued Meta, Google, TikTok and Snap, arguing that addictive platform design — infinite scroll, autoplay, push notifications, recommendation algorithms — created a student mental health crisis that schools were forced to pay for. In 2026, a Kentucky school district secured roughly $27 million in settlements from Meta and YouTube in the first bellwether case, a Los Angeles jury awarded $6 million to a young woman who argued the platforms were engineered to addict children, and Meta reportedly faces state lawsuits with claims totaling as much as $1.4 trillion. The companies deny wrongdoing and note that "social media addiction" is not a recognized clinical diagnosis — but the legal and regulatory tide clearly resembles the early tobacco litigation era.

So which is more dangerous?

The honest answer depends on which axis of danger you care about.

By destructive potential, nothing rivals the atom bomb. It is the only human invention that could plausibly end civilization within hours. As arsenals modernize and arms-control treaties collapse, that risk is not theoretical nostalgia — it is arguably higher today than at any point since the 1980s.

By realized, ongoing harm, the attention machine wins, and it is not close. Nuclear weapons have been used twice in eighty years. The infinite feed is used by roughly five billion people every single day, for hours, during the exact developmental window when young brains are forming impulse control and self-image. Its harms — degraded attention, disrupted sleep, comparison-driven anxiety, weakened real-world connection — are individually small and collectively enormous.

But the deepest difference is this: the bomb's danger is self-limiting because it terrifies us, while the feed's danger compounds because it delights us. No one has ever voluntarily detonated a nuclear weapon on themselves each evening before bed. Billions of us pick up the scroll willingly, nightly, and the business model rewards making it ever harder to put down. Fear built treaties around the bomb within years of Hiroshima; it took nearly two decades of the smartphone era before the first country dared restrict the feed even for children.

There is one final, uncomfortable symmetry. Robert Oppenheimer spent his later years warning the world about what he had helped create. Aza Raskin and Sean Parker are doing the same today. History's most dangerous inventions seem to share a signature: their own creators are the first to sound the alarm — and the last to be heeded.

The atom bomb remains the invention most capable of destroying us. But the infinite scroll may be the invention that is actually destroying us — slowly, pleasantly, and at a scale no weapon has ever reached.