EVs, obesity, and nuclear war: Vox predictions about the next 10 years

Every year, the staff at Vox’s Future Perfect section makes predictions for events we think are likely to happen (or not happen) in the year ahead. Big deal, right? Reporters and pundits make predictions all the time.

What we think sets our efforts apart from the usual journalistic prognostication is that we put a numerical probability on each of our predictions, which we use to signal how confident we are in what we’re forecasting. That matters, because numbers are more meaningful and more comparable than words like “definitely,” “maybe,” or “probably.”

Usually when we make our forward-looking predictions at the end of the year, we check back the following year to see how we’ve done. That’s partly because we think holding ourselves accountable is just good epistemic practice. (If we’re wrong, we should acknowledge it.) But we also do it because the best way to become a superforecaster is through forecasting, checking the results, and forecasting again. 

This time we’re doing something a little different. To mark our 10th anniversary, Vox’s staff dug into the turning points of the last decade — those moments in the news when history shifted course. Future Perfect took the opportunity to try to proactively predict what might be some of the meaningful turning points of the next decade. So come back for our 20th anniversary and see how we did. (Unless, of course, prediction seven comes true, in which case we may all have other things to deal with.)

Progress against global poverty has been extremely rapid in recent decades. From 1990 to 2019, the extreme poverty rate, as measured by the World Bank, fell from 38 percent to 8.9 percent globally:

The decline was especially fast in East Asia (in particular China) and South Asia (in particular India). In East Asia, the rate fell from 65.4 percent in 1990, the highest of any world region, to a measly 1.2 percent in 2019. The economic disruptions of the pandemic were a major setback across the globe, but poverty reduction seems to finally be getting back on track, at least in South Asia.

The bad news is that today, most poor people live in sub-Saharan Africa, and while poverty there fell in recent decades, it fell more slowly than in Asia. Weak, coup-prone states, persistent infectious disease burdens, and recurrent war and violence have made it harder for African countries to build large export-based industries like China, or robust service sectors like India’s. 

The most recent projections I’ve seen by World Bank researchers suggest that the world’s extreme poverty rate will fall to 6.8 percent by 2030. At that pace of progress, the rate should still be over 6 percent by 2034.

The final steps toward eliminating extreme poverty will be the hardest, and progress will likely be slower than it’s been in the past couple decades. It’s important also to remember that the $2.15 per day poverty rate used by the World Bank is very low. The US poverty rate is closer to $30 per day; by that standard, the vast majority of the world will still live in poverty in 2034. But I have faith we’ll still make progress.

The autonomy metric developed by SAE (the organization formerly known as the Society of Automotive Engineers) measures on a 0 to 5 scale how capable cars are at self-driving. Zero means the driving is totally manual; 1 means that one element, like the speed of the car, is automated, as in the adaptive cruise control systems common in new cars as of 2024; 2 means that multiple elements, like both speed and steering, are controlled by computer, as in Tesla Autopilot (which, to be clear, is not actually an autopilot).

Level 4 is where things start to get interesting. It signifies a car that can self-drive all the time, provided it’s only used in a certain type of environment (say, only in a city, or only on paved roads). This is where the technology goes from a convenience to a game-changer, and where scenarios like pushing a button to have your car find a parking spot by itself, or commuting to work while typing at a laptop and not even looking through the front windshield, start to become possible.

We already have level 4 vehicles. Sort of. On a trial basis. As of 2023, Waymo, the self-driving division of Alphabet (formerly Google) and by far the industry leader, was operating level 4 taxis in San Francisco and Phoenix. It has expanded to Los Angeles this year and is opening in Austin soon. 

Until very recently, Waymo has limited itself to surface streets, and its vehicles are pointedly not for sale. They’re for rent, and even then in a small handful of places. I would very much like to use a self-driving car on my periodic 500-mile road trips from DC to New Hampshire, but alas, we’re still a ways off.

I’m guessing that by 2034, we’ll be there. Waymo and other companies are gathering massive amounts of video, radar, and other data they can use to train cars to operate in diverse environments. Meanwhile the economic advantages of self-driving vehicles — in terms of labor costs, parking costs, idleness, and more — are enormous. 

The fact that Waymo can run vehicles reliably in several cities — albeit cities that don’t have much experience with snow and similar adverse weather — is a very encouraging sign. There’s a gap between that and hopping into my 2035 model year Honda Civic, pulling up an address on my phone, and having the Civic chauffeur me there like a sultan. But the gap is narrowing, fast.

Here, I am mirroring the UN’s World Population Prospects report, which estimates that in 2034, global life expectancy at birth will reach 75.2 years. That number will mask considerable inequality, with North America at 81.7 and Africa at 65.6.

But as with global extreme poverty, this projection also relies on decades of past progress. When UN population data began in 1950, the world had a life expectancy of 46.5 years. Over the next thirty years, it rose to over 60. In 2021, in the midst of a global pandemic pushing the number down, it was 71. 

As the world gets better at conquering infectious diseases like tuberculosis, HIV/AIDS, and malaria, and as countries grow richer and their residents better able to afford medical treatment and healthier living conditions, lifespans are growing.

As with poverty, progress in life expectancy is becoming more gradual. In the 1960s alone, global lifespans shot up by over 8 years. We’ve picked enough of the low-hanging fruit that we’re not going to see progress that fast again. But we still have a ways to go in preventing deaths from easily treatable diseases, which means that global life expectancies have not stopped their rise yet.

In its 2017–2018 survey, the CDC’s authoritative National Health and Nutrition Examination Survey found that the US adult obesity rate was 42.4 percent. (Obesity in the survey means having a body mass index — a metric that has come in for its share of criticism — of 30 or more.) While the agency could not finish its scheduled 2019-2020 survey because of Covid-19, it took the data it had collected by March of 2020 and combined it with the previous cycle to update their estimates. 

They found that the obesity rate among US adults over 20 had fallen to 41.9 percent as of that terrible spring — a time before the market for Ozempic and a promising new class of anti-obesity medications had begun to grow exponentially.

There is still plenty of uncertainty about what the long-term health effects of these drugs will be. The future holds too many unknowns: How many people get prescriptions? Will patients stick with the drugs? Will they be able to afford them? 

But Wegovy et al. are about to become even more commonplace than they already are, now that Medicare has decided to cover them for heart conditions, which about four in 10 Medicare patients have. Millions of prescriptions will be written. 

New drugs are also in the works: Ozempic was a breakthrough because it was “just” a weekly injection (and very effective). Pill forms of semaglutide are now being tested for weight loss. Even if some people do stop taking their medicine, one large new study (not yet peer-reviewed) suggests many of them may be able to keep the weight off.

There have been momentary dips in obesity before, while the long-term trend has stayed stubbornly upward. In 2000, the adult obesity rate was 30.5 percent. Now, it’s north of 40 percent.

But extrapolate the improvement from 2017–2018 to 2019–2020 and the United States could potentially make up a lot of that ground over the next decade. And this time, we have the most effective anti-obesity medications ever developed.

In 2022, the US hit a grim milestone: It was the first year the country had slaughtered more than 10 billion land animals for food. I predict that in 2033 it will slaughter 11.5 billion or more land animals. 

It’s not so much my prediction — earlier this year, the US Department of Agriculture published its meat, dairy, and egg production projections through 2033. The agency didn’t project the number of animals that’ll be raised for food each year but rather the amount of meat, eggs, and dairy that will be produced by weight, which I converted into the number of animals that’ll be slaughtered.

I’m putting my confidence level at 70 percent because there are a number of unpredictable domestic and international factors that could influence the final number: economic recessions, zoonotic diseases (like the ongoing bird flu outbreaks), consumer trends, the cost of farming inputs (like fertilizer and animal feed), and agricultural and trade policy.

But the most important factor, by far, will be demand for poultry, as chickens comprise over 90 percent of the total number of animals raised for food in the US. And that’s expected to continue to rise well into the next decade — and half-century. 

The USDA also projects that US per capita consumption of red and white meat combined will rise from 226.8 pounds in 2022 to 235.4 pounds in 2033 — a near 4 percent increase.

Given the enormous ecological and climate impact of meat and dairy production, this is all headed in the opposite direction of what climate scientists say we must do: drastically reduce livestock numbers and transition to a plant-rich food system. 

If the US were to align its agricultural policy with its climate policy, we’d move in that direction instead. I’m not a betting man, but if I were — given the power of the meat lobby — I don’t expect that to happen by 2034.

Over the next decade, cutting-edge technology will open the floodgates for human augmentation, which could fundamentally change what it means to be human. From CRISPR to biohacking, there are a growing number of methods people might want to use to enhance themselves. 

But I want to make a prediction specifically about brain-computer interfaces (BCIs): By 2034, someone will have an implantable BCI not to help with paralysis or disease, but simply because they want to be a smarter or faster version of themselves. 

So far, the FDA has greenlighted brain chips for people with medical conditions like paralysis, allowing them to type or text with just their thoughts. But it’s no secret that some of the makers of these chips — like Elon Musk’s Neuralink — want to go further. Musk has said that the ultimate goal is “to achieve a symbiosis with artificial intelligence.” And Neuralink is explicit about its dual mission: to “create a generalized brain interface to restore autonomy to those with unmet medical needs today and unlock human potential tomorrow.”

BCIs exist on a spectrum of invasiveness, with Neuralink at the most extreme end (the device is implanted directly into the brain) and wearables on the other end (they use EEG sensors and other tech that sits on the scalp or skin, so there’s no need for surgery; some are already on the consumer market). 

Between these two ends of the spectrum, there are BCIs that are less invasive than Neuralink’s but are nevertheless implantable — they may go in the skull, but not in the brain itself. This type, not a Neuralink device, will probably be used for enhancement purposes first, and I will consider my prediction right if this type is in use by 2034. 

I know even that may seem extreme. And I wouldn’t hazard such a guess in the realm of, say, CRISPR, because when it comes to changing our genetics the scientific community has been appropriately cautious in calling for moratoria. But if the ruling mantra in biomedicine is “first, do no harm,” the ruling mantra in tech is “move fast and break things.” And with companies like Meta and Apple also exploring brain-reading technology, appetite is growing: The global BCI market is expected to top $8 billion within the next 8 years. So I think this is a solid possibility.

My Future Perfect colleague Dylan Matthews, who is the best forecaster I know, wrote recently that one key to accurately predicting future events is to first establish a “base rate.” A base rate is the rate at which some event has been known to happen in the past, which is a useful starting point for estimating how likely it is to happen in the future.

The base rate for nuclear weapon use in war in a given year is, thankfully, very low. (If it weren’t, I likely wouldn’t be here writing this.) Nuclear weapons have been available to at least one country since the US successfully tested the first atomic bomb at Trinity Site in New Mexico on July 16, 1945. In all the time since, an atomic weapon has been used in war twice — in 1945, by the US at Hiroshima and Nagasaki. 

That’s one year out of 79, which equates to a base rate of 1.2 percent. Extend that over the next decade, and the base rate grows to a little more than 11 percent. And even that’s likely putting it high — for the last 78 out of those 79 years, nuclear weapons have been used in war precisely zero times, which should make us more confident that we’ll get through the next decade nuclear-free.

So why do I think there’s a 1 in 5 chance we’ll see a nuclear weapon deployed in war over the next decade? It’s based on my reading on trends in the future of war, arms control, and international relations — all of which are moving in increasingly dangerous directions.

War between states, which had basically ceased over the past several decades, is back once more. Russia’s invasion of Ukraine in 2022 made nuclear weapons a live issue again, in part because Vladimir Putin was not shy of making vague atomic threats, but also because nuclear weapons ultimately defined the contours of the conflict. 

There’s a limit to how much NATO can defend Ukraine precisely because Putin controls the single biggest nuclear arsenal in the world. And Putin himself is constrained — though how much he purposefully makes ambivalent — by NATO’s own nuclear weapons, and its stated policy of treating an attack on one member state as an attack on all. So far the balance has held, but that’s a risky place to be.

And even as the war in Ukraine raises the atomic temperature, the arms control treaties that have successfully reduced the nuclear threat since the end of the Cold War are unraveling one by one. The New START treaty, which aimed to constrain the tactical and strategic nuclear arsenals of both the US and Russia, is currently set to expire in 2026. 

Both the US and Russia have withdrawn from the Intermediate-Range Nuclear Forces Treaty, which reduced short-range nuclear missile numbers. Given the open animosity between Russia and the West — plus the growing arsenal of China, which is mostly outside the current arms control regime — and you have a three-body problem that spells heightened nuclear danger.

Most worrying of all, the nuclear taboo seems weaker than ever. Over the past several months, Iran — a country that is perpetually on the threshold of nuclear weapons — launched attacks on two nuclear-armed countries, Pakistan and Israel. North Korea, which has resisted every attempt to constrain its arsenal, now has dozens of nuclear weapons, many of which are capable of striking the US. Should Donald Trump return to the White House and make good on his promises to close the US nuclear umbrella, the risk will grow that countries in Europe, the Middle East, and Asia will seek their own nuclear weapons. That future, as President John F. Kennedy said in 1961, will be one where we all live under a “nuclear sword of Damocles.”         

I usually hope my predictions will turn out correct. This one, I pray, won’t.

In May 2022, the UN refugee agency reported that the number of people forcibly displaced from their homes globally had passed 100 million for the first time in history. Since then, the number has only grown, reaching 114 million people as of early 2024, including those displaced within their countries and those forced to flee across international borders.

The causes are manifold, with war as a major driver. More than seven in 10 international refugees come from just five countries — Syria, Ukraine, Venezuela, Afghanistan, and South Sudan — most of which are embroiled in some kind of serious conflict. For countries like Cuba, severe economic disruption can force people to flee in search of a better life, or just survival.

Environmental factors — including climate-driven droughts, famines, and wildfires — are driving others to leave their homes in places like southern Africa. And human rights abuses, such as those perpetrated against the Rohingya minority in Myanmar, are another factor.

Predicting that the number of displaced people will reach 200 million between now and 2034 means assuming that the total will rise at least 75 percent. I think that’s a more than reasonable bet. For one thing, while global population growth has been slowing, it’s still increasing a little below 1 percent per year, which means the world will likely have hundreds of millions of more people by 2034. More importantly, the bulk of that growth will be in places like sub-Saharan Africa — a region that is already highly vulnerable to climate disasters and political disruption. That means more people will likely be in harm’s way, even as that harm continues to grow.

Nor do I expect the rich nations for displaced people are trying to reach to suddenly become more welcoming — even if, precisely because fertility rates are falling so rapidly, it would likely be in their long-term economic interest to do so. The very fact that President Joe Biden is considering stringent border policies that sound like something out of the Trump playbook underscores just how toxic migration has become as a political issue. I don’t see that changing.

Lastly, another driving factor is actually a sign of global success. It sounds paradoxical, but it’s true — economic growth in poor nations can increase outward migration. As even the poorest people in the world become richer than they used to be, more people will have the means to escape their countries.

The last decade has seen a startling increase in displacement and migration. I’m willing to bet that the next decade will see even more. 

Finally, some actual good news.

With the exception of 2020, when pandemic lockdowns froze economic activity, energy-related greenhouse gas emissions have mostly gone in one direction: up. Last year, global energy emissions reached a record high of 37.4 billion metric tons of CO2.

But the peak is in sight, and it may be coming sooner than you think. The International Energy Agency (IEA) has predicted that energy emissions will top out in 2025, a forecast echoed by the energy intelligence company Rystad. Behind those estimates are predictions that global coal consumption — which has been at a record level, driven by demand from developing countries — will peak in the next few years, as will demand for oil.     

Together those two fuels contribute more than 50 percent of global energy emissions. While demand for natural gas continues to grow, as lower-carbon gas substitutes for coal, it’s still a net win for emissions. Add that to the rapid increase in solar and wind energy, along with renewed growth in nuclear energy, and there’s a more than reasonable case that the end of a multi-decade era of growing energy emissions could finally be over. 

Of course, nothing is guaranteed. As artificial intelligence ramps up, the demand for electricity to power all those generative searches and data centers could skyrocket. Already, after years of relatively flat growth, US electricity demand is projected to spike, due in large part to AI demand. And there are other potential obstacles; should the EV revolution falter, oil demand could continue to rise for years into the future. The renewable energy revolution in the US will require an equally revolutionary shift in how the country permits energy projects — and that’s far from guaranteed.

Nonetheless, I’m willing to bet that energy emissions will peak, and soon. That doesn’t mean we’ve won the war against climate change — not by a long shot. Emissions will have to peak soon and then drop, rapidly, to avert some of the more dangerous scenarios around climate change. Still, for as long as I’ve been alive, the indicators around climate change have only gone in one direction: worse. Don’t underestimate the psychological benefit of finally turning one around.

The global car market is on the cusp of its greatest revolution since, perhaps, the Model T first brought automobility to the American middle class. To decarbonize the transportation sector, the world’s second-highest greenhouse gas emitter, governments are pushing to speed up the availability and mass adoption of electric cars. 

Last year, according to preliminary data, about 18 percent of new cars sold around the world were electric, up from 14 percent in 2022 and less than 5 percent in 2020. The world’s highest electric car adoption rates are in China, where about a third of new cars are electric, and in Western Europe, but US sales have been rising precipitously, too, hitting 10 percent of new cars in 2023, more than four times the share just two years prior. 

EVs’ year-over-year share of new car sales give us a sense of how quickly they’re going mainstream, but it’s the overall makeup of cars on US roads that matters for carbon emissions. At the end of 2022, EVs made up a more sobering 1.2 percent of the national car fleet. 

As new federal and state fuel economy rules aiming to sharply reduce vehicle emissions take effect, that number will climb. One leading estimate projected that 10 percent of cars on American roads will be electric by 2030. To reach that, we’d need to add more than 20 million new EVs over the next six years. I don’t think it’ll happen that quickly, but I do think the share will hit 10 percent within a decade. 

With interest rates elevated until who knows when, Americans are holding on to old cars for longer than ever — and the average age of cars on the road has been steadily increasing for decades, a trend that’s unlikely to go away. There have also been recent signs, as Vox’s Umair Irfan has reported, that Americans’ enthusiasm for EVs is flagging. 2023 US electric car sales disappointed the expectations of some forecasters, as consumers report being wary of the high price tag, lack of convenient charging stations, and reliability issues, particularly in cold weather. This year, following concerns from the auto industry, the EPA’s new car emissions standards were downgraded from what the agency proposed last year.

The hurdles — like equipping power grids to handle tens of millions of massive EV batteries — are considerable. But mass adoption will depend on carmakers’ and policy leaders’ will to make electric cars no less cheap or convenient than their fossil-burning counterparts.