Earlier this week, Lawrence Hamilton of the University of New Hampshire published new polling evaluating the extent to which Americans agreed or disagreed with a number of statements. Of the 14 statements, seven were reflections of scientific consensus — or reality, as we understand it. The other seven required belief in or acceptance of a conspiracy involving scientists and/or the government.

On the whole, Americans were less likely to agree with the conspiratorial statements than the scientific ones. The conspiratorial statement with which the most Americans agreed was about the government hiding evidence of UFOs. Fewer people agreed with that than agreed with the least-supported scientific statement, that the Sun shines for 24 hours a day during certain times of the year at the South Pole.

Disconcertingly, Hamilton also found that Americans had grown more conspiratorial since 2021. There was a statistically significant increase in the percentage of people who agreed that the Moon landing was faked, for example, and a significant decrease in agreement with evolution, the age of the Earth and even, weirdly enough, that the Earth orbits the Sun.

We can probably attribute some of this to the general embrace of nonsense that we’re seeing in the moment. Some of it is also probably linked to the political shift since 2021, the country’s shift to the right. After all, as Hamilton writes, “conservatives tend to be more open to conspiracy beliefs about non-political as well as political topics.”

That’s shown in the data. Asked if the government has the ability to control hurricanes, which it doesn’t, 14 percent of Americans said it does. Among those who voted for Kamala Harris last year, only half that percentage thinks that the government can control hurricanes. Among Trump voters, though, nearly 1 in 5 do.

The less tied to traditional politics the respondent, the less robust the rejection of the conspiracy theory. Only about half of third party voters and those who didn’t vote last year said the government can’t control hurricanes.

One can debate which way the causal arrow points here, but it isn’t particularly surprising that confidence in the political system (as manifested in vote choice) correlates to distrust of government. Nor is it surprising that conspiratorial thinking increased as Trump’s political power resurged.

What is a bit surprising is how that conspiratorial thinking has seeped into actual policy. The most conspiratorial people have historically been kept outside of the political system but, in part because Trump and his base embrace conspiracies as a means of consolidating power, even unpopular ideas like “chemtrails” and rejecting vaccines have become the targets of administration or congressional action.

America is being devoured by conspiracy theories — a statement that 1 in 10 Americans might think was meant literally.

Photo: Airplane contrails over England during World War II. (National Archives)

In keeping with Americans’ perennial inability to properly contextualize the events of the recent past, there’s a particular narrative about the 2024 election that I find frustrating. In short, that the election marked a significant rightward shift of American voters and, by extension, by America.

The value of such a narrative to President Trump and to his ideological allies is obvious. Trump’s return to the White House brought with it the unexpected cudgel of popular support, something he lacked when he arrived in 2017. But that cudgel was flimsier than it appeared. Trump won more votes than Kamala Harris but less than half of votes cast. A central driver of those votes was inflation, not Trump’s politics.

That rightward shift also followed a leftward shift four years prior. Yes, states shifted to the right from 2020 to 2024 — but only after they shifted to the left from 2016 to 2020.

This is little consolation for Democrats, certainly, and it glosses over the actual shifts in the electorate between Trump’s first and third bids for the presidency. So let’s use Pew Research Center’s robust assessment of the electorate in each year to show how Trump’s base of support evolved over that time … and how it didn’t.

Take gender. In 2016, 53 percent of Trump’s support was male, compared to 54 percent in 2024. That’s a subtle change, but one that looks more substantial when you consider the intervening election. In 2016, the pool of Trump voters was 6 points more male. In 2020, the gap was half that size — before swelling to 8 points last year.

Over those eight years? A bit more male and a bit less female.

(Why was the pool of Trump voters more heavily female in 2020? In part because more men voted for Joe Biden. In 2016, Trump won men by 11 points. In 2020, he won them by 2 points.)

When we look at age, we see a more obvious shift. In 2016, about 20 percent of Trump’s support was under the age of 40. Last year, about 26 percent was.

If one ascribes to the thesis that Trump’s decade-plus as the head of the GOP has normalized his politics among younger Americans (as I do), feel free to use the chart above in your defense.

Below, you can see how Trump’s coalition shifted by age and gender over the three elections. Women aged 30 to 49 were 10 percent of Trump’s support in 2016. Last year, they were 13 percent — which was still a relatively modest part of the coalition.

One of the most important distinctions (and one not captured in exit polling) is how much less White Trump’s 2024 coalition was than his 2016 one.

Data on race don’t add up neatly to 100 percent, so let’s compare data across election cycles by stacking racial groups. Below, you can that White men and women made up less of Trump’s support in 2024 than in 2016 (by four and six points, respectively) and that Hispanic voters in particular made up more of it: 6 percent in 2016 compared to 1 in 10 Trump voters last year.

When I was still at The Washington Post, I wrote about how the shift to Trump among young men was heavily among non-White men — in part because younger Americans are less likely to be White. Below, you can see that reflected, to some extent. We’re slicing demographics pretty thin here, so take this with a grain of salt.

More striking is the shift by race and education. Trump’s 2016 coalition was heavily centered among Whites without a college degree, who made up almost two-thirds of his support. Last year, it was barely over half.

That decline was seen among both White men and White women without degrees.

There’s a pattern here: Whites, Whites without degrees and Whites over the age of 50 — groups that overlap heavily — all constituted at least 10 percentage points more of Trump’s coalition in 2016 as they did last year.

The question for the Republican Party is whether they can maintain that more-diverse coalition. Which, in turn, is a question about the roots of that shift: how much of it was about Trump and how much of it was about broader, lasting trends and shifts in American politics?

There’s one other question, too. If the 2028 election shows a shift back to the left, will people remember that this happened in 2020, too?

Photo: Harry Truman and his daughter Margaret Truman voting in Independence, Missouri, 1946. (National Archives)

I suspect that history will record 2025 as an inflection point in American (if not world) history, the point at which the promise of a future rooted in universal access to knowledge collapsed under the weight of infinitely accessible fictions. And if a debate emerges over the moment that embodies that collapse, I would like to offer as a contender the murder of Melissa Hortman.

Hortman is the Democratic legislator in Minnesota who was shot to death in her home by a man masquerading as a police officer. His disguise was uncomplicated, involving a clunky mask, ersatz uniform and vehicle adorned with flashing lights. The get-up was convincing enough that, when questioned by an actual police officer, the assassin was able to simply drive away.

In the abstract, this charade doesn’t have much to do with the broader collapse of objectivity. But the attack was contemporaneous with a bit of performative costuming that served the exact opposite purpose: the habit of immigration enforcement officers to hide their affiliations and to mask their faces. In Minnesota, a criminal dressed as a police officer to commit a crime. Everywhere else, law enforcement officers were dressing as criminals to enforce the law. Encountering a cop on the street might mean you anre encountering a criminal and encountering an armed, masked man might mean you are encountering a cop. The line is blurry.

This blurriness between real and fake is the defining characteristic of the moment.

That’s centrally because we have a president, Donald Trump, whose politics depend on that sort of blurriness, on generating uncertainty about otherwise certain things. This is what wanna-be authoritarians do, as Hannah Arendt explained more than 50 years ago.

“In an ever-changing, incomprehensible world the masses had reached the point where they would, at the same time, believe everything and nothing, think that everything was possible and that nothing was true,” Arendt wrote in “The Origins of Totalitarianism.” “… Mass propaganda discovered that its audience was ready at all times to believe the worst, no matter how absurd, and did not particularly object to being deceived because it held every statement to be a lie anyhow. The totalitarian mass leaders based their propaganda on the correct psychological assumption that, under such conditions, one could make people believe the most fantastic statements one day, and trust that if the next day they were given irrefutable proof of their falsehood, they would take refuge in cynicism; instead of deserting the leaders who had lied to them, they would protest that they had known all along that the statement was a lie and would admire the leaders for their superior tactical cleverness.“

We hardly need to say more than that to implicate Trump and his political and media allies. Nearly every facet of Trump’s public commentary adheres to the idea that truth is malleable, from — to pick just a few examples from this week — accusing Barack Obama of treason to claiming that drug prices are down an impossible 1,500% to insistences that unfavorable jobs numbers are a function of bias after having touted the same data as proof of his skilled leadership. If you need more convincing that Trump seeks to build and, in large part, has built an informational environment in which reality is blurred, you are probably already comfortably ensconced in the cynical world Arendt describes.

But it is overly simple to attribute all of this to Trump. While he benefits from and amplifies uncertainty and blurriness, he didn’t bring it to politics, much less to America. He saw, or stumbled onto, or — most accurately — was mired in an anti-objectivity strain of right-wing thought that has decades-old roots in anti-intellectualism and media hostility. During the Obama administration, this strain built enough momentum to be politically viable, though most Republican officials still operated under the assumption that it wasn’t or wouldn’t be for long. Trump simply proved them wrong.

The momentum he leveraged was a product of explicit efforts to blur reality, as demonstrated by Glenn Beck and others. It was also a byproduct of the emergence of the social web, a place where anyone could assert anything and amplify anyone else’s assertions. The pool of universally accessible information that was the promise of the internet was polluted with nonsense — and often nonsense that flattered people’s existing biases. The idea (and ideal) that people could become informed by looking up anything they wanted was bastardized by people seeking out only what they wanted to hear. They did their own research only to the extent that they unearthed claims that sounded true enough for them to believe what they wanted. (I’ll again point to this prescient 2009 essay on the subject.)

This is the essence of the challenge faced by traditional media outlets. People, on both the left and the right, are often less interested in an objective reality than a subjective one. It becomes a simple economic problem: demand for what news organizations do has eroded and so have the economics.

The surge in demand for bespoke realities has surged, powered by and powering a galaxy of independent voices proffering news-like information. Scrolling social media means slogging through a morass of claims and accusations rhetoric that may or may not be reliable and may or may not be true, with occasional appearances (if you so choose) by archaic institutions that actually worry about things like probability and accuracy.

For a time — what we now understand as a period of transition — social media platforms tried advantage reliable sources of information by policing falsehoods and verifying reliable actors. But agreed-upon reality was precisely what the demand was working against, so — just as Fox News shifted to fend off competition from its right 15 years ago — the social media companies (and some news organizations) embraced a blurry, subjective presentation of reality.

An important point that bears highlighting: blurriness yields economic benefits as well as political ones.

This was the state of play even before the emergence of AI. It was already the case that random Instagram accounts could make claims about healthy eating that shifted public habits even when those claims were utterly baseless. It was already the case that a guy who dedicated his life to undermining the broad, proven consensus on public health could be tapped to run the federal government’s health infrastructure after parlaying his base of own-research-doing skeptics into a bloc of voters that Trump found valuable. And into this space came AI slop.

It is clear that there are some valid uses for AI, whether or not one thinks that those uses are worth the costs in energy or shifts to the economy. But it is more abundantly clear that the emergence of AI is dramatically hastening the collapse of shared reality, to the extent that such a thing even exists.

This is the aforementioned slop. Nonsensical, artificial images aimed at social spread or making rhetorical points. False news stories created for the same purposes. The garbage that follows from AIs trained to generate authentic-sounding answers to questions generating answers that have no grounding in reality. This use of AI is an enormous blurring filter over the world.

There’s overlap with Trump’s political project. He, his administration and his allies use AI to generate a fake reality and to skew the existing one. AI images show a buff, conquering Trump astride the world or his opponents — immigrants, the media, the left — as cowering simps. It’s hard for Trump’s base and American voters to see what Trump is through the lens he and his allies present, which is entirely the point.

Again, though, the threat extends past Trump. The internet is already evolving into a mix of true and false, real and generated, both willfully and accidentally. Querying the world’s accrued knowledge through a web browser was never guaranteed to result in an accurate, concise response, but first with the social web and now with AI it’s become far harder to differentiate truth from fiction. In each case, users and robots jostle to get between you and the truth in order to convey their reality or absorb the currency of your attention.

I began here by suggesting that 2025 would be recognized as an inflection point. In other words, this doesn’t feel like an aberration, a step away from shared reality before we step back toward it. Instead, it feels as though this is the state of the world moving forward: a bit of reality buried under a lot of garbage, with scavengers picking through the dump to sell what they can.

Photo: Earth observation views of city lights at night taken during STS-76 mission from shuttle orbiter Atlantis. (National Archives)

OK, fine. I’ll address President Trump’s justification for the Texas Republican Party’s plan to shift five House seats from Democrats to the GOP.

It was offered in a phone interview with CNBC and host Joe Kernen.

Trump: "We have an opportunity in Texas to pick up five seats … I got the highest vote in the history of Texas as you probably know. And we are entitled to five more seats."

— Aaron Rupar (@atrupar.com) 2025-08-05T12:38:40.297Z

“I won Texas — I got the highest vote in the history of Texas, as you probably know,” Trump said, “and we are entitled to five more seats.”

“I don’t know,” Kernen replied. Trump (calling him Brian, perhaps confusing him with Fox News host Brian Kilmeade) suggested he run it past the fact-checkers, giving the always accommodating Kernen a chance to complain about how fact-checkers had given Joe Biden a pass while scrutinizing Trump closely.

There was some irony to his doing so! After all, the reason that Trump was the focus of so many fact-checks relative to Biden was that Trump was so much more likely to present false facts than his successor/predecessor. Just like Trump’s historic vote total in Texas was a function of the state’s much larger population.

It is true that Trump got more votes in Texas than any candidate in history. You can see that below, using data from U.S. Election Atlas.

But that is in large part because there are more people living in Texas than ever before and, by extension, more people casting votes than in any prior election in the state.

As a result, his was not the largest margin of victory in Texas. That honor goes to George W. Bush’s 2004 reelection bid. Bush got fewer total votes, but won by a larger vote count.

If we consider margins of victory in terms of the percentage of votes cast, Trump fared unusually poorly in the state. He had the 27th highest percentage-point margin of victory in 2024, substantially behind Franklin Roosevelt’s 1932 victory.

This context is beside the point for Trump, just as population growth was beside the point of his first-term boasts about the number of people working. It was true that more people were employed than at any prior point — but it was also true that the population was larger than it had ever been. During Biden’s administration, the number of people working kept setting new records, too, but this was not presented by Trump as a demonstration of his excellence at managing the economy.

Nor did Biden take advantage of his unusually strong performance in Texas — the second-highest vote total on record! — to demand that the state redraw its House boundaries to offer more seats to the Democratic Party. After all, the point of House seats is to reflect local, constituent preference, not statewide opinion on the chief executive.

For Trump, though, everything is a referendum on him — unless the numbers are bad.

There’s a car in my neighborhood (and perhaps in yours, too) that sports a confounding pair of bumper stickers. One advocates for making America great again, an idea recently popularized by marketing mogul Donald Trump. The other endorses fealty to the U.S. Constitution, an idea recently eviscerated by President Donald Trump.

I am obviously aware that “we must adhere to the Constitution” was a GOP mantra for the past 20 years or so, given how often it was invoked in opposition to Barack Obama’s tremulous expansions of executive power. But it is rather jarring to see this idea presented alongside explicit support for the president who’s doing everything in his power to relegate the Constitution to second-tier status.

Last month, YouGov conducted a poll to measure how Americans viewed the concept of “American.” What was it that made someone “American,” exactly, a status that traditionally has been applied more generously than, say, “French” or “Nepalese”?

The second-most commonly agreed upon characteristic, tied with being a citizen, was “supports the U.S. Constitution” — something that 86 percent of respondents believed was very or somewhat important to being considered an American. The most commonly agreed upon characteristic? Obeys U.S. laws.

The pollsters also asked respondents if they themselves met the conditions which were part of the survey. The result is fascinating, with views of what constitutes being an American generally increasing along with the characteristics held by the respondents themselves. You can see that below; the larger the number of people who indicate they possess a characteristic, the larger the percentage who identified that characteristic as being very or somewhat important to being an American.

Dots that sit above the diagonal line are characteristics that were more likely to be seen as important for being American than they were possessed by the respondents themselves.

The YouGov poll also allows us to break down responses by party. If we take the same measure as above and look only at Democrats, this is the result.

Notice how low “supports the U.S. president” sits on both axes.

Where things really get interesting is when we look at the responses among Republicans.

Notice that, across the board, Republicans are more likely to identify the presented characteristics as very or somewhat important for American-ness than are Democrats. In fact, we can compare those directly. On the chart below, any dot that sits above the diagonal was more likely to be seen as important by Republicans than by Democrats. You will notice that all of the dots sit above the line.

The two that sit closest to the upper right corner are, again, adherence to laws and to the Constitution, neither of which seem to be Trump’s strongest suits.

I haven’t yet discussed the most striking finding in this poll, however, the one that I mentioned in the headline. YouGov included “being White” as a possible option for being important to being American and nearly one-fifth of respondents said that it was. Among Republicans, 3 in 10 did, with a fifth saying it was “very important.”

On the plus side, I haven’t yet seen a car with an “only White people can be Americans” bumper sticker. Here’s hoping that this continues to be a sentiment that people are unenthusiastic about endorsing so publicly, however energetically the Trump administration may be putting it into practice.