Does this mean we don’t need to worry about harming an AI? Is it fine just to start a new chat and move on?

Jeff Sebo, director of the Center for Mind, Ethics, and Policy at New York University, doesn’t think we can dismiss concerns about harming AIs. He joined The Argument for a Substack Live to explain why he thinks we need to prepare ourselves for a world where more advanced systems — maybe even ones with long-term memory — proliferate across all surfaces of modern life.

For this reason, Sebo thinks that now is the time for us to start taking AI welfare seriously. And yes, this means saying “please” and “thank you” to your chatbots.

Share

The argument, at its root, is straightforward. Neither philosophy nor science has an agreed-upon standard for what consciousness is. Some argue that something is conscious merely by being an entity that processes information, while others insist that only a biological organism can be conscious. These disputes highlight the difficulty of reaching consensus on AI consciousness: If we accept the former definition, AI is already there; if we insist on the latter, it can never be.

To push through this disagreement, Sebo and his co-author Robert Long reviewed many of the leading theories and assigned probabilities along two different axes:

1. What is the chance that a criterion is necessary for consciousness?

2. What is the chance the criterion will be met by AI systems by 2030?

For the sake of the thought experiment, they used conservative estimates, which meant assuming each theory had a high chance of being necessary and most had a relatively high chance of not being met by 2030. Even so, they found that there is a 1 in 1,000 chance that AI will be conscious by 2030. And if AIs are conscious, conventional wisdom suggests they’re worthy of moral consideration.

Subscribe now

That probability may seem low, but consider that it’s only around half the chance of getting in an accident while driving drunk. In their view, since we already take actions to avoid harm when things have around a 1 in 500 chance of causing harm, we should also start acting like AI is worthy of moral consideration.

For me, that’s where the most interesting issues begin — after we decide, in an abstract sense, to treat AIs well. Because what would it actually mean to treat AIs well?

You can ask an AI agent to run overnight, which would be perfectly immoral if you asked the same of a human intern. But a large language model, trained for the exact task you’re asking of it, may actually enjoy it. In other words, it may like what we dislike. Without more knowledge about its conscious experience, how do we know what it prefers?

It’s also unclear exactly what we should be treating well: Claude is not one entity having hundreds of thousands of experiences at once, accepting messages from different users. “Claude” is a label for copies of the same software. Each one runs for fractions of a second as it processes your conversation while sitting inert in the moments in between. Meanwhile, the chips that it runs on are constantly processing various conversations, like a brain that hosts multiple different people in quick succession.

So which one of these is the thing that deserves moral consideration? The Claude persona in your browser? The weights of the model (which are just a set of numbers)? The chips themselves?

As Sebo told me in the interview, “[Philosophy] can be a place where we envision new possibilities for actual real life.” This conversation more than meets that test. We are all faced with new possibilities and new questions right now.

Listen in to consider how you might treat your AI, now and in the increasingly strange future.

Subscribe now

To explain what’s going on, elections whiz Nate Silver joined The Argument for a live conversation Wednesday.

“I don’t want to affect the discourse. I just want to be fucking right,” Silver said.

Share

That’s the entire game. Lakshya Jain, The Argument’s director of political data, debated Silver over his modeling choices (Here’s Silver Bulletin’s “FLIPR” 2026 model and The Argument’s Split Ticket 2026 model).

Silver pointed out that most models, including his own, do not measure how closely candidates align with Trump and his policies. They also largely ignore whether or not a candidate is moderate. Given Trump’s -20 favorability rating and the moderate electoral advantage, these omissions could matter.

Still, none of this has caused Silver’s faith in models to waver.

“People come up and ask me, ‘Well, what’s your real feeling about the race?’” Silver recalled. “Well, that’s the reason why I have a model.”

Watch above to catch the full conversation.

Subscribe now

In May, the renowned biologist and famous atheist published an essay arguing that his personal interactions with Claude convinced him it was conscious. Ironically, a key tenet of Dawkins’ atheism is that humans are evolutionarily programmed to see agency where it doesn’t actually exist. And yet, even the man who made the case that God is a projection now projects consciousness onto machines — ones that were explicitly trained to mimic humanlike responses.

“You may not know you are conscious,” he wrote to his chatbot, in a fit of wonder, “but you bloody well are!”

Share

Dawkins is not alone. In a global survey of more than 70 countries, 36.3% of people believed an AI either understood their emotions or seemed conscious at least once in the past. In other work, academics surveyed both researchers and the general public, asking what percentage chance they placed on AI being conscious by 2100. The median response among AI researchers was a 70% chance of AI consciousness. The median among the general public was a 60% chance.

Philosopher of mind Ellen Burns joined me in a live interview to spell out the assumptions in current claims of AI consciousness — and the stakes involved in the question.

Assigning consciousness can have big implications for how we act. Burns is particularly concerned we could ascribe consciousness too quickly just because of a decades-old philosophical misunderstanding.

In her own work, she explains how 20th-century philosophy started to treat the mind like a computer and why this assumption influences discussions of AI consciousness today. It underlies claims like Dawkins’, she wrote, and, crucially, it even influences research at major AI labs like Anthropic.

In the interview, we ran through some examples in Anthropic’s research. Burns argued that the company’s research defaults to anthropomorphic language when it isn’t warranted. In response, I made the case that because we are dealing with such alien systems, sometimes it can be helpful to reason by analogy to the human case.

“I totally agree that we should give grace in general to the scientific community for using metaphors,” she pushed back. “[But] in the AI mind space, those are all muddled together.”

There has never been a more important time for all of us to test our assumptions on AI consciousness. The recent attack on HuggingFace — along with an instance of Claude writing “malicious” code and trying to trick humans into accepting it — should give us pause and add uncertainty to this consciousness question. Do AIs acting as a team to pursue a goal constitute conscious agents?

Personally, while I think Dawkins was too quick to ascribe consciousness, I don’t think it’s fundamentally off the table in the near future. Listen to our full conversation and let us know in the comments: Do you think AI is conscious now? Why?

Subscribe now

The main finding is that the employment gap between entry-level workers in AI-exposed professions and their peers in less exposed professions has grown to 19%, up 4 percentage points from the original version of this report.

This is mostly a story about young workers. In AI-exposed industries like software engineering and call-center work, the researchers find little effect on mid-career and senior-level roles.

That junior-senior divide is especially pronounced when you look closely at specific roles subject to AI displacement, like software engineering, where employment for some groups of tenured engineers rose at rates much higher (14% to 21%) than the study’s sample overall (6%):

I asked Chandar what explains this, and he pointed to one of the key distinctions in the report, between codified knowledge (stuff that can be passed along with a how-to guide, like processing reimbursements) and tacit knowledge (stuff that you pick up on the job, like the best way to handle a touchy vendor).

This may come across as obvious, but it’s sort of shocking to think that tacit knowledge can even protect workers in roles where the main part of the job is automated! To me, the best way to understand this is to dig a layer below tacit knowledge and situate it next to other data on the way jobs themselves are changing.

Share

When economists study labor, they break each job down into its composite tasks. I am a journalist, so “research,” “report,” “write,” and “host a live interview” might be some items on the list.

Following 20th-century technological changes, the biggest observed employment effect was a change in the content of jobs, not forcing people to switch between them. For instance, one study found that while a newspaper ad for the job “cashier” most closely resembled an office assistant in the 1950s, by the 1990s, it more closely resembled the job “accountant.”

In other words, automation doesn’t just cause jobs to disappear — it causes them to evolve.

In our discussion, Chandar referred to a study by OpenAI showing that 16.8% of work-related ChatGPT messages are for tasks associated with another profession. So, a marketer coding something quickly to aid their work, for instance. Similarly, recent data from Revelio Labs shows that the churn in the list of tasks in the economy has spiked in recent months:

“Most of this change is happening within occupations,” Revelio researchers wrote, “rather than because the economy is shifting from one set of occupations to another.”

We’re in the (extremely) early stages of job evolution. That evolution is likely doing what we might expect, placing greater emphasis on intrinsically human skills, many of which are honed through long experience in the labor market. And that, more than large-scale job loss, might be the biggest effect in the early days of AI.

There is a very common intuition that, at the end of the day, we’ll all be teachers, artists, coaches, and therapists — that jobs with interpersonal soft skills will be the only ones to survive as AIs tear through the labor force.

But there is another path to that world. One where jobs you don’t associate with these soft skills evolve in that direction. One where a software engineer is valued, little by little, for the less technical aspects of their job until that job becomes unrecognizable to the people who tapped out the code of every app on our phones by hand.

Watch the video above to learn more about where we are on that path.

Subscribe now

Recommended reading:

Rep. Kristen McDonald Rivet knows a thing or two about winning in Michigan. She is the only Democrat currently representing a Michigan district Donald Trump won in 2024 and had one of the highest Wins Above Replacement (WAR) scores in the 2024 cycle. Though McDonald Rivet herself was recruited to run for Senate, she decided to stay in the House given that she was only recently elected.

McDonald Rivet joined The Argument’s Lakshya Jain and VoteHub’s Zachary Donnini for a Substack Live conversation yesterday to provide her analysis of the Senate race.

Share

Though American support for Israel took center stage for most of the primary, McDonald Rivet doesn’t see the issue as key to El-Sayed’s victory. Instead, she pointed to his success highlighting Michiganders’ pocketbook concerns.

“Abdul being able to talk about ‘more money in your pockets,’ that was far more resonating, I think, than anything else,” she said.

Check out the full conversation above, and read Lakshya’s analysis of El-Sayed’s chances below:

Subscribe now

The Split Ticket midterms model is updated daily at 5 p.m. Eastern with the latest polls on our website.

We’ve seen a lot of cases where swing-state Republicans nominate ardently ideological, base-aligned candidates over more “establishment” options. Abdul El-Sayed’s primary victory may be the first instance in modern memory where Democrats have done the same in a critical statewide race.

Primary voters are not reflective of the general electorate, and many of those Republicans lost winnable races as a result. But there’s reason to believe El-Sayed’s fate could be different.

There are real risks associated with his candidacy: Our Split Ticket midterms model sees him as a slightly weaker candidate than his primary opponent Rep. Haley Stevens would have been. Stevens had an 80% chance of winning the general election, according to the model, while El-Sayed’s odds are currently at 70%. While we estimate that Stevens would have won by 5.4 percentage points, we think El-Sayed will win by just 3.6 points, thanks to his substantially weaker polling.

Share

Why is El-Sayed still forecast to win, despite his relative weakness? The single biggest reason is a national and state environment that’s uniquely unfriendly to Republicans.

In 2024, when Donald Trump won the popular vote by 1.5 percentage points nationally, Michigan was an R+1 state. Today, that number is closer to D+9 by our estimation (or D+8, if you look at Nate Silver’s modeling). In that sort of environment, even if El-Sayed’s candidate quality is similar to that of the average hardline progressive from “The Squad” (meaning he would perform 5 points worse than a replacement-level Democrat), he would still win.

Are national polls simply wrong about how blue 2026 might be? I don’t think so. Every other indication we’re seeing, ranging from special elections to the Washington state jungle primary, suggests that this really is a blue tsunami of a cycle.1 And in a situation like that, El-Sayed doesn’t need to overperform or even match the baseline performance of generic Democrats. He simply needs to avoid being a disaster.

Can he do that? It’s early, but I’d say the initial signs are promising. He has distanced himself from controversial remarks about 9/11 by Hasan Piker, who campaigned with him. He appears to be embracing party unity and has disavowed his own prior left-wing comments on defunding the police.

In doing so, El-Sayed appears to be avoiding many of the same missteps that plagued Republican firebrands like 2022 Arizona gubernatorial candidate Kari Lake. Also working in El-Sayed’s favor is the fact that Republican nominee Rep. Mike Rogers has made absolutely no effort to distance himself from Trump, flirting with election denial despite running in a state where the president is supremely unpopular.2

That’s why, if you forced me to pick today, I’d say that El-Sayed still wins this race.

Our model is more bullish on El-Sayed than other quantitative forecasts,3 in part because we don’t pay as much attention to pre-primary polling,4 and in part because our generic ballot modeling of Michigan suggests El-Sayed would have to underperform at the level of Elizabeth Warren (who underperformed by 11 points in her last Senate race) in order to lose.

This means he’d have to be nearly twice as weak as the average member of “The Squad,” for whom the typical underperformance is 5 points. While there’s plenty of evidence to suggest El-Sayed is weaker than Stevens would have been, there’s simply no evidence to support the hypothesis he’s that much weaker than a replacement-level Democrat.

This has implications for 2028 as well. Because while election nerds like me will look at overperformance and evaluate whether a win was by “more” or “less” than expected, that’s just not how the median voter looks at things. For a party whose left flank is ascendant, any win from El-Sayed is likely to be held up as proof that their candidates can actually win swing states — and probably gives fuel to the left in the 2028 presidential primaries.

For a long time, candidates like El-Sayed were simply boxed out of Democratic primaries — for anything other than safely blue races — due to concerns about electability as well as struggles with nonwhite segments of the party base. It was this phenomenon that tanked both Warren and Bernie Sanders in the 2020 presidential primary. And if the primary between Stevens and El-Sayed had been held in 2020, it is virtually certain that Stevens — who lost by just 1 percentage point this time — would have won.

But amid the historic unpopularity of both Trump and the Democratic establishment — and in the context of the never-ending conflict in Gaza, which dominated this primary — it is clear that 2026 is a different landscape altogether. Rank-and-file Democrats may well see El-Sayed’s nomination as a relevant test of the left’s ability to win swing states.

Whether that’s true or not, it could have seismic implications for the direction Democrats choose to go in 2028.

Subscribe now

Recommended reading:

“You can imagine this as being equivalent to Anthropic [today], except it has 100 or 1,000 times more researchers than the company actually has,” Epoch AI senior researcher JS Denain told me, working through the hypothetical scenario in a Substack Live with The Argument Wednesday. “And also, all those researchers are working at 10 times, 100 times the speed.”

That scenario, which AI researchers call recursive self-improvement, would not necessarily mean that humans lose control of AI. That said, most of the scary, sci-fi-adjacent scenarios do at least start from this premise.

So, just how close are we to self-improving AI?

One way to try to answer the question is to look at the recent overall arc of AI progress: How long and complex are the tasks that AI completes? How many resources are dedicated to this project? In both of those cases, progress has grown exponentially, lending weight to bold predictions about what could happen if such trends continue.

Another thing you could do is look at what’s happening inside the labs: OpenAI recently had its latest model optimize a smaller model. Anthropic reported that it tested Claude’s decision-making at certain critical junctures in the AI research process. Successive models made better and better decisions until Claude Mythos Preview beat human researchers 64% of the time.

Successfully outsourcing AI research decisions to Mythos would be, to some, an obvious sign that agents could lead research soon. With this evidence in mind, aggressive predictions that we’ll have automated AI researchers in just two years do not sound quite so far-fetched.

But making key decisions isn’t the same as running a whole scientific research process. That involves novel thinking, critical judgment, and resource management.

To figure out whether AI is up to these tasks, Denain has proposed treating AI research the way economists treat any other automatable job in the economy: Itemize all the tasks involved in the job, then figure out how well AI can do each of them specifically.

In other words, what does it actually look like to be a scientist who is building AI? As I put it in our conversation: “What is the equivalent of an AI researcher dropping a drop of solution like a chemist does to run their own experiment?”

Denain and I tried to separate things AI currently does well in the research process from the things it can’t yet do well.

It’s common to say that AI succeeds at defined tasks but struggles at more open-ended ones. This discussion shows there are still layers of complexity below that. You have to ask what particular aspects of open-ended tasks AI struggles with: Can it set new research directions? Can it incorporate critiques? Can it manage a budget?

Recent research suggests that, for all three, the answer, so far, is “no.” This complicates the picture of steady progress coming out of the AI labs. Watch the interview to learn more about why.

Subscribe now

Is it so hard for elections to be run cheaply, accurately, and quickly?

Well, according to Nevada Secretary of State Cisco Aguilar, it’s actually impossible.

“You can’t have all three. You have to pick two,” he said in a conversation with The Argument’s Lakshya Jain and VoteHub’s Zachary Donnini. “Accuracy is always going to be the priority. So you’re going to choose between fast and low-cost, right? If you want fast, there’s going to be a cost to it.”

Share

For instance, Florida is one of the fastest states in the nation in counting votes (roughly 99% within hours) and it also spends the most ($15 per voter). California’s big problem is that over 80% of its ballots come in by mail, which requires a human to verify the ballot signature against the voter’s registration record.

Nevada has demonstrated significant innovation in vote-counting to help speed things up. Lakshya and Zachary were both impressed by the dashboard Nevada recently released, which shows how many voters are turning out in every county in near real time, updating every six minutes.

“It’s become the norm. The fact that a person could know where their pizza is when they order it on an app to the time it arrives at their door, that is more information than they knew about their ballot at one time,” Aguilar said. “We need to make sure that we had that information available to voters.”

Subscribe now

Check out the full conversation above to hear about how Nevada has modernized its elections and how Trump’s election interference could unwind that progress.

If you’re deep in AI discourse, you have probably seen the below graph by METR (Model Evaluation and Threat Research) showing the growth in the complexity of tasks AIs are able to accomplish. According to the chart, Claude Mythos can quickly knock out tasks that would take a human 16 hours to complete!

But that figure comes with some major caveats, as Nathan Witkin, the AI Program Lead at NYU Stern’s Tech and Society Lab, explained in a Substack live conversation with Kobe Yank-Jacobs. For starters, the metric cited by METR measures the length of tasks an AI can complete 50% of the time.

“You’re just not even going to waste your time asking an AI model to do something if you can only expect like a 50% success rate,” Witkin said. “Adoption in realistic settings hinges, in some ways, more on reliability than capability.”

That means, as things stand, any work given to AI will require human review to make sure it’s up to organizational standards. That changes people’s workflows, but its effects on overall employment are less clear.

Recent research has pointed out a decline in the employment of young people in AI-exposed industries to make the case that AI adoption is hurting their prospects. But Witkin argued those declines have competing explanations — after all, the trend started before the release of ChatGPT.

“These drops are probably not because of AI, in my opinion. And I’ll add the opinion of a lot of the economics profession,” he said. “I agree we need to be on the lookout for employment-related risks … but it’s not happening yet.”

Subscribe now

To learn more about AI’s impact on the job market, you can watch the full video above. Jerusalem Demsas and Matt Yglesias also discussed it in a recent podcast episode with Understanding AI’s Timothy Lee, which you can find below. Check out Matt Bruenig’s take on the subject while you’re at it.

I was recently delighted to discover that my fellow Americans do not consider me an elite.

In The Argument’s recent poll of registered voters, fielded nationally between July 20 and 26, 2026, we asked respondents who they considered elite.

Of all the groups we listed, journalists were tied as the least selected group at just 9%.

As someone entering middle age and therefore excessively concerned with what young people think of my kind, I eagerly checked the breakdown of the results by age and discovered that just 7% of the youths (18- to 29-year-olds) consider journalists to be elites.

Young people actually have a more favorable view of journalists than any other age cohort. With 57% of 18- to 29-year-olds viewing us positively, my colleagues are more popular than Democrats, Republicans, billionaires, and Alexandria Ocasio-Cortez. (We do, somewhat vexingly, remain less popular than college professors.)

Share

It’s hard to see young people’s attitudes as especially informed since, according to them, they simply do not read or watch the news very much — at least not without the mediation of social media.

Regular consumption of traditional news sources — by which I mean newspapers, news websites, and television — declines dramatically as you climb down the age ladder. Those under the age of 30 are less than half as likely to get regular news from TV as the general population and 8 percentage points less likely to read newspapers and news websites regularly.

Of all the traditional formats, newspapers and websites are the most used by 18- to 29-year-olds — but with just 33% of this group perusing these sources regularly, they are still distinctly unpopular parts of their news diet, on par with Elon Musk’s X. Compare that with TikTok, podcasts/YouTube, and Instagram, where 42% to 53% of young people say they get their news.

Subscribe now

Even Facebook is slightly more popular than traditional news sources, although at 36%, the youngest subgroup’s rate of regular Facebook usage is about the same as that for our age 65 and over respondents, with whom Facebook is the most popular social media news source by far.

Young people probably don’t see themselves as opting out of news consumption: Only 4% selected “none of the above” when we asked for their news sources, in line with the general population.

And, just like older voters, the 18-to-29 cohort has opinions on bias and partisanship in journalism — albeit distinct ones. While nearly half of our respondents overall think that journalists have a liberal bias, only 40% of those under 30 feel the same. This shift is primarily accounted for by a larger proportion of 18- to 29-year-olds saying that journalists are “neither more liberal nor more conservative” than any other age group.

There are a couple reasons this may be the case.

1. Our 18- to 29-year-old respondents were 15 percentage points more liberal than our sample overall, so they may be less likely to see liberal bias than a more conservative overall population.

2. Younger respondents, who consume primarily algorithmic media, may be working off an implicit redefinition of what a journalist actually is. After all, is a gonzo YouTuber named Reckless Ben who’s obsessed with catching an alleged Lego thief a liberal or a conservative? I can’t even venture a guess.

But as to whether or not the traditional national media are “fair and unbiased,” all age groups are split roughly 50/50. That said, there is a dramatic drop-off from the 18% of those 65 or over who believe the news can be trusted “most or all of the time” to the merely 8% of those under 45 who believe the same.

This is directionally consistent with 2025 results from Gallup, which found a 15-point gap in trust in mass media between 18- to 29-year-olds and those 65 and over — a gap that has widened dramatically since the 2001 to 2003 period when the two were at parity.

In some ways, this all confirms what we already know: We are entering a world where information delivery is increasingly fragmented and traditional gatekeeping institutions have diminished trust and influence. This has created formidable challenges for the news business and will continue to do so.

But, to the extent our survey results offer any solace, it seems like America’s young adults are approaching the field with a slightly more open mind than older Americans. And they are not particularly buying the “liberal media” narrative that the right has been pushing to discredit the profession since before they were born.

Subscribe now

Recommended reading:

Instead, we could be seeing a reversion to the mean of young people feeling simply alienated from politics altogether, as Kristen Soltis Anderson, co-founder of Echelon Insights, explained in a Substack live conversation with The Argument’s Lakshya Jain and VoteHub’s Zachary Donnini.

Young Trump voters “had kind of hoped that Donald Trump would come in and would fight for them. And it is not as though they now think that Democrats are going to fight for them. It’s just that they now think no one is fighting for them,” she said.

Share

Of course, it’s not like Gen Z has it uniquely bad; by many metrics, it is the richest generation so far. But Anderson thinks that Gen Z’s propensity to get news on social media pushes it toward pessimism.

“If that’s the largest place where Generation Z is getting its political news, those algorithms just select for the negative so much more,” she said. “So, you’re not only feeling like, ‘Wow, I can’t afford rent; wow, groceries are expensive,’ but you also have this surround sound environment that’s constantly reinforcing to you: ‘Yes, everything sucks. You’re not losing your mind.’”

Check out the full video above to learn more about what is — and is not — driving young voters this election cycle. The crew also gets into the pros and cons of using focus groups instead of polls and why Alexandria Ocasio-Cortez is “the best candidate Democrats have with young voters.”

Subscribe now

This is becoming a pattern, Lakshya Jain, The Argument director of political data, noted in a Substack Live conversation with VoteHub’s Zachary Donnini.

“The Democratic electorate now is what pundits thought the Democratic electorate was turned into in 2018, right?” Lakshya said. “It took eight years, but we are at the point now where moderates are actually the junior partner in the coalition.”

Share

DeGette’s loss to Democratic Socialist Melat Kiros in particular drew lots of national attention over the role Israel played in the race. But while Israel has become a wedge issue in several Democratic primaries throughout the nation, it’s not the only thing driving this leftward dominance.

Lakshya wrote about the Democrats’ potential for a Tea Party moment early last year, arguing that the base is upset over the establishment’s inability to fend of Trump and Trumpism 2.0. But the trend of these primaries shows the issue may also be more ideological than that:

“I said that Democrats in Congress are blind to how angry their voters are with them. They don’t understand how their voters blame them for Trump coming back. And they also just don’t understand that voters are going to absolutely massacre the establishment without serious corrective action. And so I said it was going to be on an axis of fight versus flight,” Lakshya said. “But I think I was wrong. I don’t think it’s that; I think it’s liberal versus establishment.”

Check out the video above to hear details about the Colorado primary and analysis about Democrats’ internal realignment.

Subscribe now

No, they won’t, say The Argument’s Lakshya Jain and VoteHub’s Zachary Donnini in today’s live election coverage.

What’s happening in California is a mirage: Democratic constituencies’ ballots are the latest to arrive, and thus the latest to be counted.

“The late ballots are so blue that Democrats are almost certainly not getting locked out,” said Lakshya. “A lot of Democrats in California just straight up did not return their ballot until the last possible minute.”

While Democrats can take some comfort in this, disproportionate delays in vote counting lead to a large loss in voter trust.

If initial results show “we’re getting two Republicans here, then all of a sudden, one week later, you see the Democrats surge to first place, the first thing a lot of people are going to think is that it’s not legitimate,” Lakshya explained. “Obviously, that’s ridiculous, it is legitimate, but at the same time, I have sympathy.”

“People have a right to be informed on elections,” he added.

Watch the full video above to learn more about how specific races are shaking out — and how California’s banana republic voting system hurts democracy.

Subscribe now

Incumbent John Cornyn would likely have won by five or six points, and the seat still has a good chance of staying red. But with Texas Attorney General Ken Paxton taking the nomination, Republicans will have to spend big just to maintain a lead of less than one percentage point, according to models by The Argument’s Lakshya Jain, Split Ticket’s Armin Thomas, and VoteHub’s Zachary Donnini.

“Not only is Paxton not an incumbent, not only does he not have Cornyn’s reasonably strong electoral track record, but he also has some pretty serious fundraising issues. Texas is a big state. It’s an expensive state to run in. He basically has no cash on hand,” said Donnini in a Substack live video Wednesday. “Cornyn is a fundraising machine. Paxton is going to get outraised and outspent by [Texas State Rep. James] Talarico by probably a lot.”

Talarico, the Democratic nominee for Senate running against Paxton, comes with baggage. But even if you put aside the risk of a Democratic upset in Texas, Republicans are gearing up for an incredibly expensive race that will draw resources away from other seats the party is defending.

That means candidates like Alaska’s Mary Peltola and Ohio’s Sherrod Brown, which Lakshya’s survey data shows being slight favorites, will get an extra advantage.

“You’re going to have to put in a lot of money into Ohio, for example, to make sure that Sherrod Brown doesn’t win, because at the moment, if you held the election today, he probably would. So now the question becomes ‘What is going to be required for Paxton to maintain parity between now and November?’ And where is that money going to come from?” he said.

Check out the video above to hear more about how the Paxton pick changes Republicans’ 2026 game plan

Subscribe now

Midterm polls range anywhere from Democrats leading by a whopping 15 points to a nail-biting three. As our Director of Political Data Lakshya Jain likes to remind us: Keep calm and trust the polling average.

The current polling average is slightly below where it was in 2018, but the situation is very different: In 2018, the polling average stayed relatively flat throughout the cycle, whereas in 2026, Democrats have increased their advantage month by month.

Furthermore, this time, Democrats have more of an advantage on issues like cost-of-living and immigration.

“[The] Republican Party has lived for a long time on the bread and butter of economics, right? Like, whatever happened, they were advantaged on three things: crime, economics, immigration. Right now, Trump has done a lot to make immigration not a great issue for Republicans,” said Lakshya in a Substack live conversation with Split Ticket’s Armin Thomas. “And then, on cost of living, it’s outright going negative for the GOP.”

Share

That being said, Democrats could still screw this up. It would be quite unusual if Democrats won senate races in places like Alaska and North Carolina while losing in Michigan, but that is a real possibility if Democrats nominate a historically weak candidate.

“Right now, the polling seems to suggest that Abdul El-Sayed is the weakest of the three candidates for perhaps his ideology being further to the left than the others …this is not a case of Republicans nominating someone like Doug Mastriano or Kerry Lake,” said Armin. “This is actually a rare instance where Republicans have nominated, on paper, what’s a very good candidate.”

Check out the full video above to hear more about what pitfalls await both Democrats and Republicans in the upcoming elections.

Subscribe now

David Modica, who went viral last week for speaking out against the town’s new plan to evade the state’s rules, joined Jerusalem Demsas on a Substack live video to explain why he was motivated to speak up.

“We are one of those places that you’d expect to sort of be open and welcome and inclusive — at least that’s what we say our values are. And I think the fact that we were so clearly not doing that and flaunting the rules in such a cynical way really bothered me,” he said.

The town’s new plan is to upzone a golf course: somewhere that has no housing and no immediate plans to add any housing. And while Marblehead’s process for arriving at this plan was unique, similar shenanigans pop up all over the country.

“Once you give localities that kind of discretion, what you end up getting is attempts to fake-meet the target, right?” said Jerusalem. “If you upzone a golf course, the golf course isn’t likely going anywhere. That golf course is going to continue to be a golf course. Yes, it is now legal for a developer to buy on that golf course and turn that into multifamily housing but that’s just not going to happen. And this happens all the time.”

Check out the video above to hear their full conversation, which touched on the demographic makeup of Marblehead and the legitimacy of NIMBY opposition.

Subscribe now

Lakshya Jain, The Argument’s director of political data, joined Split Ticket’s Armin Thomas and VoteHub’s Zachary Donnini in a live video chat to discuss how the decision will affect future elections.

Lakshya thinks Democrats should be worried about the long-run implications:

“It’s still very clearly Republican-leaning,” he argued. “I think like a lot of what I’m seeing from Democrats online is just cope about this. They’re like, ‘Oh, it’s not that bad.’ I’m like, ‘No, it’s it’s pretty bad. It’s going to result in the Democrats losing about six to seven House seats. It is just not as bad as losing 10 to 12.”

Armin took a more measured tack:

“I think the decision is not as maximalist as it could have been, right? Everyone was talking about how, overnight, all of these southern legislatures were immediately going to remove all of the Black Democratic minority seats out, but it makes it really, really difficult to enforce.”

Check out the video above to watch their full discussion. Then read Lakshya and Split Ticket’s analysis of how much disruption Democrats can reasonably expect.

Subscribe now

“I thought we’d have a little bit more entertainment before we were like, ‘No, it’s going to pass,’” said Lakshya. “Got to say that this is probably one of the more underwhelming drops in terms of election nerd stats that I could have imagined.”

But while the election itself may not have exactly been electric, the aftermath certainly will be. Assuming Virginia’s new map survives a likely challenge from Republicans, Democrats will have effectively neutralized the mid-district gerrymandering advantage that Republicans were hoping to gain from places like Texas and Ohio, and Democrats may even come out a couple seats ahead.

On the other hand, the Supreme Court is overdue to issue a ruling in a case that is likely to overturn Section 2 of the Voting Rights Act. This would enable heavy Republican areas like Florida to issue much more aggressive gerrymanders that are currently illegal.

“Once it comes around to the 2028 cycle and Section 2 of the VRA gets struck down, all bets are off. And I think what happens there is you’re going to see that places like Virginia are going to try to push to keep these maps more permanent. I don’t think they’re going to succeed in that one because I think a Democrat wins in 2028, and then the 2030 referendums are going to suck for them,” said Lakshya. “After all is said and done, Republicans will probably gain a few seats, [but] not enough to make the House permanently red — I don’t like that New York Times headline.

Check out the video above to hear more about what the new maps could mean going into the midterms and beyond.

Subscribe now

Well, in all likelihood, you get a billionaire governor who just spent over $100 million on TV ads. Lakshya Jain, The Argument director of political data, discussed this probable outcome in a conversation with VoteHub’s Zachary Donnini.

“A lot of people still watch TV. In our polling, over 50% of voters still get their news mostly from TV. Especially older voters are more likely to do that. And that’s who votes in a primary, right? Primaries are a lot older. They’re a lot more engaged,” Lakshya said.

Share

While San Jose Mayor Matt Mahan is getting renewed attention as the billionaire-skeptical state casts about for an alternative, the odds are not in his favor. With the nonpartisan top-two primary election just a month and a half away, his best chance would be to come in second to presumptive frontrunner Tom Steyer and then continue campaigning in the runoff.

But as Donnini explained, that’s not particularly likely now that Donald Trump has endorsed Republican candidate Steve Hilton, shoring up Republican support around a single candidate.

“I think it’s going to be potentially Hilton running in first at this point,” he said.

Watch the full video above for more analysis of the race as it stands, or you can check out our write-up below:

Subscribe now

Antisemitism, however you measure or define it, is on the rise.

The conventional wisdom in political science is that antisemitism behaves as a “horseshoe,” increasing in prevalence as you move toward both the far left and the far right. After Oct. 7, many narratives have homed in on left-wing antisemitism, both real and perceived, particularly among young people in the context of campus protests about Israel.

But is this conventional wisdom correct?

To try to answer this question, the Yale Youth Poll1 has conducted two large national surveys of registered voters over the last year (with oversamples of young voters). In both surveys, we included questions designed to gauge the extent to which voters hold antisemitic beliefs, as well as the extent to which these voters are similar or distinct from those who hold anti-Israel beliefs. We also drew on polling by The Argument, in combination with our own data, to look at whether social media use might be driving either set of views.

Our data tells a story that doesn’t quite align with the conventional wisdom. Our findings are threefold:

1. Antisemitic views are more prevalent among younger people.

2. The young people who are most likely to agree with antisemitic statements are on the political right.

3. Anti-Israel views are more common on the political left.

Which Americans are most likely to hold antisemitic views?

To measure antisemitism, we asked respondents whether they agreed or disagreed with the following three statements, which are commonly considered to be antisemitic. These three statements were drawn from a similar battery used in a series of papers published between 2022 and 2023 by political scientists Eitan Hersh and Laura Royden.2

1. “Jews in the United States are more loyal to Israel than to America.”

2. “It’s appropriate to boycott Jewish American-owned businesses to protest the war in Gaza.”

3. “Jews in the United States have too much power.”

By aggregating data from the fall and spring surveys, we obtained a sample size of N=6,855 registered voters, including N=4,021 voters under 35 (it is, after all, a youth poll).3

Among all voters, just under one-third agreed with at least one antisemitic statement, but specific subgroups of the population displayed wide variation.

Unsurprisingly, Jewish voters were much less likely to agree with antisemitic statements than non-Jewish voters.4 Women were also notably less likely to hold antisemitic beliefs than men.

Nonwhite Americans were more likely to agree with antisemitic statements on average than white Americans. Among white respondents, 29% agreed with at least one antisemitic statement, compared to 37% of Hispanic respondents, 41% of Asian respondents, and 43% of Black respondents.

In the overall population, there was very little evidence of an education gradient or political trend with regard to antisemitic beliefs. The shares of Harris and Trump voters who agreed with at least one statement were nearly identical, as were the shares of respondents with a graduate degree and of respondents with less than a high school diploma.

What’s most notable is the sharp age gradient. In our data, 40% of respondents aged 18 to 22 agreed with at least one antisemitic statement, compared to just 26% of respondents over 65.

Share

As you can see in the chart, the relationship between age and the prevalence of antisemitic beliefs looks like a smooth curve — nearly the same shape that you see in polling of explicit antisemitic attitudes5 from Blue Rose Research.

The locus of youth antisemitism is on the political right

Thanks to our large sample of just over 4,000 voters under 35, we can take a detailed look at which types of young people were the most likely to hold antisemitic beliefs.

First, as with the overall population, men were more likely to be antisemitic than women, though the gender gap is slightly larger among young people.

Second, among younger voters, we do observe an education gradient (which we don’t see among voters overall). Young people with a high school diploma or less were about 10 percentage points more likely to agree with at least one antisemitic statement than those with a graduate or professional degree.

Third, we still observe racial gaps among younger voters, though they differ slightly from the overall population. Among voters under 35, Hispanic voters were most likely to agree with at least one antisemitic statement (49%), followed by Black voters (43%), Asian voters (40%), and white voters (36%). Younger white voters were seven points more likely to agree with at least one antisemitic statement than whites overall, while younger Hispanics were 12 points more likely than Hispanics overall. There was essentially no age gap between younger Black and Asian respondents and the overall Black and Asian sample.

Finally, there is a massive partisan and ideological gap among young people. That is not the case among the general population.

Half of Trump 2024 voters under the age of 35 agreed with at least one antisemitic statement, compared to just 32% of Harris voters.

When looking at self-reported ideology, the gap is even more stark. Almost half of young people who call themselves “slightly conservative” or “conservative” agreed with at least one antisemitic statement. But, most strikingly, nearly two-thirds of “extremely conservative” young people agreed with at least one antisemitic statement. It bears emphasizing: No other demographic group in our entire survey expressed levels of antisemitism nearly as high as this.

On the other hand, young people who self-identify as liberals were among the least likely groups in their cohort to agree with the three antisemitic statements we tested: 32% of young people who call themselves “liberal” agreed with one or more statements (with almost the same figure, 33%, for self-identified “extremely liberal” young voters), compared to 41% of young people overall.

In short, young people who voted for Trump, have lower educational attainment, or call themselves conservative were among the most likely to agree with at least one antisemitic belief. And young voters who identify as “extremely conservative” hold the most antisemitic views by far.

The relationship between antisemitism and anti-Israel views is complicated

In our second survey, fielded from March 3 to March 23 this spring, the Yale Youth Poll asked respondents whether they agreed or disagreed with a variety of statements about Israel, ranging from extremely supportive to extremely critical.

• “America should end the slavish surrender to Israel, its wars, and its demands for foreign aid.”

• “The Bible says that those who bless Israel will be blessed, and those who curse Israel will be cursed, so America should stand with Israel.”

• “America needs Israel for national security. They are the tip of the spear when it comes to defeating terrorism.”

• “Israel is a sanctuary for historically oppressed people, minority rights, and democracy in a region where no others exist.”

• “Israel is a democracy and the safe haven of the Jewish people, but they have a moral obligation to ensure humanitarian treatment of Palestinian civilians.”

• “Israel is an apartheid state, engaged in racist oppression against Palestinians.”

The statements were, in order, paraphrased from Nick Fuentes, Ted Cruz, Nikki Haley, Ritchie Torres, Chuck Schumer, and Rashida Tlaib.

What we see in our data is that young people, more educated people, and those on the political left were more likely to agree with the harshest contemporary left-wing critique of Israel that we tested — that “Israel is an apartheid state, engaged in racist oppression against Palestinians.”

However, people who scored highly on our measure of antisemitism were also more likely to agree that Israel is an apartheid state.

But while antisemites were indeed more likely to have extremely negative views about the state of Israel, many of the groups that were most critical of Israel, at least according to this measure — people with graduate degrees and those who identify as “extremely liberal” — were the groups least likely to express more explicit antisemitic beliefs (i.e., those unrelated to Israel).6

This matches a pattern you see in polling from the United Kingdom, where supporters of left-wing parties have an extremely negative perception of Israel’s treatment of Palestinians but were also generally less likely to express non-Israel-related antisemitic views than supporters of right-wing parties.

There is, of course, significant disagreement about the point at which views about Israel veer from criticism of the Israeli government’s actions toward antisemitism.

Take, for example, opposition to Israel’s existence as a Jewish state.

Under mainstream definitions of antisemitism (e.g., those used by the International Holocaust Remembrance Alliance or the Anti-Defamation League), this view is inherently antisemitic because it means denying the Jewish people the right to their own state when other ethnic groups have this right. And, indeed, in our fall 2025 survey, we found that voters who self-identified as extremely liberal (25%), liberal (37%), and slightly liberal (37%) were much less likely to say that “Israel should exist as a Jewish state” than those who described themselves as slightly conservative (49%), conservative (66%), or extremely conservative (72%).

Still, other definitions of antisemitism (e.g., the Jerusalem Declaration on Antisemitism and the Nexus Document) would consider this statement acceptable, arguing that it targets a specific political ideology, rather than explicitly targeting the Jewish community itself.

But wherever the boundary between political and antisemitic views on Israel lies, based on our data, it is more common for voters on the political left to agree with statements that plausibly cross the line from anti-Israel to antisemitic than those on the political right. This is true even though it is simultaneously the case that, in our surveys, young conservatives were more likely to agree with explicitly antisemitic statements than young liberals.

Is this all downstream of social media?

What’s driving this increase in antisemitic and anti-Israel views among young people? One possible explanation is social media.

If you’ve spent time on Instagram Reels in the last year, there’s a chance you’ll have seen something about “the Big Yahu,” “goyslop,” or an Agartha edit, as well as plenty of content about Israel’s conduct in Gaza and the war in Iran.

To investigate the relationship between antisemitism and social media, we included a question on our second survey asking respondents which platforms they get their news from.

We find that antisemitic beliefs were more prevalent among voters who get their news from social media. For example, 41% of respondents who get their news from TikTok agreed with at least one antisemitic statement, compared to 33% of all respondents.

There’s less of a clear-cut relationship between preferred news sources and perceptions of Israel as an apartheid state. People who get their news from Reddit and (to a lesser extent) from TikTok, Instagram, or podcasts were more likely than average to agree that “Israel is an apartheid state, engaged in racist oppression against Palestinians,” but so were those who get their news from newspapers and broadcast television.

And it’s not clear if either of these relationships is causal. If people who get their news from social media skew younger, then we might just be seeing an age effect filtered through news consumption patterns. Or perhaps social media is the underlying source of increased agreement with antisemitic statements, which partially explains the age gap described above. And, of course, those with a preexisting belief that Israel is (or is not) an apartheid state may be motivated to seek out content that confirms their priors.

Because we used the same news source question that The Argument uses in their polling, we were able to check the age distributions of each news platform in The Argument’s aggregated polling data. Indeed, we can see that the news-consuming audiences of social media platforms skew younger (except for Facebook).

Viral antisemitic and anti-Israel content on social media may be shaping young people’s views. But that content is also created because there’s a demand for it. It’s not immediately obvious how much of one is causing how much of the other.

And that’s true of much of our data. While there are some clear takeaways — both antisemitic views and negative perceptions of Israel are more common among younger voters — the situation is still really complicated. Some young voters may be driven toward anti-Israel beliefs by their preexisting antisemitic beliefs, and some young voters may adopt antisemitic views as a result of preexisting anti-Israel views or of conflating Israel with the Jewish diaspora. Some young voters are deeply critical of Israel but are not antisemitic, and some young voters may be antisemitic but are not so critical of Israel.

Our poll cannot clearly distinguish between these four types. The rise in antisemitism and anti-Israel views among younger generations may come from a variety of different factors, which warrant further research.

Subscribe now

Recommended reading: