Agatha and her coven! Will she backstab? Probably! | Chuck Zlotnick/Courtesy of Marvel Do you think Patti LuPone would be able to explain who Tony Stark is? Does she know anything about the multiverse? Could she cobble together three facts about Kang the Conqueror? Or the difference between Captain Marvel and Ms. Marvel? Is she familiar with Wanda Maximoff and how she became Scarlet Witch? Did she watch WandaVision?   Like the probability of Patti LuPone showing up to a Madonna concert, the answer to all the above is no, never, or absolutely not.  LuPone will tell you herself that she has no idea how anything in the Marvel universe works. That’s what makes her perfect for Agatha All Along, where she plays the kooky fortune teller Lilia Calderu. It might seem counterintuitive, but the fact that even one of the show’s stars lacks baseline Marvel knowledge goes to show how easy it is to get into the new WandaVision spinoff. It’s a perfectly silly spooky season teaser, even for non-Marvel fans.  Viewers require very little background knowledge to understand the show. That’s given the series a serendipitous story-telling cushion, allowing everyone involved a bit more creative slack to show us how the most maniacal 400-year-old witch to ever haunt the planet got her groove back. It’s a journey that takes us to a spoof of Mare of Easttown, an uncanny play on Bravo’s Real Housewives, a winking riff on The Wizard of Oz, and so many weird places in between.  Agatha is more interested in creating witchy nonsense than sticking with superhero seriousness. Each episode is crammed with spells and potions and magic, rather than reminding you what the Avengers are up to. There’s comedic, spooky world-building happening here in a way that Marvel’s regular fare rarely allows.  All you need to know for Agatha is that Agatha Harkness (Kathyrn Hahn) is a magic user who messed with the very powerful Wanda Maximoff (Elizabeth Olsen), a.k.a. the Scarlet Witch, and lost. Wanda took away all her power, zapped all her memories, and sentenced her to live in a New Jersey town called Westview. Stuck in the suburbs with no magic and no identity, Agatha’s career is in shambles.  Her rock bottom is where Agatha All Along picks up.  Thanks to the help of Teen (Joe Locke), a character that goes by the name “Teen” because of a spell that doesn’t allow him to say his actual name, and triggered by Wanda Maximoff’s death (see: Doctor Strange in the Multiverse of Madness), Agatha breaks out of her trance. Like anyone living a life of quiet desperation in suburban New Jersey, she realizes she needs to regain her powers, and to do so, Agatha will need to assemble a gang of witches — a divination expert (LuPone), a potion brewer (Sasheer Zamata), a protection mage (Ali Ahn), and an earth magic enchantress (Debra Jo Rupp) — to find the “Witches Road” and pass its various trials.  A group of misfits walking down a road and having to help each other along the way? It’s an inverted riff on another very famous story about a road, a wicked witch, a group of new companions, and a troublesome redhead with all the pieces mixed and matched. Unfortunately for her coven, Agatha is very into double-crossing. Fortunately for her coven, everyone knows that Agatha is a double-crosser. Getting to the end of the road will require teamwork and keeping Agatha in check — things neither Agatha nor her coven seem capable of doing.  When Agatha All Along was greenlit, Marvel mania was still riding high on the back of 2021’s WandaVision, the potential for a multiverse of parallel timelines and worlds, and the promise of a new era of Avengers. Controversy surrounding Jonathan Majors’s domestic abuse allegations and a series of middling movies has fully derailed that momentum, but the checks were already signed and filming wrapped for Agatha in May 2023.  Agatha doesn’t have to set the table for the Marvel Cinematic Universe and the series doesn’t feature a character that prominently figures into Marvel’s bigger plans. This could all change, but there’s currently no expectation for the Tony Award-winning LuPone to be busting out of an Iron Man suit to take on Doctor Doom (as fantastic as that sounds). Nor does there seem to be a big plan for Hahn’s Agatha to link up with any of the existing Avengers.  Without the pressure of being an integral piece of the Marvel design or needing the characters to save the world (as is the MO of so many Marvel projects), showrunner Jac Schaeffer and Agatha’s creative team have been allowed a bit more freedom to create a show on their terms. That materializes in the show’s penchant for silly musical moments and multiple costume changes. An episode later this season (critics were given four episodes to review) finds the women trapped in a Nancy Meyers-esque coastal mansion, complete with poison wine and a treacherous luxury bathroom.  The only predictable thing about the series is that not everyone in the coven is going to make it to the end, and no one, not even Agatha, is guaranteed a happy ending. Perhaps the witches will all turn on Agatha. Perhaps Agatha will, again, betray the witches who trust her. For once, anything and everything could happen in a Marvel show, and that’s an enchanting possibility.

New York Gov. Kathy Hochul waves during the first day of the Democratic National Convention at the United Center on August 19, 2024, in Chicago, Illinois. | Kevin Dietsch/Getty Images There’s a hot new trend this back-to-school season: cellphone bans. At least eight states have enacted regulations limiting cellphone use in schools so far this year, and many more individual districts and schools have implemented similar policies on their own. The changes are driven by bipartisan concern that teenagers are unable to break away from their phones and concentrate in class, or even just talk to people in real life, as well as growing concern about the pervasive mental health challenges posed by social media. “At first I thought it was going to be really annoying, but it’s actually not that bad,” said Lev Zitcer, a freshman at Science Leadership Academy in Philadelphia, which is limiting phone use this year. “I think there’s like a different level of communication that comes with, like, being bored.” New York Gov. Kathy Hochul agrees. She’s been campaigning to enact a cellphone ban across the Empire State. “I’ve talked to schools where they have banned cellphones. We’ve found out that there are a lot of challenges involved. But if you get ahead of it, we can be successful,” Hochul told Today, Explained co-host Sean Rameswaram. Below is an excerpt of Sean’s conversation with Hochul, edited for length and clarity. There’s much more in the full podcast, so listen to Today, Explained on Apple Podcasts, Spotify, or wherever you get podcasts. Sean Rameswaram You’ve got a lot on your plate. You’ve got that — we saw you at the DNC. Obviously, a lot of election stuff in the air right now. There’s been a lot of controversy around congestion pricing. But we’re here to focus on getting cellphones out of the classroom. Why is this an important issue for you right now? Kathy Hochul This is the end of a long journey that I started about a year-and-a-half, two years ago, when I was seeing data about teenagers really struggling after the pandemic. And I wanted to know what’s going on with the kids. And I have a lot of teenage nieces and nephews and I’m seeing things happening that are not positive. I started talking to teenagers, going around to schools, every corner of the state, convening them and finding out that they are so affected by the bombardment of addictive algorithms on social media throughout the day. And also the extreme cases of FOMO. They need to know what’s happening — the girls meeting in the restroom without them? Is a party being planned and they’re not there? So these kids’ stress levels are off the charts. But meanwhile, this is during the school day when they’re supposed to be paying attention to a teacher, learning something and ultimately graduating. So it’s a huge distraction. I know the opposition, what they’re going to say — they’re going to tell me that they need to be able to reach their children if there’s a crisis. And I’ll tell you right now, that was my first reaction, like, well, okay, they may not need to have access to the internet, social media during the day, but they certainly need a cellphone to contact their parents if there’s a mass shooting — every parent’s nightmare. But what I heard from law enforcement disabused me of that notion … They said if there is a crisis on campus, no matter what it is, the last thing you want are your kids reaching for their cellphone, trying to communicate with them, take pictures, getting video of it. You want them to pay attention to the head of the classroom, their teacher, to lead them to safety. And I was persuaded the second I heard that from law enforcement. Sean Rameswaram Interesting. As you alluded to, a lot of schools are doing this. Schools across the country are doing this, states across the country are doing this. Blue states, red states, they’re taking different approaches. What approach do you want to take in New York? Kathy Hochul We’re winding down in our information-gathering process. First of all, I can’t do anything as governor without the legislature. So I’m basically building the case, building the data, the narratives, the information from the surgeon general, other experts, authors who study this, building the argument that, first of all, our young people are better off without access to a cellphone during the school day. And teachers, 72 percent of teachers in this country are saying it’s a big distraction. It is not helping the learning process. In fact, it’s hindering. So we are looking at and I’m going to be proposing this in the next few months before they meet again in January, more likely a cellphone ban … It is easier to lock them up at the beginning of the school day, get them out at the end of the school day instead of: “Who’s going to make sure they have them locked up again when they come back from recess? Who’s going to make sure they’re locked up after lunch? What if they go to the restroom?” … The teachers don’t want to be the phone police. They want to teach. Sean Rameswaram I want to ask you about how parents might feel about that, because parents seem to be a big part of this equation. And it seems generally, and we’ve heard it from even, you know, a parent on our team, parents want to be able to communicate with their kids throughout the school day. And I’m wondering, one, are you hearing that from parents in New York state, and two, how you’re going to sell this to them when the time comes? Kathy Hochul That’s a good question. And here’s what I would say: Talk to the other parents who came to our meetings, but especially talk to the parents who are teachers. We get a lot of them. And people who understand what has happened to their child in this setting, that they’re a different person than they would otherwise be because of this constant communication to others when they’re supposed to [not be] distracted, they’re supposed to be learning. One mom said, “My son is being bullied throughout the day. My husband now leaves work early, so he’s there to be there when he gets home from school to make sure he doesn’t take his own life.” Because the intense pressure on kids being bullied through their phones, through the social media platform, it’s intense. Now, this is maybe a rare case. I’m not saying it’s common, but people don’t realize the pressure they’re under about how the kids look, what they say. It’s a tough environment to be a teenager under any circumstance, but you exacerbate it when you have all these outside external factors that are hitting them at the same time. When I was growing up, yes, there were bullies. Yes there were mean kids. You walked down the other hall, you avoided them, right? You can’t avoid being bombarded with messaging throughout the day. And we need our kids to be liberated. Sean Rameswaram Where do you think we’ll end up as a country? … Do you think we’ll end up in a place where every school will be doing some version of this? Kathy Hochul My view is that if we never start out with an expectation that they’re allowed in schools, this will be the first generation we liberate from that, and then the subsequent ones will not have that same pressure … It’s all about listening to the kids. They want us to save them. And I’m the adult who’s going to be willing to do that.

Democratic presidential nominee Kamala Harris attended a town hall style forum hosted by Oprah Winfrey in Michigan Thursday night. Harris received questions on most of the 2024 campaign cycle’s top issues, including guns and immigration — but a segment on abortion proved to be an emotional centerpiece that has continued to generate conversation. That moment largely focused on a 28-year-old Black woman from Georgia named Amber Thurman, featured in a recent ProPublica report. Thurman died in August 2022 after doctors hesitated to treat her following a complication from a medication abortion. After that year’s Supreme Court ruling in Dobbs v. Jackson Women’s Health Organization, which overturned the right to an abortion guaranteed by Roe v. Wade, Georgia implemented a strict abortion law that severely limited the abortion-related care available to anyone more than six weeks pregnant.  Thurman was at least nine weeks pregnant; state law allows Georgia doctors to perform the procedure she required — a dilation and curettage, or D&C — only when the mother’s life is in danger. Doing so before then could result in a doctor going to prison for up to 10 years. At the time, the hospital reportedly had no guidance or policy in place about how to navigate the law and ascertain whether a pregnant person’s life was in danger. However, the ProPublica report suggests Thurman’s doctors waited too long — 20 hours after she went to her local hospital — before beginning to operate on her. A state review found Thurman’s death was “preventable,” and that’s a theme her family stressed as they spoke with Harris and Winfrey. “They just let her die because of some stupid abortion ban,” Thurman’s older sister said. “They treated her like she was just another number.” Harris offered her condolences to Thurman’s family, and used the moment to argue that Thurman’s story underscores the need for a change for greater abortion rights — as well as the sort of abortion policy Democrats are running on. Democrats are running on expanding abortion rights In the wake of Dobbs, more than a dozen states have passed strict abortion bans; nearly a dozen others, including Georgia, have laws that severely limit who is able to access an abortion.  Georgia’s law, the LIFE Act, was initially passed in 2019 and upheld last year. It outlaws abortion once embryonic cardiac activity is detectable, something that usually occurs around five or six weeks of gestation. It does allow abortions past that point for “medical emergencies,” but is vague about when doctors should declare an emergency, other than defining them as a “condition in which an abortion is necessary in order to prevent the death of the pregnant woman or the substantial and irreversible physical impairment of a major bodily function of the pregnant woman.” In many states, there have been efforts to overturn abortion laws — or keep them from being instituted. In Michigan, where the town hall was held, voters enshrined the right to abortion into the state’s constitution in 2023. In the wake of Dobbs, ballot initiatives to protect abortion access in Kansas, Kentucky, Michigan, Vermont, Montana, and California all passed. And this year, a new slate of states will decide whether to protect access. “There are 10 states with ballot initiatives for this November,” said David Cohen, a law professor at Drexel University who focuses on gender and abortion access. “Five of those states would change the current law in that state … going from [ending] a complete abortion ban [in] South Dakota and Missouri, to alleviating a six-week ban in Florida, a 15-week ban in Arizona, and a 12-week ban in Nebraska.” Democrats have tied themselves to these initiatives, hoping that they boost turnout. The party successfully campaigned on abortion in the 2022 midterms, and made abortion a factor in several special elections that were Democratic wins. They hope to make the issue a central part of this year’s election too. According to the Pew Research Center, abortion is a top five issue for Democratic voters, and a top 10 issue for voters overall. Harris has repeatedly attacked former President Donald Trump as being responsible for the end of Roe, arguing as she did Thursday, “The former president chose three members of the United States Supreme Court with the intention that they would overturn the protections of Roe v. Wade — and they did as he intended.” Harris, meanwhile, has said that, as president, she would approve federal legislation protecting the right to abortion. The current model for that legislation is the 2023 Women’s Health Protection Act, which would prevent state governments from imposing restrictions on abortion rights pre-viability. (Of course, Harris would probably need a Democratic majority in both the House and Senate — which currently seems unlikely — for federal abortion protections to pass.)  For his part, Trump has bragged about being the president who overturned Roe, and has argued that abortion policy should be left to the states. He has said he would not approve a federal abortion ban if given another term. He has also sought to distance himself from Project 2025, the conservative vision for the US that includes draconian restrictions on women’s health care, rights, and freedom.  But that’s not to say that a second Trump term couldn’t make even abortion more difficult to access, including through the method he used the first time around: court appointments. 

Shohei Ohtani #17 of the Los Angeles Dodgers looks on after hitting a two-run home run against the Miami Marlins during the third inning at loanDepot park on September 17, 2024, in Miami, Florida. | Sam Navarro/Getty Images During a Thursday Los Angeles Dodgers game versus the Miami Marlins, baseball phenom Shohei Ohtani hit a record that no other player has reached.  In that game, Ohtani became the first baseball player to reach the elusive “50/50” milestone, which translates to hitting 50 home runs and stealing 50 bases in one season. This new stat surpasses records set by then-Seattle Mariner Alex Rodriguez in 1998, when he achieved a “42/42,” and Atlanta Braves outfielder Ronald Acuña Jr., who hit 41 home runs and stole 73 bases in 2023. It’s particularly impressive because most players are either muscular power hitters or speedy base stealers, not both. HISTORY!SHOHEI OHTANI IS THE ONLY MEMBER OF THE 50/50 CLUB. pic.twitter.com/F1T5D4n6QD— Los Angeles Dodgers (@Dodgers) September 19, 2024 This latest record also only adds to Ohtani’s dominance in Major League Baseball. First signed to the Los Angeles Angels in 2017, Ohtani has long been considered uniquely talented because of how good he is at both hitting and pitching, a rare combination. Given his strengths as a “two-way player” – of a caliber not seen since the likes of Babe Ruth – he secured one of the most lucrative contracts in the sport when the Dodgers offered him a 10-year $700 million deal in 2023. Since joining the MLB, Ohtani has become the first player in recent memory to be in the top 15 for both home runs scored and strikeouts pitched in a single season.  In addition to being the only member of the 50/50 club, Ohtani turned in a historically strong game on Thursday. In six at-bats, Ohtani hit three home runs, two doubles, and a single, scoring every time. His hits also led to 10 runs batted in (RBIs), setting a new Dodgers record. And on top of that, he stole two bases. (That means he didn’t just hit the 50/50 mark, he actually now has a record 51 home runs and stolen bases.)  Notably, Ohtani is also still recovering from an elbow surgery that’s left him unable to pitch this season.  All that has made Thursday’s game a neat encapsulation of what has made Ohtani a star. He’s demonstrated uncommon versatility in the game as a commanding pitcher and hitter, and now a record-breaking base-stealer, too. Players recovering from surgery often have slow seasons, but that just has not been the case for Ohtani. And as the Washington Post notes, it can take other players several games to do what Ohtani did Thursday in one. That Ohtani was able to achieve so much in Thursday’s game speaks to why he’s one of the most-hyped athletes in Major League Baseball, and already considered by many fans to be one of the greatest players of all time.  What the record means The record is a testament to Ohtani’s unique power as a hitter, as well as his speed.  Ohtani had one of his strongest offensive games of the season on Thursday, ultimately helping the Dodgers land a spot in this year’s playoffs with the runs he scored. He’s also refined his ability to steal bases, improving his “running mechanics” and broader offensive techniques, according to the Wall Street Journal. As ESPN notes, players in Ohtani’s current position — designated hitter, an athlete who stands in to bat for the pitcher — tend to be slow. Before this year, Ohtani hadn’t stolen more than 26 bases in a single season. Following an injury in 2023, Ohtani was forced to take a break from pitching and instead used that time and energy to improve his base stealing. His latest success also comes after a dramatic sports gambling scandal earlier this year involving his former interpreter Ippei Mizuhara, who allegedly stole money from Ohtani to cover debts and pleaded guilty to bank fraud. After an investigation by the League, Ohtani was ultimately exonerated of wrongdoing.  Ohtani’s record-shattering game this week firmly established his ascendancy in the sport and was broadly cheered by other baseball greats including the Oakland Athletics’ Jose Canseco, who in 1988 was the first player to hit “40/40,” and fellow players like Tampa Bay Rays’ Taylor Walls, who said that his team was watching Ohtani’s game while playing its own.  “This guy is unreal,” basketball legend Lebron James posted on X. Ohtani’s colleague on the Dodgers, second baseman Gavin Lux, best summed up Thursday’s game and Ohtani’s style of play. “That has to be the greatest baseball game of all time. It has to be,” Lux said after the game. “There’s no way. It’s ridiculous. I’ve never seen anybody do that even in Little League, so it’s crazy that he’s doing that at the highest level.”

In 1888, Eastman Kodak patented roll film and released the Kodak No. 1 box camera. The company’s business model — selling film, then processing and printing the photos taken on that film for their consumers — made photography available to the masses for the first time. Before then, photography was a complicated process requiring knowledge of chemistry and expertise working big, bulky equipment. Later, when Kodak introduced the Brownie in 1900 and sold it for a dollar, photography went fully mainstream. The company dominated the film sales and development market during the 20th century and successfully marketed its automatic cameras as crucial to capturing fleeting moments — at home and on vacation. But digital camera sales began to outpace film camera sales in the early 2000s, and Kodak failed to keep up. They filed for bankruptcy in 2012 but do still exist and sell film, albeit to a much smaller market.You can find this video and all of Vox’s videos on YouTube.

North Carolina Lt. Gov. Mark Robinson delivers remarks prior to Republican presidential nominee former President Donald Trump speaking at a campaign event at Harrah’s Cherokee Center on August 14, 2024, in Asheville, North Carolina.  | Grant Baldwin/Getty Images Mark Robinson, the GOP nominee for governor in North Carolina endorsed by former President Donald Trump, was already known to be an extreme right-wing candidate who appeared to be alienating voters in a key battleground state. New revelations about his inflammatory comments on a porn site and his sexual exploits have now shaken even some of his Republican colleagues.  CNN reported on Thursday that Robinson regularly posted on a porn website’s message board between 2008 and 2012. In those posts, he identified himself as a “black NAZI!”, supported a revival of slavery, described himself as a “perv” for enjoying transgender porn, and admitted to “peeping” on women in public showers as a teenager. The posts were found on the site “Nude Africa.” Despite now supporting legislation that would ban abortion at around six weeks of pregnancy, he also said on the forum that he would not care if a celebrity had an abortion, though he would “wanna see the sex tape!” And though he would become North Carolina’s first Black governor if elected, he referred to Martin Luther King Jr. as a “commie bastard.” Robinson’s email has also been connected to an account on Ashley Madison, an online dating website for people seeking to have an affair. Robinson, North Carolina’s current lieutenant governor, denied writing the posts, and his campaign said he has not made an Ashley Madison account. Nevertheless, several North Carolina Republicans, including some running in competitive races this fall, subsequently pressured him to drop out of the contest to succeed incumbent Democratic Gov. Roy Cooper, who is term-limited. He has opted to stay in the race, in which he is running behind his Democratic rival, North Carolina Attorney General Josh Stein, by between 5 and 14 percentage points in polls conducted over the last month. In doing so, Robinson might drag down not just down-ballot Republicans in a closely divided purple state, but also potentially Trump.  “We have few examples of reverse coattails where a down-ballot candidate hurts the top of the ticket,” Whit Ayres, a GOP pollster, told Vox. “But if anyone could do it, it’s this character.” Robinson is extreme even for MAGA The current scandals don’t mark the first time Robinson has been mired in controversy.  He has hurled hateful remarks at everyone from Michelle Obama to the survivors of the Parkland school shooting. He’s called the LGBTQ community “filth.” He threatened to use his AR-15 against the government if it “gets too big for its britches,” and he wants to outlaw all abortions as well as return to a time when women couldn’t vote. He’s also ridiculed the Me Too movement, women generally, and climate change. It seems Robinson is willing to entertain all manner of conspiracy theories, too. He’s a Holocaust denier and has a history of antisemitic remarks. He’s suggested that the 1969 moon landing might have been fake, that 9/11 was an “inside job,” that the music industry is run by Satan, and that billionaire Democratic donor George Soros orchestrated the Boko Haram kidnappings of school girls in 2014. In spite of all of this, Robinson was not only able to win his party’s nomination for the state’s most powerful position, but he did so by a margin of more than 45 percent over his rivals. The other Republican candidates, trial lawyer Bill Graham and state treasurer Dale Folwell, raised concerns about Robinson’s electability, but ultimately neither could compete with his name recognition nor his MAGA bona fides in a state that twice voted for Trump. Will Robinson hurt the GOP’s chances this fall in North Carolina? Robinson might function as an ideal foil for Democrats — not just in the governor’s race, but also in the presidential and down-ballot races. Robinson might struggle to capture the more than 35 percent of GOP voters who opposed him in a contentious primary. Republican leaders certainly seem concerned. “We knew that [he was extreme], but I still think the revelations over the last 24 hours are stunning,” former Rep. David Price (D-NC) told Vox. “I believe the Republican leaders know that as well. A corner has been turned in terms of Robinson being able to count on even the most faithful Republican supporters.”  Rep. Patrick McHenry (R-NC) told the Hill that the reports about Robinson are “not good.” And Rep. Richard Hudson (R-NC), chair of the National Republican Congressional Committee, said they were “very concerning.”  “My hope is that [the] lieutenant governor can reassure the people of North Carolina that the allegations aren’t true,” he said. “He said they’re not true. I think he needs to have the opportunity to explain to the people in North Carolina exactly how these allegations aren’t true.” The fact that even Republicans are distancing themselves from Robinson suggests that the tide may turn further in Democrats’ favor in North Carolina.  That doesn’t mean, however, that Robinson will deliver Democrats a major victory. Trump and Vice President Kamala Harris are practically tied in the state, according to FiveThirtyEight’s polling average. And Republican candidates have won three presidential elections in North Carolina since Barack Obama surprisingly took the state in 2008 — including Trump twice. The state has stayed on the red side of purple, though it does have a tradition of split-ticket voting. That helped power Cooper’s two election wins and has added to the belief that the state may be in reach for Democrats.  “I don’t think celebration is called for because so much more is at stake than just the governor’s race,” Price said. “We know how tight this is.”

In this photo illustration, the sign of OpenAl o1, the first in a planned series of “reasoning” models that have been trained to answer more complex questions, is displayed on a smartphone screen on September 13, 2024, in Suqian, Jiangsu Province of China. An underappreciated fact about large language models (LLMs) is that they produce “live” answers to prompts. You prompt them and they start talking in response, and they talk until they’re done. The result is like asking a person a question and getting a monologue back in which they improv their answer sentence by sentence. This explains several of the ways in which large language models can be so frustrating. The model will sometimes contradict itself even within a paragraph, saying something and then immediately following up with the exact opposite because it’s just “reasoning aloud” and sometimes adjusts its impression on the fly. As a result, AIs need a lot of hand-holding to do any complex reasoning. This story was first featured in the Future Perfect newsletter. Sign up here to explore the big, complicated problems the world faces and the most efficient ways to solve them. Sent twice a week. One well-known way to solve this is called chain-of-thought prompting, where you ask the large language model to effectively “show its work” by “‘thinking” out loud about the problem and giving an answer only after it has laid out all of its reasoning, step by step.  Chain-of-thought prompting makes language models behave much more intelligently, which isn’t surprising. Compare how you’d answer a question if someone shoves a microphone in your face and demands that you answer immediately to how you’d answer if you had time to compose a draft, review it, and then hit “publish.” The power of think, then answer OpenAI’s latest model, o1 (nicknamed Strawberry), is the first major LLM release with this “think, then answer” approach built in.  Unsurprisingly, the company reports that the method makes the model a lot smarter. In a blog post, OpenAI said o1 “performs similarly to PhD students on challenging benchmark tasks in physics, chemistry, and biology. We also found that it excels in math and coding. In a qualifying exam for the International Mathematics Olympiad (IMO), GPT-4o correctly solved only 13 percent of problems, while the reasoning model scored 83 percent.” This major improvement in the model’s ability to think also intensifies some of the dangerous capabilities that leading AI researchers have long been on the lookout for. Before release, OpenAI tests its models for their capabilities with chemical, biological, radiological, and nuclear weapons, the abilities that would be most sought-after by terrorist groups that don’t have the expertise to build them with current technology.  As my colleague Sigal Samuel wrote recently, OpenAI o1 is the first model to score “medium” risk in this category. That means that while it’s not capable enough to walk, say, a complete beginner through developing a deadly pathogen, the evaluators found that it “can help experts with the operational planning of reproducing a known biological threat.”  These capabilities are one of the most clear-cut examples of AI as a dual-use technology: a more intelligent model becomes more capable in a wide array of uses, both benign and malign.If future AI does get good enough to tutor any college biology major through steps involved in recreating, say, smallpox in the lab, this would potentially have catastrophic casualties. At the same time, AIs that can tutor people through complex biology projects will do an enormous amount of good by accelerating lifesaving research. It is intelligence itself, artificial or otherwise, that is the double-edged sword. The point of doing AI safety work to evaluate these risks is to figure out how to mitigate them with policy so we can get the good without the bad. How to (and how not to) evaluate an AI Every time OpenAI or one of its competitors (Meta, Google, Anthropic) releases a new model, we retread the same conversations. Some people find a question on which the AI performs very impressively, and awed screenshots circulate. Others find a question on which the AI bombs — say, “how many ‘r’s are there in ‘strawberry’” or “how do you cross a river with a goat” — and share those as proof that AI is still more hype than product.  Part of this pattern is driven by the lack of good scientific measures of how capable an AI system is. We used to have benchmarks that were meant to describe AI language and reasoning capabilities, but the rapid pace of AI improvement has gotten ahead of them, with benchmarks often “saturated.” This means AI performs as well as a human on these benchmark tests, and as a result they’re no longer useful for measuring further improvements in skill. I strongly recommend trying AIs out yourself to get a feel for how well they work. (OpenAI o1 is only available to paid subscribers for now, and even then is very rate-limited, but there are new top model releases all the time.) It’s still too easy to fall into the trap of trying to prove a new release “impressive” or “unimpressive” by selectively mining for tasks where they excel or where they embarrass themselves, instead of looking at the big picture.  The big picture is that, across nearly all tasks we’ve invented for them, AI systems are continuing to improve rapidly, but the incredible performance on almost every test we can devise hasn’t yet translated into many economic applications. Companies are still struggling to identify how to make money off LLMs. A big obstacle is the inherent unreliability of the models, and in principle an approach like OpenAI o1’s — in which the model gets more of a chance to think before it answers — might be a way to drastically improve reliability without the expense of training a much bigger model.  Sometimes, big things can come from small improvements  In all likelihood, there isn’t going to be a silver bullet that suddenly fixes the longstanding limitations of large language models. Instead, I suspect they’ll be gradually eroded over a series of releases, with the unthinkable becoming achievable and then mundane over the course of a few years — which is precisely how AI has proceeded so far.  But as ChatGPT — which itself was only a moderate improvement over OpenAI’s previous chatbots but which reached hundreds of millions of people overnight — demonstrates, technical progress being incremental doesn’t mean societal impact is incremental. Sometimes the grind of improvements to various parts of how an LLM operates — or improvements to its UI so that more people will try it, like the chatbot itself — push us across the threshold from “party trick” to “essential tool.”  And while OpenAI has come under fire recently for ignoring the safety implications of their work and silencing whistleblowers, its o1 release seems to take the policy implications seriously, including collaborating with external organizations to check what their model can do. I’m grateful that they’re making that work possible, and I have a feeling that as models keep improving, we will need such conscientious work more than ever.  A version of this story originally appeared in the Future Perfect newsletter. Sign up here!

Marine animals known as sea squirts, shown here, produce defense compounds that can damage cancer cells. Scientists have used them to produce anticancer drugs. | Reinhard Dirscherl/ullstein bild via Getty Images Stuck to rocks, shells, and piers in oceans around the world is a strange little creature called a sea squirt. It resembles a squishy potato and has two valves poking out, which it uses to suck in and expel seawater. Sea squirts are special for a few reasons. They tend to shoot water out of their valves when you squeeze them. And like oysters and clams, they filter ocean water, helping keep it clean. They also produce chemical compounds to defend themselves that are known to damage cancer cells. Scientists have used those compounds to develop drugs for patients with certain kinds of cancer.  Sea squirts are among an endless list of animals, plants, and microbes that stand to improve human lives.  Researchers estimate that an astonishing 70 percent of antibiotics and cancer treatments in use today are rooted in natural organisms, from plants to snakes to sea sponges. The first medication to treat HIV came from a Caribbean sea sponge. The cosmetic drug Botox is derived from a bacterium. The enzyme used to stonewash jeans was originally derived from wild microbes in salt lakes in Kenya.  Collectively, these natural derivatives — and the profits they generate — are considered the benefits of a planet with healthy biodiversity. And maintaining these benefits is a key justification for protecting nature: It can literally save our lives. But a key question that has long been a source of division among global conservation leaders is who, exactly, should reap those biodiversity benefits. There’s a long history of what some advocates and researchers call biopiracy. It typically refers to when companies take organisms from poor nations and Indigenous communities, such as medicinal plants, and use them to develop commercial products, failing to share the benefits back with them. Those benefits include things like money but also access to those products and research results. Until recently, the solution to exploitative innovation was, at least in theory, relatively straightforward. Under a United Nations treaty called the Convention on Biological Diversity (CBD), countries can require that foreign researchers sign benefit-sharing agreements before granting them permission to retrieve wild organisms. Essentially, under those agreements, you can’t walk out of a country with medicinal plants without first agreeing on how you’ll compensate that country and its people. But there’s an enormous loophole to this approach that keeps getting bigger.  Major advances in biotechnologies have made it easier to sequence and analyze DNA. Now, researchers and companies no longer rely only on physical samples to make stuff derived from biodiversity. They increasingly make products — drugs, vaccines, better crops, and so on — using DNA and RNA sequences, or other genetic data. This digital biodiversity data, referred to by the arcane term digital sequence information, or DSI, is found abundantly in scientific databases that are free for anyone to use. And it’s not regulated by the CBD treaty. That means industries can create vaccines and other commercial products using DSI without sharing the benefits from those products — the benefits of biodiversity — with whatever country or local community the sequence information originates from. It’s kind of like what platforms like Napster did to the music industry: Instead of having to buy CDs from stores and funneling money to musicians, once music was digitized, you could eventually stream unlimited music online for free. If you find DSI confusing, that’s because it is.  Experts who have been debating for years about how to regulate DSI don’t even agree on how to define the term. Does it cover just DNA and RNA sequences? Would it also include 3D images of proteins and epigenetic data (i.e., changes to how genes are expressed)? They also don’t agree on how benefit sharing should work. Some countries, especially those with smaller economies, want genetic information tracked from its place of origin all the way to the final products. Others say that’s essentially impossible. This debate will come to a head this October in the Colombian city of Cali. Countries that are party to CBD — which notably does not include the US, in part because conservative lawmakers tend to dislike global treaties — will convene for their biannual meeting to discuss global conservation issues. One of their main goals this year, at what will be known as COP16 — but not that COP — is to hammer out a plan to bring more accountability to the use of DSI around the world.  On one hand, such a plan seems impossible to put in place. Companies hold a tremendous amount of power and typically want fewer regulations, not more. But it could also be a massive opportunity. If developed nations and industries shared some of the money and knowledge that is derived from digital biodiversity data, it could be used to conserve nature in the places where it is most vital — and most at risk. Who benefits from nature? The debate and tensions around DSI are rooted in inequality. Put simply, rich nations have loads of scientific resources, whereas many poorer nations have loads of less-explored biodiversity. And up until now, the relationship between the two groups has been lopsided.  Decades ago, a US pharmaceutical company developed anticancer drugs with the help of a plant from Madagascar called the rosy periwinkle; the company didn’t share its profits with the people of Madagascar. You can find similar stories with the antifungal spray Neemax, derived from a tree in India, and muscle relaxants made with compounds from curare, a group of poisonous plants from the Amazon.   “Scientists from the global north have frequently extracted data and samples from the Global South without the permission of the people there, without collaborating meaningfully — if at all — with local scientists, and without providing any benefit to the countries where they conduct their work,” a team of researchers wrote earlier this year.  Global environmental leaders recognized this problem decades ago. When they established the Convention on Biological Diversity in 1992, still the world’s most important biodiversity agreement, they made benefit-sharing one of three main goals of the treaty, along with conserving biodiversity and using it sustainably. Under the agreement, benefits derived from plants and animals should, at a minimum, be shared with the countries and local communities where that biodiversity is found — and especially with the groups who have safeguarded it, such as Indigenous communities. Nearly two decades later, CBD made the requirements around benefit-sharing more concrete and enforceable through an agreement called the Nagoya Protocol, named after the Japanese city where it was adopted. The agreement essentially affirms that countries have the legal right to regulate access to physical plants, animals, and other elements of biodiversity within their borders. All countries are also supposed to make sure that any bits of biodiversity they — or their companies — use that come from other nations are collected with the consent of that country.  Share your feedback Do you have feedback on this story or tips for the author? Reach out to Vox reporter Benji Jones at benji.jones@vox.com. The Protocol has, at best, a mixed record. Middle-income nations, like Brazil, or those with a lot of donor support, have established systems that work. In many poorer nations, however, access is still poorly regulated or unregulated. In general, very little money has flowed into countries via the Nagoya Protocol, said Marcel Jaspars, a professor at the University of Aberdeen and a leading expert on DSI in the Global North. DSI only adds to these benefit-sharing woes. When environmental leaders crafted the CBD and the Nagoya Protocol, digital biodiversity data wasn’t as easily accessible or as useful as it is today; these agreements don’t even mention DSI. It’s widely understood that CBD and the Protocol only pertain to physical materials — microbes, plants, compounds from a sea squirt — not genetic sequences. That leaves the use of DSI, now a massive source of scientific innovation, largely unregulated.  What DSI is and how it works DSI is one of the most confusing concepts in the environmental world, which is already racked with confusing terminology and technical jargon. Here’s the gist: After researchers collect plants, animals, and other organisms, they commonly sequence their DNA, or part of it, and upload that information to a database. The largest global collection of DNA and RNA sequences — which is the subject of much of these discussions — is (take a breath) the International Nucleotide Sequence Database Collaboration. It houses billions of genetic sequences and is free for anyone to use. Downloading the data and using it to develop commercial products does not trigger the legal obligations under CBD that harnessing a biological sample would. You’re basically harvesting information from a computer instead of from the environment. Scientists use DSI for a mind-bending array of projects. Consider the Moderna Covid-19 vaccine. The company used nearly 300 genetic sequences, according to the patent, many of which were drawn from open-access databases, to produce the shot (which the company was able to design in just two days). Researchers also use DSI to figure out how unique a particular genetic sequence might be, or what it might do — as in, what physical trait it might be linked to. This is incredibly valuable for the biotech and agriculture industries. A seed company, for example, might have a crop that appears more drought tolerant in their private collection. They can sequence the plant’s DNA and cross-reference its genetic information with online databases. Those databases often list information about the role of different sequences. Ultimately, this can help the company identify which particular sections of the plant genome might be associated with an ability to survive droughts, a valuable trait. Artificial intelligence, including projects like Google’s AlphaFold, makes these sorts of predictions even easier. Conservation scientists also benefit from DSI in a big way. They increasingly rely on an approach called environmental DNA (eDNA) to catalog what species live in a particular area, such as a stream or the forest floor. Researchers will gather samples of water or soil and filter out bits of DNA that were shed by animals or microbes into the environment. Then they’ll look for a direct match with those sequences in open-access databases, revealing what those animals are. If the species are rare or otherwise considered important, this information could, say, help justify protecting habitat.  This is to say: DSI is useful! There’s a good reason it’s open to everyone. It both enables and speeds up research, some of which is literally life-saving. At the same time, it maintains inequities and furthers exploitation when the people who prosper from it are largely in wealthy economies (a problem that’s especially pronounced and worrying when it comes to developing vaccines.) “DSI makes it possible to get all kinds of commercial advantages,” said Michael Halewood, an expert in genetic resource policy at CGIAR, a global agriculture organization. “That creates a big gap that needs to be closed. We all agree on the inequities of the situation. What’s a sensible way to close that gap without undermining science?” What a plan to regulate DSI might look like Even as COP16 looms, the debate around DSI is still a mess. There’s a lack of trust between country negotiators, leading to an enormous amount of uncertainty about how this digital data might be regulated on a global scale.  Consensus has, however, grown around the idea that companies, entire sections of the economy, or perhaps even consumers should pay into a fund that supports conservation and development, especially in the Global South. From there, two big questions rise to the top: Who, exactly, pays to use DSI, and who ultimately receives those payments? A handful of low- and middle-income countries want what negotiators refer to as a “track and trace” system. That would entail tracing genetic sequences, i.e., DSI, from open-access databases to specific products, such as medicines or drought-tolerant crops, that generate value. So if a company in Europe makes a drug using DNA from a sea squirt found in, say, Panama, it would have to share benefits from that drug — money and also access to the drug itself, perhaps — back with the Panamanian people. Many academic researchers and developed countries call this a nonstarter. The chain of development for drugs and other products is long and tangled and relies on hundreds if not thousands of individual sequences. Even if it’s clear that an end product is based on a specific bit of genetic code, the research process to find that sequence — which could involve scientists across dozens of institutions, all using their own bits of DNA — relies on sifting through unimaginable amounts of genetic material. That makes it hard to determine who, exactly, should receive the benefits. Plus, many of the sequences in global databases don’t come with location information; geotagging wasn’t required until somewhat recently. That makes it even harder to direct benefits to a specific country.  Scientists also fear that a complex tracking system would slow the pace of innovation, and be incredibly expensive to operate. “By making track and trace a necessity, the system will cost more than it generates, almost certainly,” Jaspars said. Many of the groups who oppose a track and trace approach, including many developed economies and academic scientists, prefer what they describe as a “sector” approach. This would require companies in sectors of the economy that are highly dependent on DSI — such as agriculture and pharmaceuticals — to put a small percentage of their profits or sales (or other measure of value) into a DSI fund. That money would likely be dished out to countries or specific projects for the benefit of conservation and human development. Proponents of this approach argue that it would allow money from DSI to flow quickly; it wouldn’t be contingent on companies profiting from specific DSI-based products. It’s also simpler because it doesn’t involve tracking sequences.  While this sector approach has a lot of support from scientists and rich countries, it’s still not clear how to determine which industry sectors, or parts of sectors, would need to pay up. The corporate world, meanwhile, has serious concerns about requiring payments from broad sectors of the economy, according to Daphne Yong-D’Hervé, who leads global policy at the International Chamber of Commerce. Different companies use vastly different quantities of DSI, she said. And generally speaking, trying to regulate DSI as separate from physical materials is problematic, Yong-D’Hervé said. Organisms and their genetic sequences are often used collectively during R&D.  Ultimately, she said, what corporations want is a simple, unified system to use DSI and physical materials that gives them a license to operate worldwide. Without paying too much, of course. “Businesses support the principle of benefit sharing, but this has to be implemented in a way which is aligned with scientific and business realities, is simple, and does not discourage investments in research and innovation,” Yong-D’Hervé told Vox. Then there are some wealthy countries, such as Japan and Switzerland, that seem to be in favor of the status quo. They prefer a deal that encourages companies that actively use digital biodiversity data to contribute to a DSI fund, but without the legal obligation to do so.  But advocates for lower-income countries say this, too, would be a nonstarter — a continuation of exploitation.  “We are asking for accountability,” said Nithin Ramakrishnan, a senior researcher at Third World Network (TWN), a group that advocates for human rights and benefit sharing. To TWN, Ramakrishnan says, the priority is getting an agreement that makes sharing benefits from DSI mandatory.   Will this ultimately help biodiversity? None of these proposals are perfect, and they represent only a handful of the issues pertaining to DSI that countries disagree on. There are questions about sharing benefits other than money and access to drugs, such as lab equipment. Some researchers and advocates are also concerned about who will manage the DSI fund and about the databases that store genetic data. The big databases are largely hosted by organizations in developed nations, so poorer countries have little control over how they operate, Ramakrishnan said.  Making everything more complicated is the reality that there are other international treaties — including those pertaining to crops and the high seas — that are also trying to figure out how to manage access to DSI. Regulating genetic data on a global scale will likely only work if all of these treaties are aligned and define DSI in the same way.  And even if countries come to an agreement, it’s not clear they’ll be able to enforce it at the national level. (This is a problem for other international agreements. The 2015 Paris Accord, for example, lacks teeth because it doesn’t have a strong enforcement mechanism). Can environmental officials get entire corporate sectors to pay up? In some countries, these payments might also require governmental approval.  Also not helping: The US, the world’s premier scientific power, is not a party to CBD, so it wouldn’t be bound by any framework that officials finalize in Cali. (However, some of the big US pharmaceutical companies have told Jaspars they are “open to sharing benefits.”) So yes, crafting an effective plan will not be easy. Then again, the payoff of such a system could be huge — it could be lifesaving.  Researchers estimate that the gap in funding for biodiversity conservation globally is somewhere around $700 billion a year; that’s a key reason why biodiversity is in peril. And critically, any payments for using DSI could help close that gap, especially if they’re generated from entire sectors.  Far more support is needed for things like restoring coral reefs, managing parks, and preventing wildfires, much of which Indigenous groups and local communities have already been doing. These efforts help ensure that biodiversity, and all the secrets it still holds, is left intact.  “There’s a whole incredible world still to discover,” said Sarah Laird, co-director of People and Plants International, a nonprofit environmental organization. “We know a lot, but there are things we can’t even imagine out there. There are amazing opportunities.”

A Trump supporter holds a “stop the steal” sign while gathering on the steps of the Colorado State Capitol to protest the election on January 6, 2021, in Denver, Colorado. | Michael Ciaglo/Getty Images Are we barreling toward a legitimacy crisis in this election? The polls show a tight race between Vice President Kamala Harris and former President Donald Trump, suggesting that the 2024 election, like the 2020 one, may be decided by narrow margins in a few battleground states. And like last time, polls suggest a sizable proportion of Republican voters seem poised to reject the results if Trump comes up short.  If Trump loses, about a quarter of Republicans said they think he should do whatever it takes to ensure he becomes president anyway, according to a September PRRI poll.  That may include resorting to violence: Among Republicans who don’t believe Biden’s win in 2020 was legitimate, almost one-third said in an August poll by the SNF Agora Institute at Johns Hopkins University that they expected “a lot” or “a great deal” of political violence after the November election. That doesn’t mean that violence is inevitable — but after the deadly January 6, 2021, insurrection, the possibility cannot be dismissed out of hand.   These beliefs may stem from the fact that, among Republicans, Trump proved by far the most trusted source of information about election results, well above local and national news outlets. In an Associated Press/NORC/USAFacts poll from earlier this month, more than 60 percent of Republicans said they believe Trump himself is the best place to get the facts about results.  The problem with all this, of course, is that the former president has been very consistent in falsely claiming that he won the 2020 election, and in casting doubt on the legitimacy of US elections ahead of the 2024 contest.  During his first debate against Harris, Trump again refused to acknowledge that he lost the 2020 election, walking back his comments in a podcast interview earlier this month in which he said he “lost by a whisker.” And despite facing criminal charges for pressuring election officials to overturn the results in 2020, Trump has not indicated that he will accept the results in November.  “We have to have good elections. Our elections are bad,” he said during the debate. Trump has also threatened to prosecute “those people that CHEATED” in the 2024 election and subject them to long-term prison sentences if he wins.  His running mate, Sen. JD Vance (R-OH), also said in a podcast appearance this month that he wouldn’t have certified the results of the 2020 election if he had been in Congress at the time. “I would have asked the states to submit alternative slates of electors and ask the country to have a debate,” he said.  Trump’s long-running insistence that he won in 2020 appears to be having an effect over time, with several surveys measuring greater buy-in of his lies about the election from voters today than in the past. A December Washington Post/University of Maryland poll found that 36 percent of US adults did not believe Biden was legitimately elected, compared to 29 percent two years prior. And in a Pew Research poll conducted earlier this month, 27 percent of US adults said that Trump did nothing wrong in trying to overturn the election results, up from 23 percent in April. The pervasiveness of Trump’s lies about the election seems to have also contributed to Republican fears about the future of democracy, which polling suggests are more acute than among Democrats. Republicans were much more likely to say American democracy isn’t working and that it would end in their lifetimes than Democrats in a March Quinnipiac poll. Overall, surveys suggest that many Americans believe there’s a real possibility that Trump won’t accept the election results. An August ABC/Ipsos poll found that only 29 percent said they believed Trump was prepared to accept the results regardless of the outcome.  Add it all up and the picture is troubling: If the election is close — and all indications are that it will be — we seem to be set up for a genuine legitimacy crisis if Trump were to lose. It’s yet another indication of US democracy’s perilous state and how Trump has bent American politics to his will. Can the center hold?  Trump’s words and actions — and his followers’ beliefs — stand in contrast to Harris’s and her supporters.  In her Democratic National Convention speech last month, Harris committed to a peaceful transfer of power. Democrats show the same commitment: The August ABC/Ipsos poll found that 92 percent were prepared to accept the results, regardless of the outcome, in contrast to 76 percent of Trump supporters.  That said, recent reports about Trump allies laying the groundwork to undermine the election results in some key states could certainly end up raising questions for some Democratic voters about the legitimacy of the results if he wins.  In Georgia, for instance, members of the state elections board associated with the Trump-aligned “stop the steal” campaign have signed off on new rules that allow local elections officials to make “reasonable inquiry” into election irregularities — without defining what might constitute a “reasonable” inquiry. That could potentially lead to frivolous challenges that could complicate the certification of what is expected to be a close election in Georgia.  Moreover, the polarized state of the country means a Trump victory, especially a narrow one, would likely prompt the kind of massive peaceful protests the country saw in the wake of the 2016 election (though, of course, accepting the results and peacefully protesting them is entirely different from storming the Capitol and attempting to stop the peaceful transfer of power).  As we near Election Day, it’s well worth taking a step back and assessing the state of American democracy in 2024. One of the two major party candidates refuses to accept his defeat four years ago and shows every sign of doing the same thing again this year if he loses. His supporters are right there with him.  That the election remains close raises the risk of a true crisis but is also a reminder of how, because of Trump, what was once unthinkable is now reality. 

Traci Coston, a serious young Christian TikToker, stares directly at the camera.  “One day, an animator was messing around, and he created this picture of a little minion.” Coston points up to an image of a sausage-bodied minion hanging limply from a wooden cross.  Easily among the most recognizable anti-heroes in film history, minions are the stars of a $5 billion animated movie franchise. The enduring success of the franchise can be attributed, at least in part, to their extremely online fans, who have been feverishly producing minions memes since the release of the first movie, Despicable Me, in 2010.  “Listen to this,” she continues, “a minion didn’t die for you, but somebody actually did. Jesus actually died for you.”  Coston had sacrificed the Minion Jesus for my attention, and the crucified minion had done its job. I had not scrolled away. His death was the key that unlocked the door to my carefully curated TikTok algorithm, and Coston had walked right through.   The 40-second video was as sacrilegious as it was sincere. It was deeply funny and deadly serious. I typed “Minion Jesus” into the search bar and dove headlong into the wild world of high-kitsch TikTok evangelism. What I learned was that Coston isn’t a lone wolf. A dozen other TikTokers had recently posted Minion Jesus videos, all repeating the same near-verbatim message in front of the exact same minion torture scene.  The scriptedness of the videos was unusual. On TikTok, sounds and trends go viral, and participation in memes tends to be iterative and improvisational. In the comment sections of these scripted Minion Jesus videos, hundreds of TikTokers gathered to discuss the mystery. Who wrote this script? Do these people know each other? Are they AI deep-fakes? Are they joking?  For the Today, Explained podcast, I decided to get to the bottom of it.  The road to Minion Jesus is long and rambling. It takes us from a tiny town in Louisiana all the way to the steps of White House. But our tale begins in the halls of Facultad de Artes Plásticas, a small art school in Arteaga, Mexico. The Minion Jesus origin story  In 2021, a young graphic design student named Americo Cruz sketched out a Minion Jesus on paper, opened up his 3D modeling software, and got to work. Two days later, he posted it to his personal Facebook page. Cruz told Vox he took inspiration from a 2015 meme: a photograph of a stuffed toy minion engulfed in flames and fastened to a cross. “The truth is that it’s just a parody — a simple joke,” he said. “At the end of the day, it is just the absurd humor that we have in Gen Z.” Cruz hadn’t seen the TikTok videos until I reached out to him for this story. How they’d gotten hold of Cruz’s image, though, seemed relatively straightforward. Minion Jesus went moderately viral in 2021, and it has since been shared thousands of times on Facebook, X, Reddit, and Instagram — all platforms where images are easy to access and download. In all likelihood, someone had stumbled across the image online, downloaded it, and penned the script.  Less clear was who, and whether the subsequent videos were a part of some sort of coordinated campaign. The Minion Jesus TikTokers post similarly structured content every day, most of it concerned with convincing the viewer to convert to Christianity. Several also make conservative political content.  They don’t all follow one another on their various social media platforms, but I soon realized they all have one connection in common: a blond, brawny evangelist named Taylan Michael Seaman. The makings of a Christian content farm Seaman is the 26-year-old self-proclaimed millionaire behind a coaching program called Kingdom University. From a small town in central Louisiana, he offers online courses to any would-be social media evangelist willing to “invest” thousands of dollars into his ministry. In exchange, he teaches them his “Viral Video Framework,” a basic digital content marketing strategy couched in the language of proselytism. The program offers its students a bold guarantee: “at least 100,000 new followers in 90 days.” As you may have guessed by now, the Minion Jesus video is a Kingdom University product.  @vox If you scroll away from this video then you’ll miss our producer explaining “Minion Jesus.” #minions #minionjesus #christianity #religion ♬ original sound – Vox – Vox According to several enrollees who spoke to Vox under the condition of anonymity, the coaches alert their students when a particular piece of content — like Minion Jesus — is doing well. They are then encouraged to follow suit, which is the reason the videos are often so eerily similar. Seaman’s personal content skews a bit more pessimistic. His videos are often concerned with demonic possession, and they tend to be more overtly political and caught up with conspiracy theories than the content created by his students. His message is also often about money. This is no coincidence. An explicit focus on wealth-building is integral to his ministry, and his videos contain clear nods to a midcentury offshoot of Pentecostalism commonly known as the Word of Faith Movement. How Word of Faith became an evangelical powerhouse In the 1960s, ​​itinerant evangelist Kenneth Hagin rose to fame by performing (and possibly inventing) the miracle of Holy Laughter in churches around the United States. At the wave of his hand, congregants would fall to the floor shrieking and convulsing in rapturous hysterics. This kind of ecstatic religious expression is a defining feature of the Pentecostal tradition, which originated in the early 20th century. At the height of Jim Crow, Pentecostalism emerged in California as a multiracial, working-class movement. The services were frenzied experiences, and congregants often jumped out of their seats during services, dancing in the aisles and crying out in what a 1906 Los Angeles Times story referred to as a “weird babel of tongues.”  Half a century later, Kenneth Hagin’s Word of Faith movement capitalized on that fervor. The laughing evangelist told his followers that poverty was “the penalty for breaking God’s law,” and true faith was the key to wealth. As he traveled the country spreading his Holy Laughter, Hagin asked for large donations from congregants and promised them miraculous returns on investment. Fifty years before TikTok appeared in the app store, Hagin began broadcasting his sermons over the radio and selling reel-to-reel tapes to anyone willing to pay for them. In 1974, Hagin founded Rhema Bible Training College, an unaccredited bible school in Oklahoma. Today, there are Rhema training centers in 14 countries, and course listings include classes on “how to apply proven techniques in the areas of multimedia, web and management to grow your congregation” and “the importance of the corporate governing documents, IRS and state legal requirements, liability issues, setting of compensation and benefits packages.”  In 2002, two Rhema graduates founded Faith Church Ruston, a nondenominational church in Ruston, Louisiana. Their son-in-law and protégée is the godfather of the Minion Jesus TikTok, Taylan Michael Seaman. The movement turned toward Christian nationalism “It’s the blood, the blood, the blood — the word, the name, the blood — the word, the name, the blood,” shouted Kenneth Copeland, an 84-year-old prophet with an estimated net worth of $300 million. One year into the Covid-19 pandemic, he was receiving a vision from God on a megachurch stage. For a brief moment, the prophet saw heaven, crying out: “There is Brother Hagin! There is Brother Hagin!”  By 2021, Word of Faith founder Kenneth Hagin had been dead for 18 years. In his absence, this new Kenneth had risen up: Kenneth Copeland is now the figurehead of the Word of Faith movement, a millionaire prophet with pinprick pupils and visions of blood. In the 1960s, a young penniless Copeland attended one of Hagin’s seminars and allegedly offered to trade his car for a collection of Hagins tapes. He managed to walk away with a few without giving up his vehicle and promptly memorized them. Within a few years, Hagin was acting as the young man’s mentor. Under the guidance of this new “spiritual father,” Copeland took the Word of Faith movement to new heights. He founded a Christian broadcast network, built a multimillion-dollar empire, and now claims that his flagship show reaches 885 million viewers a day. In the decades after Hagin’s death, Copeland’s theology and conservative politics began to meld. The Word of Faith movement in its current form has dovetailed neatly with another quasi-Pentecostal movement called the New Apostolic Reformation — a fusion of prosperity gospel theology and Christian Nationalist politics. Where the Word of Faith movement leveraged the faithful to amass wealth, the New Apostolic Reformation is leveraging that wealth to build political power.  Andy Kroll, who covers fringe religious movements in America for ProPublica, told Vox that where right-wing Christian political movements in decades past have encouraged Christians to effect change from within the political system, the New Apostolic Reformation’s vision for America is more ambitious and less democratic.  New Apostolic Reformation’s most prominent figure is Lance Wallnau, a regular guest on Kenneth Copeland’s television network, the movement’s loudest megaphone. For the last two decades, Wallnau has been scribbling out a vision he calls “The Seven Mountains Mandate” on dry-erase boards across the country. He argues that conservative Christians should “take dominion” over American life by assuming positions of power in seven key spheres of influence: arts and media, business, church, education, family, government, and science and technology.  “This segment of Christianity is not the largest by number, but it’s the fastest-growing,” says Kroll. Across the country, pastors are using Wallnau’s language to communicate a new kind of Christian Nationalism to their congregations.  You cannot, of course, spell “dominion” without the word “minion.”  “How To Have DOMINION!!” In July, the man behind the Minion Crucifixion TikToks livestreamed a sermon to his YouTube followers. For an hour and a half, Taylan Michael Seaman paced back and forth across Faith Church Ruston’s stage, microphone in hand, and delivered a sermon he titled “How To Have DOMINION!!”  Once fringe, this vision for America has found a champion in former president Donald Trump. Wallnau was one of the first Christian leaders to endorse Trump, and Trump has spent the past eight years borrowing liberally from Wallnau’s language and theology to appeal to his base.  Hours after Donald Trump survived the July 13 assassination attempt, Wallnau made a special appearance on Flashpoint, a Christian political show on Kenneth Copeland’s broadcast network. “There are some attacks, when they happen, God intervenes and he moves in a miraculous way,” said Wallnau, pointer finger raised, “and I pray for a miraculous intervention of God now on President Trump.” The next day, Trump logged into his Truth Social account and typed, “Thank you to everyone for your thoughts and prayers yesterday, as it was God alone who prevented the unthinkable from happening. We will FEAR NOT, but instead remain resilient in our Faith and Defiant in the face of Wickedness.” Who could have more influence than … minions?  In the tradition of Hagin, Copeland, and Wallnau, Seaman has come to recognize that he is only as powerful as his flock is large. His YouTube audience? 3 million.  “When I go online and I look up Minion Jesus, and when I read about Taylan, I see an archetype,” said Kroll. “Someone who is common throughout this Wild West of Christianity. Someone who’s really thinking about how to break through on whatever the latest communication platform is.” At this point in his career, Seaman is nowhere near as influential or wealthy as the trailblazers that came before him. There is also a chance his subscriber count has been inflated by purchased followers. This is a common practice, and the fact that Taylan’s Kingdom University guarantees enrollees 100,000 followers in a matter of months does raise red flags. Regardless, he is effectively exploiting a loophole that his predecessors could only have dreamed of.  Put simply, I didn’t go searching for Minion Jesus. Minion Jesus came looking for me.  Hagin, Copeland, and Wallnau have demonstrated that there are myriad ways to become a millionaire Christian influencer.  Perhaps the simplest one to date is Seaman’s. Game the system. Crucify a minion. Let the algorithm do the rest.

Solar power deployment is exceeding even the most optimistic forecasts. Every day, the sun’s rays send 173,000 terawatts of energy to Earth, 10,000 times the amount used by all of humanity. Which is to say, the potential for solar energy is immense, and we’re nowhere near the limit. That’s why solar energy is such an appealing prospect, particularly as an alternative to the fossil fuels that cause climate change. And over the past decade, solar energy technology has vastly improved in performance and plummeted in cost.  As a result, photovoltaic panels have cropped up like dandelions across fields and rooftops at a stunning pace. Yet even the people most plugged-in to the energy industry and most optimistic about solar power continue to underestimate it. In fact, it’s a long-running joke among energy nerds that forecasters keep predicting solar will level off as it continues to rocket up to the sun. “Solar does continue to surprise us,” said Gregory Nemet, who wrote How Solar Energy Became Cheap, in an email. “It seems like it shouldn’t at this point. It’s been roughly 30 percent growth each year for 30 years. And costs continue to fall so new users — and new uses — continue to emerge.” In the past year, solar power has experienced Brobdingnagian growth, even by solar standards. According to a new report from Ember, an energy think tank, the world is on track to install 29 percent more solar energy capacity this year — a total of 593 gigawatts — compared to last year, which was already a record year. This is more than one-quarter of the electricity produced by every operating coal plant in the world combined. In 2020, the whole world had installed just 760 GW of solar in total. Yes, this deserves all the italics I’m using.  That solar power installations are going up as the technology improves and prices come down isn’t too surprising, but the sustained surge is still stunning.  “When you look at the absolute numbers that we’re on track for this year and that we installed last year, it is completely sort of mind-blowing,” said Euan Graham, lead author of the report and an electricity data analyst at Ember.  Several factors have aligned to push solar power installations so high in recent years, like better hardware, economies of scale, and new, ripe, energy-hungry markets. Right now, solar still just provides around 5.5 percent of the world’s electricity, so there’s enormous room to expand. But solar energy still poses some technical challenges to the power grid, and the world’s ravenous appetite for electrons means that countries are looking for energy wherever they can get it.  So if you’re concerned about climate change, it’s not enough that solar wins; greenhouse gasses must lose.  Why’s everything so sunny for solar?  Solar energy has a lot going for it, particularly photovoltaic panels. They’re modular and they scale up and down easily — there isn’t much difference between a panel that’s one of a dozen on a suburban rooftop and a panel that’s one of thousands in a megawatt-scale power plant spanning acres. They’re mass-produced in factories using well-established processes, namely semiconductor fabrication. That means tiny improvements in cost and performance in individual panels add up to massive advantages in aggregate.  And for solar, gains have been anything but tiny: Solar electricity prices have dropped 89 percent since 2010 while silicon solar panels have surged in efficiency from 15 percent to more than 26 percent over the last 40 years. Solar’s scalability means that curious developers can try it out with less upfront investment before ramping up. Most solar installations use off-the-shelf components, so when a homeowner or a utility does decide to step into the sunlight, they can start making power quickly. “That development time is absolutely minimal compared to something like building a nuclear power station, but also even just a wind farm, which can take five to seven years or so from the initial permitting to first power coming out,” Graham said.  Even if you don’t care about climate change, solar energy has become one of the cheapest, fastest ways to sate your appetite for electrons. Texas, the biggest oil and gas producer in the country, is also the national leader in adding solar power to its grid, surpassing California. But what happens when the sun sets? Solar does have some drawbacks. The sun does sink below the horizon every day, and solar energy’s output varies with weather and the seasons — dipping when it’s cloudy and when the days get shorter. Banking electricity when it’s abundant to use when it’s scarce would resolve this problem, and, well, there’s good news on that front too.  Energy storage technologies like batteries are also getting way better and cheaper. The price of batteries has tanked 97 percent since 1991. Because of better technology, falling costs, and more markets for saving power,  the US is on track to double its grid energy storage capacity compared to last year. More than 10 gigawatts of solar and storage came online in 2023 across the country and that’s likely to double this year. “Energy storage is at an earlier stage [than solar] but we are likely to see rapid expansion in that segment, especially in regions where solar and wind penetration are high already such as California and Texas,” said Steve Piper, director of energy research at S&P Global Commodity Insights, in an email.  ​Combined solar-plus-storage energy projects are already cheaper than new fossil fuel power plants in many parts of the world, and costs are poised to fall further.  Good morning with good news: Solar plus storage or wind plus storage are NOW much cheaper than coal or gas generated electricity in most of the world.By 2027, clean solar or wind PLUS STORAGE are cheaper than dirty fossil fuels in the entire world!https://t.co/ZtyhVOZNFV pic.twitter.com/l9pz28hYhs— John Raymond Hanger  (@johnrhanger) September 15, 2024 Even knowing all this, energy experts keep underestimating the potential of solar. “Forecasters recognize that with regard to solar PV we are in a phase of rapid expansion and adoption,” Piper said. “In a period like this, being off about the rate of expansion by even a little bit will still result in a large forecast error.” The details of solar energy’s expansion are even more surprising Not every country is riding the solar power rocket to the sun just yet. Individual countries have seen peaks and dips in solar installations based on how well their economies are doing and how strong their policy incentives are, like feed-in tariffs, net metering, and tax credits.  In the past couple of years, the global story has really been about China. Add up every solar panel installed in the US in history and you get how much China installed last year alone, almost 60 percent of all new solar installed in the world. The sheer scale of this deployment broke a lot of forecasters’ models.  “No research shop necessarily predicted the pace at which China was going to grow their solar capacity over the last year or so,” said Michelle Davis, head of global solar at Wood Mackenzie, an energy market analysis firm. “Everyone’s been revising them upward in order to correct for the data that’s been coming out of China.”  Photovoltaics are also a key part of China’s export strategy, and last year, China cut wholesale panel prices in half. That in turn has led to a huge surge in exports and knock-on solar power booms in other countries. Pakistan, the fifth-most populous country in the world, imported 13 gigawatts of Chinese solar modules in the first half of this year alone. That’s almost one-third of Pakistan’s total installed electricity to date.  Davis cautioned that imports of solar panels don’t necessarily mean they’ll all be installed, but it’s definitely a sign that solar is growing and its impact may be greater there than in larger or wealthier countries. While the solar energy additions in developing countries may be smaller in absolute numbers, they’re proportionately a larger share of the grid.  “Those developing parts of the world are growing at a more rapid rate on a smaller base,” Davis said. “The big Kahunas in the solar world are China, Europe, and the United States. Those markets are maturing, though, and they’re not growing as fast.” There are some clouds in the sky This can’t keep going on forever, right? Well, again, solar is still in the single digits in the global electricity supply, and it’s often the cheapest, fastest, and easiest way to generate power. That momentum isn’t going to dissipate anytime soon. In the US, the Federal Reserve’s recent interest rate cut means it will likely be even cheaper to get a loan to finance solar power, giving it another boost.   Some challenges have also emerged. If you want to add more solar to the US power grid right now, you need to take a number and get in line. There are hardware limits to how much intermittent power you can add to the aging electricity network, and making the necessary upgrades to accommodate it costs money and takes time. Delays are getting longer: In 2015, a typical energy project waited about three years in an interconnection queue. In 2023, that wait time was almost five years. Getting the permits to build more large-scale solar is also a tedious process. Many countries are facing similar hurdles.  In addition, the US is bolstering its domestic clean energy sector with trade barriers, including tariffs on Chinese solar panels. That may give an advantage to US producers but it raises overall costs and imposes supply chain constraints. The US is also investing $40 million to bring more of the solar energy supply chain within its borders.  And solar power didn’t just fall out of a coconut tree; it exists in the context of a global economy that’s still 80 percent powered by coal, oil, and natural gas. Overall global energy consumption is growing, and not everyone is discerning about where they get their heat and electricity. As a result, fossil fuel demand is also rising, though it may peak before the end of the decade. To meet international climate change targets of limiting warming to less than 3.6 degrees Fahrenheit this century, greenhouse gas emissions have to fall at a much faster clip and effectively zero out by 2050.  Analysts are anticipating that solar energy will help bend that curve. According to Wood Mackenzie, total global solar capacity is going to almost quadruple in the next decade. It’s not certain whether the world will reach its climate goals, but solar will continue to spread as sure as the sun will rise. 

Teamsters General President Sean M. O’Brien speaks during a rally with workers and union members as part of an “Amazon Teamsters Day of Solidarity” in support of the unionization and collective bargaining of Amazon delivery drivers at the Teamsters Local 848 on August 29, 2024, in Long Beach, California. | Patrick T. Fallon/AFP via Getty Images The International Brotherhood of Teamsters, one of the US’s largest and best-known labor unions, has declined to endorse a candidate for president, citing major political divides among its membership, as well as dissatisfaction with each major candidate’s stances on key union priorities.  In a tight presidential race where the vast majority of labor unions have endorsed Vice President Kamala Harris, the Teamsters’ decision not to do so could be seen as a high-profile rebuke, especially since Harris is part of a Democratic administration that has, for the most part, been highly supportive of labor union rights. Whether the decision actively hurts her, however, remains to be seen. The Teamsters represent more than a million workers from many industries, including UPS drivers and workers in construction, healthcare, and sanitation. Their members are located throughout the United States. But Teamsters have a particularly strong base in the Midwest and in swing states that could be decisive in November. Ahead of the decision, the Teamsters polled members on their presidential preferences. In a September telephone poll, 58 percent of Teamsters members supported Republican candidate Donald Trump, and 31 percent said they’d like the union to endorse Harris. In straw polls conducted at town halls prior to President Joe Biden’s decision to exit the race in July, 44 percent of members supported Biden, versus 36 percent for Trump.  While the poll results suggest a general preference for Trump, a number of Teamsters local unions and subgroups have vocally come out in support for Harris. The Teamsters National Black Caucus endorsed Harris in August, not waiting for the national organization’s decision. In the wake of Wednesday’s announcement, local unions in battlegrounds like Michigan, Wisconsin, and Pennsylvania — among others — also endorsed Harris.  Now both camps are claiming to have the Teamsters behind them, Trump because of the poll results, and Harris because of the local endorsements.   Will the Teamsters’ political decision carry weight in November? The lack of endorsement may not ultimately matter much when it comes to how Teamsters will vote. The polls and local endorsements so far have made it clear that most Teamsters have a set preference this election cycle, and it’s not clear that an endorsement from the national union would have done much to change that.  Peverill Squire, a political science professor at the University of Missouri, told Vox that where union endorsement really matters is on the ground, campaigning. “Union endorsements are valuable because they usually come with access to resources, particularly volunteers to knock on doors and work phone banks,” Squire said. “Leadership may be able to persuade some members to vote the way they would like, but the real value is in campaign assistance.” Besides the setback of having fewer volunteers, there’s also some possibility that the non-endorsement could sway non-union voters sympathetic to labor issues.  “There’s the broad symbolism” of the lack of endorsement, Vanderbilt University professor of sociology and political economy Larry Isaac told Vox.  Teamsters national president Sean O’Brien said, “We sought commitments from both Trump and Harris not to interfere in critical union campaigns or core Teamsters industries — and to honor our members’ right to strike — but were unable to secure those pledges.” That statement, along with the larger choice not to issue an endorsement, makes an explicit argument that neither candidate would look out for organized labor.   Prior to Wednesday’s announcement, there were indications that the union might not endorse at all, in a break with other major unions — which wasn’t a total shock given that the Teamsters have often been out of step politically with the rest of the US labor union movement, Isaac said. The United Auto Workers, the American Federation of Teachers, the AFL-CIO, and many smaller unions across industries and across the country have endorsed Harris.  Still, the Teamsters’ announcement may be enough to influence some voters. Americans support unions at higher levels than they have in decades — and some data suggests voters were more likely to support candidates they perceived as being pro-union in the 2022 elections. Through her connection to Biden, Harris does have a good record on labor; Biden stood at the picket line with UAW workers when they were striking last summer, and has generally supported the right to strike and sought to undermine corporate power. But the administration also forced Teamsters members back to work under what union leadership said was a less-than-ideal contract ahead of a railroad strike in 2022. But Harris’s own policies and values distinct from Biden’s aren’t all that well known, though she has supported the Protect the Right to Organize (PRO) Act, which would strengthen protections around workers’ right to organize and bargain collectively. But Teamsters leadership was reportedly unimpressed with what they heard during a meeting with Harris earlier this week.  Trump, on the other hand, appointed labor antagonists to the National Labor Review Board and approved legislation that restricted workers’ right to organize and strengthened workplaces’ ability to break up a union during his presidency. Rather than focusing on pro-labor policy, Trump has spent the 2024 campaign courting the working class with his populist message. Still, Trump too, apparently did not impress Teamsters in his roundtable meeting with members.