She refused to give up her password, and was, according to the prosecutors’ description, so demoralized by the earlier hacking theft of the majority of her funds that she told the men to simply shoot her. Instead, they stole her engagement ring, two iPhones, a laptop, the charger for the neurostimulator used by the other member of the household as a treatment for Parkinson’s disease, and whatever cash they could find, then left.

For their next victim, the prosecutors describe how the group targeted a man who Seemungal knew to be a fellow SIM swapping hacker and who he believed had in fact robbed him of a significant sum of cryptocurrency in 2021. To prepare for that robbery in September of 2022, they began repeatedly sending their target pizza deliveries in the hope of conditioning him to come to his door without suspicion. When the moment of their planned theft came, however, their target wasn’t home, so they instead lay in wait, then drew guns on their target when he arrived at the house.

Over the next hour, the group bound their victim’s hands behind his back with bootlaces and demanded he hand over access to his crypto accounts. When the account he gave them access to had only a small sum of crypto, they put him in the backseat of their rented Cadillac, struck his face with their guns, drove away, and began extorting his friends and father for crypto payments. Eventually, about 120 miles from their victim’s home, the men took their victim out of the car and told him to kneel. He instead escaped, as one of the men fired a gun from the moving car, though it’s not clear if the shot was intended to hit the victim or merely scare him. One of the group—who has not yet been charged—would later say that St. Felix had suggested they kill their captive.

A few months later, prosecutors write, the group carried out their next attack against another victim they believed to be a wealthy crypto-focused hacker, this time in Texas. On a road trip from Florida to start surveilling their target, St. Felix had fled from law enforcement in Louisiana, flipped his car at more than 90 miles per hour, and broken his leg. The other members of the Florida crew had been arrested after that crash. So the break-in was carried out by a newly recruited team based in the Houston area.

Just a few days before Christmas of 2022, the Texas group broke into their target’s home, bound his family members’ hands with zip ties, and repeatedly hit him in the face demanding he give them access to his cryptocurrency. Prosecutors say they shoved knives and forks under his mother’s fingernails and struck her in the face with a gun. They burned their target’s arm with a hot iron to coerce him to hand over his crypto account details, and at one point attempted to punch him in the genitals.

The victim eventually told his torturers that he had buried a device storing his cryptocurrency in the backyard. (In fact, that hardware wallet, holding $1.4 million in crypto, was in a moving box in the home that the thieves never found.) When the thieves brought their victim to the backyard to locate the device, he climbed a fence and escaped. The burglars stole $150,000 in cash as well as some jewelry, then left.

One Final Job

In early 2023, after those relatively unsuccessful attempts at extortion, an associate of Seemungal’s allegedly began feeding the group tips, hacking into potential targets’ email to see the size of their crypto holdings, and sending those leads to the home invasion crew. One Telegram chat obtained by prosecutors shows a discussion of potential targets, including someone with $1.2 million in Texas and another person with $600,000 in Tennessee.

“Then on the other extreme, [the left-wing New Popular Front] have been so vocal about all the taxation measures they want to bring back that it looks like we’re just going back to pre-Macron period,” Varza says. She points to France’s 2012 “les pigeons” (or “suckers”) movement, a campaign by angry internet entrepreneurs that opposed Socialist president François Hollande’s plan to dramatically raise taxes for founders.

Maya Noël, CEO of France Digitale, an industry group for startups, is worried not only about France’s ability to attract overseas talent, but also about how appealing the next government will be to foreign investors. In February, Google said it would open a new AI hub in Paris, where 300 researchers and engineers would be based. Three months later, Microsoft also announced a record $4 billion investment in its French AI infrastructure. Meta has had an AI research lab in Paris since 2015. Today France is attractive to foreign investors, she says. “And we need them.” Neither Google nor Meta replied to WIRED’s request for comment. Microsoft declined to comment.

The vote will not unseat Macron himself—the presidential election is not scheduled until 2027—but the election outcome could dramatically reshape the lower house of the French Parliament, the National Assembly, and install a prime minister from either the far-right or left-wing coalition. This would plunge the government into uncertainty, raising the risk of gridlock. In the past 60 years, there have been only three occasions when a president has been forced to govern with a prime minister from the opposition party, an arrangement known in France as “cohabitation.”

No AI startup has benefited more from the Macron era than Mistral, which counts Cédric O, former digital minister within Macron’s government, among its cofounders. Mistral has not commented publicly on the choice France faces at the polls. The closest the company has come to sharing its views is Cédric O’s decision to repost an X post by entrepreneur Gilles Babinet last week that said: “I hate the far-right but the left’s economic program is surreal.” When WIRED asked Mistral about the retweet, the company said O was not a spokesperson, and declined to comment.

Babinet, a member of the government’s artificial intelligence committee, says he has already heard colleagues considering leaving France. “A few of the coders I know from Senegal, from Morocco, are already planning their next move,” he says, claiming people have also approached him for help renewing their visas early in case this becomes more difficult under a far-right government.

While other industries have been quietly rushing to support the far-right as a preferable alternative to the left-wing alliance, according to reports, Babinet plays down the threat from the New Popular Front. “It’s clear they come with very old-fashioned economical rules, and therefore they don’t understand at all the new economy,” he says. But after speaking to New Popular Front members, he says the hard-left are a minority in the alliance. “Most of these people are Social Democrats, and therefore they know from experience that when François Hollande came into power, he tried to increase the taxes on the technology, and it failed miserably.”

Already there is a sense of damage control, as the industry tries to reassure outsiders everything will be fine. Babinet points to other moments of political chaos that industries survived. “At the end of the day, Brexit was not so much of a nightmare for the tech scene in the UK,” he says. The UK is still the preferred place to launch a generative AI startup, according to the Accel report.

Stanislas Polu, an OpenAI alumnus who launched French AI startup Dust last year, agrees the industry has enough momentum to survive any headwinds coming its way. “Some of the outcomes might be a bit gloomy,” he says, adding he expects personal finances to be hit. “It’s always a little bit more complicated to navigate a higher volatility environment. I guess we’re hoping that the more moderate people will govern that country. I think that’s all we can hope for.”

One of the key ingredients that made ChatGPT a ripsnorting success was an army of human trainers who gave the artificial intelligence model behind the bot guidance on what constitutes good and bad outputs. OpenAI now says that adding even more AI into the mix—to help assist human trainers—could help make AI helpers smarter and more reliable.

In developing ChatGPT, OpenAI pioneered the use of reinforcement learning with human feedback, or RLHF. This technique uses input from human testers to fine-tune an AI model so that its output is judged to be more coherent, less objectionable, and more accurate. The ratings the trainers give feed into an algorithm that drives the model’s behavior. The technique has proven crucial both to making chatbots more reliable and useful and preventing them from misbehaving.

“RLHF does work very well, but it has some key limitations,” says Nat McAleese, a researcher at OpenAI involved with the new work. For one thing, human feedback can be inconsistent. For another it can be difficult for even skilled humans to rate extremely complex outputs, such as sophisticated software code. The process can also optimize a model to produce output that seems convincing rather than actually being accurate.

OpenAI developed a new model by fine-tuning its most powerful offering, GPT-4, to assist human trainers tasked with assessing code. The company found that the new model, dubbed CriticGPT, could catch bugs that humans missed, and that human judges found its critiques of code to be better 63 percent of the time. OpenAI will look at extending the approach to areas beyond code in the future.

“We’re starting work to integrate this technique into our RLHF chat stack,” McAleese says. He notes that the approach is imperfect, since CriticGPT can also make mistakes by hallucinating, but he adds that the technique could help make OpenAI’s models as well as tools like ChatGPT more accurate by reducing errors in human training. He adds that it might also prove crucial in helping AI models become much smarter, because it may allow humans to help train an AI that exceeds their own abilities. “And as models continue to get better and better, we suspect that people will need more help,” McAleese says.

The new technique is one of many now being developed to improve large language models and squeeze more abilities out of them. It is also part of an effort to ensure that AI behaves in acceptable ways even as it becomes more capable.

Earlier this month, Anthropic, a rival to OpenAI founded by ex-OpenAI employees, announced a more capable version of its own chatbot, called Claude, thanks to improvements in the model’s training regimen and the data it is fed. Anthropic and OpenAI have both also recently touted new ways of inspecting AI models to understand how they arrive at their output in order to better prevent unwanted behavior such as deception.

The new technique might help OpenAI train increasingly powerful AI models while ensuring their output is more trustworthy and aligned with human values, especially if the company successfully deploys it in more areas than code. OpenAI has said that it is training its next major AI model, and the company is evidently keen to show that it is serious about ensuring that it behaves. This follows the dissolvement of a prominent team dedicated to assessing the long-term risks posed by AI. The team was co-led by Ilya Sutskever, a cofounder of the company and former board member who briefly pushed CEO Sam Altman out of the company before recanting and helping him regain control. Several members of that team have since criticized the company for moving riskily as it rushes to develop and commercialize powerful AI algorithms.

Dylan Hadfield-Menell, a professor at MIT who researches ways to align AI, says the idea of having AI models help train more powerful ones has been kicking around for a while. “This is a pretty natural development,” he says.

Hadfield-Menell notes that the researchers who originally developed techniques used for RLHF discussed related ideas several years ago. He says it remains to be seen how generally applicable and powerful it is. “It might lead to big jumps in individual capabilities, and it might be a stepping stone towards sort of more effective feedback in the long run,” he says.

Two activists disrupted the Amazon Web Services Summit in Washington, DC, on Wednesday to protest Project Nimbus, Amazon and Google’s $1.2 billion cloud computing contract with the Israeli government.

The protest, which interrupted the keynote speech from Dave Levy, AWS worldwide public sector vice president, is the latest in a series of recent protests that have taken aim at Project Nimbus.

The first activist, who appeared to be a young man in a video shared with WIRED, stood on a chair waving a Palestinian flag and demanded an end to Project Nimbus.

“Dave Levy, why is Amazon contracting for a government that every mainstream human rights organization agrees is an apartheid state?” he yelled. “Why is Amazon providing cloud services for a government that is committing a genocide and that is committing the crime of apartheid?”

The man was promptly escorted out by security and two officers with DC’s Metropolitan Police Department. Shortly after, a second activist, who appeared to be a young woman in a video shared with WIRED, stood on a chair while waving a banner reading, “LET GAZA LIVE.”

“Forty-thousand dead, Dave Levy!” she yelled. “You have blood on your hands with the technology that powers the indiscriminate slaughter of Palestinians! You can do tech for good, but your technology is powering genocide! How do you feel knowing that genocide runs on Amazon?”

This activist, too, was promptly escorted out by security.

Human Rights Watch and Amnesty International have both stated that Israel is committing the crime of apartheid. Since Israel began its military campaign on Gaza last fall, more than 39,000 Palestinians, including more than 15,000 children, have died, according to the Palestinian Ministry of Health. Israel’s military campaign followed Hamas’ October 7 attack on southern Israel, which killed more than 1,100 Israelis.

Israel is currently being charged with genocide in the International Court of Justice in a case brought forward by South Africa. In May, the International Criminal Court filed arrest warrants, alleging war crimes, for Israeli prime minister Benjamin Netanyahu, Israeli defense minister Yoav Gallant, Hamas leader Yahya Sinwar, and two other Hamas officials. Israel has repeatedly denied accusations of genocide and other crimes.

Both activists represented No Tech for Apartheid, a coalition formed in 2021 to protest Project Nimbus. The group is made up of tech workers and organizers with Muslim grassroots group MPower Change and anti-Zionist Jewish group Jewish Voices for Peace.

In a statement released by No Tech for Apartheid after the protest, the group said that while they have been protesting Project Nimbus since 2021, for Google and Amazon to continue with the contract “in the midst of this genocide reaches a new level of horror.”

“We’re here to interrupt business as usual until they cut ties,” the statement said.

Amazon did not immediately respond to WIRED’s request for comment.

No Tech for Apartheid has spearheaded several major protests in recent months. In March, group member and then-Google cloud engineer Eddie Hatfield interrupted Google Israel managing director at Mind the Tech, a Google-sponsored conference highlighting the Israeli tech industry. Hatfield was fired days later.

In April, Google employees with the group staged a sit-in protest in company offices in New York and Sunnyvale, California, with simultaneous protests happening outside. In response, nine employees were detained by police and more than 50 employees were fired in two waves of dismissals. Some of the fired workers filed an unfair labor practice charge with the National Labor Relations Board in response, and the case is ongoing.

In recent weeks, as part of another No Tech for Apartheid effort, more than 1,100 college students from more than 120 universities have signed on to a pledge vowing to not work or intern for Google or Amazon until they drop Project Nimbus.

Hello, and welcome to WIRED Travel, WIRED’s new destination for coverage of all things transportation, aviation, vacation, and more. Of course, innovations in travel technology, from next-generation electric vehicles to fuel-sipping, climate-friendly aircraft, are all mainstays of WIRED’s coverage already. But we’re excited to pull it all together in one place and have the opportunity to bring you even more.

WIRED Travel will introduce you to companies moving beyond fossil fuels and people embracing the “slow travel” movement. We’ll bring you up-to-date, practical tips and tricks to get where you’re going, like the best apps for solo travel and the gear you should pack before you go. As climate change and extreme weather challenge our transportation systems, we’ll offer a closer look at how experts are addressing global infrastructure challenges and how transit companies are adapting to a changing world.

If your summer plans include a quick getaway to more pleasant shores, or you just want to get a head start on the best deals for your family holidays at the end of the year, we’ve got you covered. Remember though, we’re on this trip together! If you have questions or feedback, or even want to contribute your voice to the section, email us at [email protected], and mention travel in the subject line. We’d love to hear from you.

In late March, a smoke shop in Dyersburg, Tennessee, announced the arrival of a new product in its store: a disposable nicotine vape with an LCD display that can be connected to a smartphone via Bluetooth. Marketed under the brand name RAMA, the strawberry- and kiwi-flavored vape looks more like a cell phone from the early 2000s than a typical e-cigarette. It allows users to customize the screen background, see how many puffs of nicotine are left, and even track the device’s location using an accompanying app. “NEVER LOSE YOUR VAPE AGAIN!!!!” Mk Smoke Shop said in a Facebook post.

Far from a one-off novelty, the RAMA model is part of a wave of technologically sophisticated and highly potent disposable vapes that have begun appearing on shelves in smoke shops and convenience stores across the United States in recent months, according to industry data, social media posts, and other records viewed by WIRED.

Almost exclusively manufactured in China, the vapes are colorful and come in eye-catching metallic finishes, squishy silicone textures, and rounded shapes that fit comfortably in a person’s hand. But what really sets them apart are LCD screens, which make the devices even more harmful for the environment than normal disposable vapes. And like the vast majority of all e-cigarettes available in stores, they are technically illegal and haven’t been approved for sale by the US Food and Drug Administration.

These so-called smart vapes are the product of an innovation boom taking place in China’s $28 billion e-cigarette export industry. It was spurred, in part, by the United States’ lax enforcement of nicotine regulations. The US accounts for nearly two-thirds of Chinese vape exports, according to the China Electronics Chamber of Commerce. From 2020 to 2023, the CDC foundation estimates that sales of non-tobacco-flavored vapes in the US surged more than 60 percent, increasing from 11.2 million to 18 million units.

As competition for the American market intensified, vape producers in Shenzhen needed to find ways to make their products stand out. So they developed vapes that were more affordable, better designed, and delivered higher doses of nicotine compared to their predecessors. In many cases, these innovations allowed them to move up the value chain for e-cigarettes.

Robert Jackler, an emeritus professor of head and neck surgery at Stanford University and the founder of an interdisciplinary research group studying the impacts of tobacco advertising, said that American companies have long manufactured vapes in Shenzhen. But after the Chinese government banned the sale of flavored vapes in 2022, Chinese suppliers began focusing more on marketing their own products directly to overseas customers.

“They cut out the Americans,” Jackler says. As of last year, the Associated Press reported there were over 9,000 kinds of vaping products available for sale in the US, a nearly threefold increase since 2020.

The proliferation of disposable flavored vapes from China has alarmed lawmakers in both the US and Europe. Regulators say they are especially worried about the impact the devices are having on children, who may find the sweet flavors and flashy designs they come in particularly appealing.

Brussels has accused Microsoft of illegally abusing its dominance in the business-software market at the expense of smaller rivals, following a complaint at the height of the pandemic by US competitor Slack.

The European Commission said on Tuesday it found that Microsoft was restricting competition by selling its video-conferencing software Teams together in bundles with the company’s other popular office tools such as Office 365 and Microsoft 365 since at least 2019.

“We are concerned that Microsoft may be giving its own communication product Teams an undue advantage over competitors, by tying it to its popular productivity suites for businesses,” the EU’s competition chief Margrethe Vestager said in a statement. “If confirmed, Microsoft’s conduct would be illegal under our competition rules.” The charges announced on Tuesday are only a “preliminary view,” meaning the commission has sent a “statement of objections” to Microsoft and the company has 10 weeks once it receives all the details to respond.

The Microsoft charges arrive in the same week as the European Commission also charged Apple with breaking the European Union’s new digital markets act for failing to let app developers communicate freely with their users. Over the past decade, the EU has become the de facto Big Tech regulator, forcing US giants to alter the way they operate and issuing fines of billions of dollars.

In an attempt to placate Brussels, Microsoft started excluding Teams from some Office bundles in July of last year. However, the commission said today that those changes were insufficient and expressed concern about how easy it was to use rival conferencing software in tandem with Microsoft’s other tools, a practice known as interoperability.

“Having unbundled Teams and taken initial interoperability steps, we appreciate the additional clarity provided today,” said Brad Smith, vice chair and president of Microsoft, in a statement shared with WIRED. The company plans to work to find solutions to address the commission‘s remaining concerns, he added.

If Microsoft and the EU cannot reach an agreement, the commission has the power to levy fines of up to 10 percent of the company’s annual worldwide turnover and can impose remedies on the company.

The commission opened its investigation into Microsoft Teams following a complaint by Slack in July 2020, when there was fierce competition for the remote workers who relied on office software due to pandemic lockdowns. “This is much bigger than Slack versus Microsoft,” Jonathan Prince, then vice president of communications and policy at Slack, said at the time. “This a proxy for two very different philosophies for the future of digital ecosystems, gateways versus gatekeepers.”

On Tuesday, Sabastian Niles, president and chief legal officer of Slack’s parent company Salesforce, described the European Commission’s position as “a win for customer choice and an affirmation that Microsoft’s practices with Teams have harmed competition.”

German video conferencing company Alfaview, which filed a complaint to the commission following Slack, also welcomed the decision. The measures Microsoft has taken so far to unbundle Teams have been ineffective, Niko Fostiropoulos, CEO and founder of Alfaview, said in a statement. “Microsoft offers existing enterprise customers who opt out of Teams in the overall package only a minimal discount of €2 ($2.10),” he said. “This does not provide sufficient incentives to switch to another video conferencing service.”

Apple has become the first Big Tech company to be charged with breaking the European Union’s new digital markets rules, three days after the tech giant said it would not release artificial intelligence in the bloc due to regulation.

On Monday, the European Commission said that Apple’s App Store was preventing developers from communicating with their users and promoting offers to them directly, a practice known as anti-steering.

“Our preliminary position is that Apple does not fully allow steering. Steering is key to ensure that app developers are less dependent on gatekeepers’ app stores and for consumers to be aware of better offers,” Margrethe Vestager, the EU’s competition chief, said in a statement.

On X, the European commissioner for the internal market, Thierry Breton, gave a more damning assessment. “For too long Apple has been squeezing out innovative companies—denying consumers new opportunities and choices,” he said.

The EU referred to its Monday charges as “preliminary findings.” Apple now has the opportunity to respond to the charges and, if an agreement is not reached, the bloc has the power to levy fines—which can reach up to 10 percent of the company’s global turnover—before March 2025.

Tensions between Apple and the EU have been rising for months. Brussels opened an investigation into the smartphone maker in March over failure to comply with the bloc’s competition rules. Although investigations were also opened into Meta and Google-parent Alphabet, it is Apple’s relationship with European developers that has long been the focus in Brussels.

Back in March, one of the members of the European Parliament who negotiated the Digital Markets Act told WIRED that Apple was the logical first target for the new rules, describing the company as “low-hanging fruit.” Under the DMA it is illegal for Big Tech companies to preference their own services over those of rivals.

Developers have seethed against the new business terms imposed on them by Apple, describing the company’s policies as “abusive,” “extortion,” and “ludicrously punitive.”

Apple spokesperson Rob Saunders said on Monday he was confident the company was in compliance with the law. “All developers doing business in the EU on the App Store have the opportunity to utilize the capabilities that we have introduced, including the ability to direct app users to the web to complete purchases at a very competitive rate,” he says.

On Friday, Apple said it would not release its artificial intelligence features in the EU this year due to what the company described as “regulatory uncertainties.” “Specifically, we are concerned that the interoperability requirements of the DMA could force us to compromise the integrity of our products in ways that risk user privacy and data security,” said Saunders in a statement. The features affected are iPhone Mirroring, SharePlay Screen Sharing enhancements, and Apple’s first foray into generative AI, Apple Intelligence.

Apple is not the only company to blame new EU rules for its decision to delay the rollout of new features. Last year, Google delayed the EU rollout of its ChatGPT rival Bard, and earlier in June, Meta paused plans to train its AI on Europeans’ personal Facebook and Instagram data following discussions with privacy regulators. “This is a step backward for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe,” the company said at the time.

The music industry has officially declared war on Suno and Udio, two of the most prominent AI music generators. A group of music labels including Universal Music Group, Warner Music Group, and Sony Music Group has filed lawsuits in US federal court on Monday morning alleging copyright infringement on a “massive scale.”

The plaintiffs seek damages up to $150,000 per work infringed. The lawsuit against Suno is filed in Massachusetts, while the case against Udio’s parent company Uncharted Inc. was filed in New York. Suno and Udio did not immediately respond to a request to comment.

“Unlicensed services like Suno and Udio that claim it’s ‘fair’ to copy an artist’s life’s work and exploit it for their own profit without consent or pay set back the promise of genuinely innovative AI for us all,” Recording Industry Association of America chair and CEO Mitch Glazier said in a press release.

The companies have not publicly disclosed what they trained their generators on. Ed Newton-Rex, a former AI executive who now runs the ethical AI nonprofit Fairly Trained, has written extensively about his experiments with Suno and Udio; Newton-Rex found that he could generate music that “bears a striking resemblance to copyright songs.” In the complaints, the music labels state that they were independently able to prompt Suno into producing outputs that “match” copyrighted work from artists ranging from ABBA to Jason Derulo.

One example provided in the lawsuit describes how the labels generated songs extremely similar to Chuck Berry’s 1958 rock hit “Johnny B. Goode” in Suno by using prompts like “1950s rock and roll, rhythm & blues, 12 bar blues, rockabilly, energetic male vocalist, singer guitarist,” along with snippets of the song’s lyrics. One song almost exactly replicated the “Go, Johnny, go” chorus; the plaintiffs attached side-by-side transcriptions of the scores and argued that such overlap was only possible because Suno had trained on copyrighted work.

The Udio lawsuit offers similar examples, noting that the labels were able to generate a dozen outputs resembling Mariah Carey’s perennial hit “All I Want for Christmas Is You.” It also offers a side-by-side comparison of music and lyrics, and notes that Mariah Carey soundalikes generated by Udio have already caught the attention of the public.

RIAA chief legal officer Ken Doroshow says Suno and Udio are trying to conceal “the full scope of their infringement.” According to the complaint against Suno, the AI company did not deny that it used copyrighted materials in its training data when asked in prelitigation correspondence, but instead said that the training data is “confidential business information.”

Many leading generative AI companies are under intense scrutiny for how they train their tools. It’s common for these companies to argue that they are shielded by the “fair use” doctrine, which permits infringement in certain circumstances. It remains to be seen whether the court system will agree; major players like OpenAI are already facing a host of copyright infringement lawsuits from artists, writers, programmers, and other rights holders.