The bro-ligarchy blows up

It did not take any particular skill in forecasting to predict, at the end of 2024, that the unprecedented partnership between Donald Trump and Elon Musk would come to a dramatic ending. Both Trump and Musk are independently famous for their erratic leadership styles and abrupt purges of once-close allies, and neither shows any long-term patience for anyone who opposes them. 

Still, when I predicted here in December that the bro-ligarchy would collapse in 2025, I can’t say I had envisioned what transpired over the past few days.

As Musk’s time as a special federal government employee wound down last month, it seemed as if the partnership between Trump and Musk would be preserved. As recently as May 30, Trump said that Musk is "really not leaving" and was "going to be back and forth."

But just five days later, Musk snapped. On Wednesday, he called Trump’s “big, beautiful bill” — a massive spending package that is now working its way through Congress — “a disgusting abomination.” In a series of X posts, Musk complained that the bill was “pork-filled,” and suggested that it would raise the national debt to unsustainable levels.

Musk is just as likely to have been angry about the pork that wasn’t in the bill — namely, an electric vehicle tax credit that significantly benefits Tesla; and a deal for the Federal Aviation Administration to use Musk’s Starlink satellite network for air traffic control, which he had sought to no avail. Axios’ Mark Caputo reported that Musk was also angry that Trump withdrew the nomination of Musk ally Jared Isaacman to lead NASA. (Musk has also suggested he had no issue with the cancellation of the EV tax credit.)

All of that set the stage for a particularly 2025 phenomenon: two extraordinarily powerful billionaires lambasting one another from the social networks that they own. 

“Time to drop the really big bomb,” Musk posted to X on Thursday afternoon. Trump “is in the Epstein files. That is the real reason they have not been made public. Have a nice day, DJT!” 

Musk was apparently referring to files pertaining to an investigation of the financier Jeffrey Epstein, who died by suicide in a federal jail in 2019. It has long been established that Trump knew Epstein socially. In February, Attorney General Pam Bondi released a set of files related to the investigation that it turned out had mostly already been made public. Right-wing influencers, who are obsessed with Epstein and the possibility that other elites are linked to the sex-trafficking allegations against him, were hugely disappointed. 

The FBI has been preparing to release additional files, the New York Times reported today. With his post, Musk suggested without evidence that Trump is behind the hold-up. As the Times notes, simply being mentioned in the many thousands of pages of files doesn’t necessarily mean anything substantive. But it could nonetheless be used as fuel for conspiracy theories and world-building. 

Trump, for his part, seemed to shrug off Musk’s allegations in a series of posts on Truth Social. (The social network, being accustomed to receiving significant traffic, briefly went down on Thursday afternoon.) He wrote:  “Elon was ‘wearing thin,’ I asked him to leave, I took away his EV Mandate that forced everyone to buy Electric Cars that nobody else wanted (that he knew for months I was going to do!), and he just went CRAZY!”

Trump then suggested he would seek to cancel Musk’s contracts with the government. 

“The easiest way to save money in our Budget, Billions and Billions of Dollars, is to terminate Elon’s Governmental Subsidies and Contracts,” he wrote. “I was always surprised that Biden didn’t do it!”

In the immediate aftermath of this blow-up, it’s tempting to focus only on what both men stand to lose. (And we’ll get there.) But I think it’s just as important to remember what Trump and Musk got from this partnership. Trump got an unprecedented $290 million in donations from Musk, along with constant promotion of his candidacy on X (and constant criticism of opponent Kamala Harris.) And once he was elected, Musk served as a kind of heat shield for Trump in Washington, absorbing huge amounts of criticism that might otherwise have been leveled solely at the president as he and DOGE decimated federal agencies.

Meanwhile, when Musk joined the Trump administration, 11 federal agencies had “more than 32 continuing investigations, pending complaints or enforcement actions into Mr. Musk’s six companies,” according to the Times. The Musk-led assault on the federal bureaucracy dramatically weakened those agencies, throwing the future of many of those investigations into doubt. Meanwhile, Musk set up Starlink in the White House at the same time Trump ambassadors pushed Starlink in Africa. He used his proximity to Trump to boost the valuation of X, which he absorbed into xAI at an inflated valuation, and introduced Grok AI to the federal government. He attempted to block an OpenAI deal with the United Arab Emirates unless that same deal would benefit xAI. 

A report by Senate Democrats found that Musk’s cost-cutting measures protected him from up to $2.37 billion in legal liability from US agencies.

This is why I could never take seriously reports that Musk planned to step back from politics and become “super focused” on his companies. He was super focused on his companies the whole time. 

In any case, Musk has now called for Trump’s impeachment, said that Trump’s tariffs will cause a recession, and posted a poll wondering if he should start a new political party. After Trump suggested he might cut Musk’s federal contracts, Musk said SpaceX would “begin decommissioning its Dragon spacecraft immediately.” Dragon is currently the only means to take US astronauts to and from the International Space Station.

It seems certain all of this will continue to evolve at a rapid clip over the next several days. As dramatic as today’s outbursts seem, it also does not seem out of the question that Trump and Musk will eventually mend fences. 

But that’s not going to happen today — and maybe not for a long time. With that in mind, let’s look at some potential consequences for key players.

What does it mean for Musk?

In the short term: more trouble for those businesses he’s super focused on. Tesla stock declined 14 percent on Thursday, erasing $150 billion in market value. Just as worryingly, from Musk’s perspective, may be that all those federal investigations that he seemed to have effectively sidelined may resume with new vigor. One of Musk’s lobbying priorities has been to pave the way for Teslas to be used as self-driving “robotaxis,” even though they lack LIDAR sensors and other safety features found on Waymos and other autonomous vehicles. He’s going to have a harder time getting that one through Congress now.

Things could get much worse. Steve Bannon, a former top aide to the president who has long detested the Tesla CEO, told the Times that he is advising Trump to deport Musk. He also said the federal government should investigate recent reports of Musk’s drug use.

I wrote in October that billionaires should resist cozying up to authoritarians: while it can feel as if they are reducing risks to themselves and their businesses, surrendering to a corrupt system is just as likely to endanger them. Musk is the first to learn this lesson, but I suspect he will not be the last. 

What does it mean for Trump? 

As president of the United States, Trump has less to lose. He’s already won the election, and has much less need for Musk’s money. (Though Musk did promise to make a $100 million payment to his political action committee in March; Trump will now need to find that money elsewhere. Time to launch a new memecoin!)

To the extent that Musk absorbed some criticism that might otherwise have been directed at Trump, the president will now have to bear the brunt of that himself. He may also need a new way to get astronauts into space, should Musk follow through with his threats about SpaceX and Dragon.

It’s possible Musk’s allies will side with the CEO over the president, reducing Trump’s support in Silicon Valley. But I suspect most of them will remain fairly quiet: JD Vance is their guy, not Trump. If they can grin and bear it until the end of Trump’s term, they will. 

One question that I suspect is being debated in the White House right now: what to do about DOGE? That operation was staffed by Musk’s people; the loyalty-obsessed Trump may now view them as an insider threat. Their sudden exit could introduce yet more chaos into the federal bureaucracy.

What does it mean for AI?

Had they stayed close, I envisioned a world in which someday Trump might entrust Musk with oversight of powerful AI systems. It was so easy to imagine Musk convincing the president that AI had suddenly become too dangerous, and needed to be nationalized under his control. 

That possibility appears to have just vaporized. Meanwhile, OpenAI has managed to develop a seemingly chummy relationship with the president. For that reason, I’d say Thursday is the best day Sam Altman has had in quite some time. 

Let's do a mailbag!

Lately I've been feeling like I don't spend enough time talking to you all about what you're interested in. What questions do you have about tech and democracy? What do you want to know a little more about? What's dominating your group chat?

Email casey@platformer.news and put Mailbag in the subject line, and if we get enough good ones I'll write up some answers sometime soon.

Sponsored

Put an End to Those Pesky Spam Calls

There are few things more frustrating than dashing across the room to answer your ringing phone, only to see "Potential Spam" on the caller ID (probably for the third time today). If you want to cleanse your phone of this annoyance (and increase your personal security), you have three options:

1. Throw your phone into the ocean

2. Individually block each unknown caller

3. Stop spammers from getting your number in the first place with Incogni

We highly recommend option 3, and not just because electronic garbage is bad for aquatic life. Incogni’s automated personal information removal service hunts down your breached personal information, then removes it from the web. Plus, Incogni will reduce the number of spam emails in your inbox.  

Platformer.news readers can get 55% off an annual plan using code PLATFORMER – Get started with Incogni right here.

On the podcast this week: Kevin and I break down the Musk-Trump blow-up. Then, former DOGE employee Sahil Lavingia joins us to discuss his 55 days working for the government. And finally, the Times' Pete Wells joins us to discuss how top chefs have begun to use AI.

Apple | Spotify | Stitcher | Amazon | Google | YouTube

Frontiers in content moderation

• OpenAI said it had discovered 10 operations misusing its services in the past three months, four of which were likely based in China. Among other things, they used ChatGPT to write comments for social media intended to sow division in the United States. (Shannon Bond / NPR)
• A profile of AI-powered “kiss and hug” apps explores how they skirt the App Store’s content moderation guidelines while letting users transform any photo into videos of the subject kissing people, wearing skimpy clothing, and other borderline content. (Victoria Song / The Verge) 
• The Oversight Board found that Meta routinely ignores reports that celebrity likenesses are being misused to promote scams for fear of accidentally removing a genuine endorsement. The company also prohibits its content moderators from taking direct action on these posts for the same reason. No wonder Jamie Lee Curtis had such a rough time recently. Embarrassing. (Karissa Bell / Engadget) 

The chatbots are hallucinating

• Amanda Guinzburg writes on her Substack about a common frustration with chatbots: they keep making things up. This feels like it should be easy to solve at the product level: stop the bot from telling people it has read links sent to them when the bot has not.  
• Not all bots are equal when it comes to hallucinations, though. The Washington Post gave a 115-question test to ChatGPT, Claude, Copilot, Meta AI and Gemini. It found that all but one — Claude — hallucinated at least sometimes. Meta’s Llama performed the worst. (Geoffrey Fowler / Washington Post) 
• One reason bots hallucinate is that they lack needed context to answer questions accurately. So it’s noteworthy that as of this week, ChatGPT can connect to your Google Drive and Dropbox files, among other online storage services. That should help. (Hayden Field / The Verge)

Governing

• Apple’s rollout of AI services in China in partnership with Alibaba is reportedly being held up by a regulator, citing geopolitical uncertainties between China and the US. (Zijing Wu, Cheng Leng and Michael Acton / Financial Times)
• About 1,000 people have left CISA during the second Trump administration, downsizing the agency’s workforce by almost a third. (Sam Sabin / Axios)
• The Trump administration plans to reorganize the Biden era AI Safety Institute into the new Center for AI Standards and Innovation. (Thomas Brewster / Forbes)
• Joshua Kushner’s Thrive Capital and investment firm Capital Group reportedly recently visited China to meet with companies and funds and to discuss AI. (Lulu Yilun Chen / Bloomberg)
• A hacking group, the Com, has breached the networks of at least 20 companies in the US and Europe by impersonating IT personnel to break into the companies’ Salesforce tools, Google said. (Margi Murphy / Bloomberg)
• Meta’s push into defense tech shows how the “tides have turned” in Silicon Valley, CTO Andrew Bosworth said, claiming there's a “silent majority” in the industry wanting to pursue more defense projects. (Riley Griffin / Bloomberg)
• Anthropic released a new set of AI models, Claude Gov, which it says is tailored for US national security customers. (Kyle Wiggers / TechCrunch)
• Anthropic CEO Dario Amodei makes a case for a national transparency standard for AI frontier developers. (Dario Amodei / New York Times)
• Future hackers who can “vibe-hack” could launch multiple zero-day attacks simultaneously using AI, making cyber attacks more severe and accessible. (Matthew Gault / Wired)
• Two FCC members, Democrat Geoffrey Starks and Republican Nathan Simington, said they are leaving the agency this week, leaving three vacancies while the agency reviews a major media merger. (Kelcee Griffis / Bloomberg)
• OpenAI is fighting a court order instructing it to preserve all ChatGPT user logs, including deleted and sensitive chats through its API business offering, in the copyright lawsuit against it brought by news organizations. (Ashley Belanger / Ars Technica)
• Reddit is suing Anthropic for using its data without a licensing agreement. (Meghan Bobrowsky / Wall Street Journal)
• Apple has given governments around the world data related to thousands of push notifications sent to devices, which can identify a person’s device or include unencrypted content. (Joseph Cox / 404 Media)
• TikTok has banned the hashtag “SkinnyTok” globally following pressure from the French government over concerns it was used to promote unhealthy eating. (Théophane Hartmann / Euractiv)
• Microsoft is offering European governments a cybersecurity program for free to boost its defences against cyber attacks. (Supantha Mukherjee / Reuters)
• China’s Hainan province is piloting a program that would grant some corporate users access to the global internet — that is, outside the Great Firewall — as part of its efforts to become a global free-trade port. (Ben Jiang / South China Morning Post)
• A look at how misinformation ran rampant in India during a crucial moment amid its conflict with Pakistan. (Karishma Mehrotra / Washington Post)

Industry

• Nintendo has sold out of the Switch 2 console at launch. (David Keohane / Financial Times)
• A film portraying the OpenAI drama of Sam Altman’s chaotic firing and rehiring within five days is reportedly in development at Amazon MGM Studios, titled “Artificial.” (Lauren Forristal / TechCrunch)
• X updated its developer agreement to say developers can’t use content from X or its API to train AI. (Emma Roth / The Verge)
• Meta has reportedly offered Disney, A24, and other smaller production companies millions of dollars to create exclusive content for a premium VR headset set to release next year. (Ben Fritz and Jessica Toonkel / Wall Street Journal)
• Meta’s Aria Gen 2, its experimental smart glasses, has several improvements, including a better eye-tracking system. (Emma Roth / The Verge)
• An interview with Google DeepMind CEO Demis Hassabis on developing AGI, the upcoming systems that are as smart as humans, and competing with China. (Steven Levy / Wired)
• Microsoft’s LinkedIn chief Ryan Roslansky will now also lead the teams building email and productivity apps as the company looks to speed up deployment of AI tools. (Matt Day / Bloomberg)
• Amazon formed a new team within its consumer product development arm, which will focus on agentic AI, the company said. (Annie Palmer / CNBC)
• Amazon is reportedly developing software for humanoid robots that could eventually replace human delivery workers. (Rocket Drew / The Information)
• Anthropic’s Claude Explains blog is generated mostly by Claude AI with oversight from human editorial teams, the company said. (Kyle Wiggers / TechCrunch)
• Windsurf, the vibe-coding startup reportedly being acquired by OpenAI, said Anthropic significantly restricted its first-party access to the Claude 3.7 Sonnet and 2.5 Sonnet AI models. (Maxwell Zeff / TechCrunch)
  • “I think it would be odd for us to be selling Claude to OpenAI,” Anthropic co-founder and chief science officer Jared Kaplan said. (Maxwell Zeff / TechCrunch)
  • Code generation startups are finding sky-high valuations as more companies look to use AI to help and replace human software engineers. (Anna Tong and Krystal Hu / Reuters)
• French AI startup Mistral is releasing its vibe coding client, Mistral Code. (Kyle Wiggers / TechCrunch)
• AI startup Manus unveiled its text-to-video generation feature using its AI agent, in a bid to take on OpenAI and other players. (Saritha Rai / Bloomberg)
• AMC Networks partnered with Runway AI to generate marketing images and pre-visualize shows. (Steven Zeitchik / Hollywood Reporter)
  • A look at how Hollywood is already using AI behind the scenes. (Lila Shapiro / Vulture)
• Morgan Stanley’s in-house tool built on OpenAI’s GPT models, DevGen.AI, has saved developers 280,000 hours and reviewed nine million lines of code this year, an exec said. (Isabelle Bousquette / Wall Street Journal)
• Reddit is rolling out an update that will let users decide which content from subreddits they participate in will appear on their profiles, including posting and commenting history. (Sarah Perez / TechCrunch)
• Figma is launching a new Dev Mode MCP server that will give LLMs and agentic coding tools more context on how designs are created, helping AI translate designs into applications. (Jess Weatherbed / The Verge)
• A Q&A with Gaia Marcus, director at the Ada Lovelace Institute (which researches AI and its power), on why there’s an urgent need to think about the kind of society being built in the age of AI. (Melissa Heikkilä / Financial Times)
• A look at the perspectives of a growing number of creatives and academics rejecting the use of AI, citing a lack of authenticity and creativity. (Emine Saner / The Guardian)

Those good posts

Today, for reasons I hope you will understand, we are temporarily relaxing our rule against including X posts.

(Link)

Talk to us

Send us tips, comments, questions, and your predictions for what comes next: casey@platformer.news. Read our ethics policy here.