Good morning. How about that SpaceX IPO, eh? Word on the street here in Los Angeles is that all those SpaceX employees are going to push up the already-high price of real estate in the region.
Despite the company’s official relocation to Texas, about a third of SpaceX’s 22,000-strong workforce remains in Hawthorne, a small city 12 miles south of central L.A. Many of them own shares in the company. Many of them are part of the more than 4,000 that became paper millionaires when SPCX closed at about $161 on Friday.
Where to put one’s winnings? L.A. real estate is rarely a bad bet, but based on my conversations with aerospace execs in the area, there’s broad hope that some of the more entrepreneurial SpaceX vets take their newfound riches and reinvest it in new ventures in “The Gundo,” “Space Beach,” and beyond. Fingers crossed.
Today’s tech news below. —Andrew Nusca
P.S. Speaking of L.A. chatter, how electrifying was that USA-Paraguay World Cup game? I could watch those Balogun goals on repeat for hours.
Want to send thoughts or suggestions to Fortune Tech? Drop a line here.
Anthropic rushes to put out its latest fire in Washington
Somewhere, the Amodei siblings are surely thinking: If only we were able to build our AI company without so much politics.
Anthropic has reportedly sent a squad of staffers to Washington in an attempt to smooth over the administration’s abrupt move to ban foreign governments, companies, and individuals from using Anthropic’s new Fable 5 and Mythos 5 AI models.
Senior Anthropic executives and U.S. officials including Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross “spent several hours on calls Saturday” discussing the models, according to the Wall Street Journal.
Both sides reportedly seemed keen to resolve the issue, but it’s not yet clear what a compromise might entail. Access to both models—Fable is the “safe for general use” version of Mythos—remains suspended.
It’s not the first clash between the two entities. Months ago, the U.S. military labeled Anthropic a “supply chain risk” after it refused to honor a request to lift safety guardrails.
What’s odd about this latest clash is that Anthropic coordinated with the White House on the deployment of Mythos. But according to Business Insider, soon after the release of the models, “top administration officials developed fresh doubts that the AI's guardrails were as secure as the company had suggested.”
Those doubts came courtesy of Amazon researchers, who showed that they were able to evade some of Fable’s safeguards.
“We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” Anthropic wrote. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.” —AN
Apple’s Siri AI: ‘Just good enough’
By now, there isn’t a soul in Silicon Valley who doesn’t know that Apple has been working furiously to catch up to the industry on AI.
Now, the question: Do its new Siri AI and Apple Intelligence features stack up?
According to longtime Apple watcher Mark Gurman, the features are “far from groundbreaking” but “good enough to start pulling the company out” of the hole it’s dug for itself.
If you missed it, Apple’s new features include web-informed answers, additional systemwide actions, and what the company calls “onscreen awareness” and “personal context understanding.” (Much of this was first announced in 2024; it’s due for wide release this fall.)
Gurman says those last two in particular “sped up workflows” across devices for things like moving a calendar appointment, finding a particular email, and retrieving a person’s recommendations originally made via email or text message.
It also seems like the new Siri is connecting the dots a lot better than the old one. “The new Siri can handle many of the tasks an everyday consumer would throw at OpenAI’s ChatGPT, Google Gemini or Anthropic’s Claude,” Gurman writes—essential since that style of interaction is increasingly the default for users.
Issues remain, including sluggishness, dropped queries, and a lack of understanding of what’s on the user’s screen. But it’s still a beta release after all.
“Based on my usage of several AI tools over the past three years,” Gurman writes, “the new Siri is roughly competitive with where the leading chatbots were about six months ago.” You know what? We’ll take it. —AN
Alibaba may buy Pupu for $1.5 billion
Alibaba has reportedly offered $1.5 billion to buy Pupu, the Fujian fresh grocery delivery company.
The sum, as reported by Bloomberg, is more than double the bid from PE-backed Sun Art Retail and is part of a broader move to take market share from rival Meituan, which in February announced its $717 million acquisition of fresh grocer Dingdong’s China business.
But don’t let the fierce competition in the fresh grocery category distract you—this is just one battle in a broader war between China’s Big 3 e-commerce kings.
For years, Alibaba, Meituan, and JD.com have been locked in a destructive price war, driven by subsidies, that grew so economically damaging that the Chinese government stepped in.
Since then, the companies have turned toward consolidation as a strategy for growth, putting a different kind of pressure—antitrust—on Beijing.
Meituan’s Dingdong deal is pending approval by regulators; expect the same for the possible sale of privately held Pupu, one of the last major independent grocers standing. Last year Pupu raked in more than RMB 30 billion (approx. $4.43 billion) via a 10-city network in China. —AN
More tech
—Satya Nadella: “The real opportunity [in AI] is not in picking the best model but instead in building a learning loop on top of models where human capital and token capital compound.”
—Health wearables are having a moment, with Oura and Whoop on track for IPOs.
—Sundar Pichai tells Stanford grads: “We don't get to choose the world we graduate into, but we do get to choose how we frame our circumstances.”
—ByteDance may buy AI chips from Shanghai’s Iluvatar CoreX and/or Beijing’s Baidu.
—Frank founder Charlie Javice, convicted of defrauding JPMorgan Chase, seeks a pardon from President Trump.
—Chinese drivers are using DIY workarounds to foil Tesla’s distracted-driving controls.
—AI was the reason behind about 88,000 of the almost 400,000 U.S. job cuts that have taken place this year so far.
