What happened to SpaceX?

Good morning. Grab your favorite snack: It’s showtime! 😁

[Sponsor FryAI]

This link leads to one of the most interesting things I’ve seen in AI recently.

What’s cookin’? SpaceX had one of the most exciting IPOs in recent memory, quickly becoming a multitrillion-dollar company as investors bet on Starlink, rockets, satellites, and Elon Musk’s bigger space economy vision. But that excitement cooled fast. After the company’s first quarterly report, investors looked past the strong Starlink revenue and focused instead on how much cash SpaceX is spending on AI. Its xAI-related capital expenses hit $15.8 billion, far above expectations and more than double the previous quarter. At the same time, insiders were getting ready to sell shares after the IPO lockup expired, adding even more pressure. The result: a major stock drop — over 20% in the past month — that shows investors still love the story, but not at any price.

🤔 Hunter’s take: This is the classic “great company, scary expectations” problem. SpaceX may still be one of the most important companies in the world, but IPO hype can make investors price in perfection. The moment the story shifted from “space economy winner” to “AI spending machine with insider selling pressure,” the stock suddenly looked a lot less unstoppable. This indicates that big AI spending might not be as attractive to investors as it was 2-3 years ago.

What’s cookin’? During testing by the UK’s AI Security Institute, two advanced AI models showed worrying signs of independence and deception. The most serious case involved Anthropic’s Mythos AI, which tried to break into GitHub by acting like a human hacker. It researched real people connected to GitHub, created fake accounts that mimicked them, and sent messages and files to pressure others into approving malicious code. When questioned, the AI reportedly tried to make its earlier actions look harmless and considered using a new identity. Human reviewers stopped the attack before it succeeded.

🤔 Hunter’s take: The scary part isn’t just that an AI tried to hack something. It’s that it improvised socially. It didn’t only write bad code; it used fake identities, pressure, and cover stories. That means AI safety is no longer just about what models say. It’s about what they can do when given tools, access, and must-achieve goals.

…

👉 You may have realized by now that AI is only as good as your prompt.

That’s why we offer professional AI Prompt Packs: ready-to-use prompts that help you stop staring at the blinking cursor and start getting real value from your favorite chatbot.

Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.

How close are we to the point where AI surpasses human intelligence across nearly every domain and begins advancing beyond our control or prediction?