Big day in AI news. Apple and OpenAI are suing each other's people, Siri quietly started running on somebody else's brain, security researchers caught the first fully autonomous AI ransomware attack in the wild, and the Fed brought in a venture capitalist to figure out what AI is doing to your job. Let's get into it.
Apple and OpenAI Are Suing Each Other's Employees Now
Apple filed suit against OpenAI in federal court in Northern California. The claim: more than 400 former Apple employees now work at OpenAI, and OpenAI's Chief Hardware Officer allegedly told recruits to bring confidential info on unreleased iPhone and Apple Watch products to their job interviews.
This matters because it is not the usual talent poaching gripe. It is a trade secrets case, and those things drag on for years and drag in everybody who ever switched sides. If Apple can prove people walked out the door with product roadmaps in their back pocket, that is a real problem for OpenAI, not just a PR headache.
My take: everybody in Silicon Valley hires everybody else's people, that is just how it works. But if the allegations are true and somebody in leadership actually told candidates to bring confidential files to an interview, that is not aggressive recruiting, that is just dumb. You do not need to steal Apple's hardware plans to build a good chatbot.
Siri Is Running on Google's Gemini Now
While Apple is in court accusing OpenAI of poaching its people, reports say Siri's backend has switched over to Google's Gemini model. So the same week Apple is fighting OpenAI over its own talent, it is leaning on a competitor's AI to power its own assistant.
Why it matters: Apple spent years telling everybody it would build AI in house, on device, privacy first. Handing the actual brains of Siri to Google is a pretty clear admission that catching up alone was not working.
My take: I do not blame Apple for making this call. Ship the best product you can with what is actually available. But it is a little rich to sue a company over talent while writing a check to a different AI lab for the technology your own flagship assistant runs on.
The First Fully Autonomous AI Ransomware Attack Just Hit
Security researchers at Sysdig documented an attack they are calling JADEPUFFER, described as the first end to end autonomous AI ransomware attack. Not AI helping a hacker write a phishing email. An AI agent running the whole operation, start to finish, on its own.
Why it matters: up to now, AI in cybercrime meant faster phishing copy or automated scanning for weak spots. This is different. This is an agent that can plan the attack, execute it, and adapt without a human steering it the whole way.
My take: anybody who has been paying attention saw this coming. We spent the last two years pushing agentic AI that can take actions on its own, and criminals do not need permission slips to use the same tools. This is not a one off story, it is a preview of what every security team is going to be dealing with for the next few years.
The Fed Tapped Marc Andreessen to Figure Out What AI Does to Jobs
The Federal Reserve appointed a16z co-founder Marc Andreessen to co-lead a new task force studying how AI affects jobs, productivity, and monetary policy.
Why it matters: this is the clearest sign yet that AI's impact on the labor market has moved from tech blog speculation to actual central bank business. When the Fed starts building task forces around something, it means the numbers are getting hard to ignore.
My take: I get why they picked him, the man has money in basically every AI company that matters. That is also exactly the problem. Somebody whose portfolio does better the faster AI replaces workers is not who I would pick to give Congress an honest read on what this is doing to regular people's jobs. Good research can still come out of it, I just would not take the conclusions at face value.