Big Tech keeps promising artificial intelligence will make life better for everyone.
But the people building it keep quitting, and some sound downright spooked on the way out.
And one former OpenAI safety official just told a New York Times host what his old company really built.
The Safety Man Who Walked Out the Door
David Robinson used to sit on OpenAI’s Safety Systems team. He also served as the company’s head of safety transparency, which sounds an awful lot like the guy whose whole job was leveling with the public about what goes on inside the building.
He doesn’t work there anymore.
Robinson resigned and then wrote an op-ed for The Atlantic calling the culture across the AI industry “broken.” His charge was simple enough. The companies racing to build this technology care more about speed than about the safety precautions that ought to come first.
Then he sat down with Ezra Klein on The Ezra Klein Show, and things got a whole lot stranger.
“I believe in humanity. I think we are going to survive. I think there are a lot of things that could happen with this technology that could be very, very harmful,” Robinson told Klein.
That’s a peculiar way to reassure anybody. Most folks don’t open a conversation about a product that helps people write emails by promising the human race will make it through.
Klein Asked What It Is, and Robinson Didn’t Flinch
Klein pointed out that the tech crowd can’t even agree on what AI is. Some industry leaders call it plain old software. Others reach for something much weirder.
So Klein put the question to him straight. “So what is it, man?” he asked.
“An alien mind,” Robinson said.
And he didn’t soften it one bit. He kept right on going.
“It is a thing we grew. We did not engineer it. We engineered the systems around it that grew and then try to keep it safe. We grew a mind. Fundamentally, nobody knows why pre-training works, which is the big hard part where we take lots of inputs and create this basically intelligent thing,” Robinson said.
Read that one twice. A man who ran safety transparency at one of the most powerful AI companies on the planet admits nobody knows why the core process works. The engineers built the greenhouse, something sprouted, and now they bolt extra pieces onto it after the fact to make it more useful and, they hope, safe.
But Robinson isn’t some lone voice crying in the wilderness. Anthropic CEO Dario Amodei recently said the industry “must slow down the pace” of AI. He warned that rogue AI agents could take over the entire internet within a year unless the industry gets the right safety measures and oversight in place.
And an Anthropic researcher named Jacob Coxon resigned with a warning that AI could “kill us all” within a decade.
That’s three insiders. Three people who actually understand the machinery, all ringing the same alarm bell in different keys.
The Pink Slip Comes Before the Robot Uprising
Here’s the part the cheerleaders in Silicon Valley like to skip. The doomsday talk grabs the headlines, but most Americans will feel this thing in their wallets long before they ever worry about a rogue program running the internet.
Amodei himself has warned that AI could wipe out half of all entry-level white-collar jobs and push unemployment somewhere between 10% and 20% within one to five years. The man running one of the companies building this stuff made that prediction. Nobody had to drag it out of him.
Think about a kid fresh out of college with a business degree and a mountain of student loans. Or a paralegal in her fifties who has spent 25 years learning her trade. Or the grandkids, who were told to stay in school, get a desk job, and stay safe.
But the same Big Tech outfits asking Americans for patience keep putting up giant data centers in rural red counties. Residents and critics say those facilities gobble up power and water while pushing electricity bills higher for families who never asked for any of it.
And let’s not forget who these companies are. They throttled, demonetized, and banned conservatives for questioning COVID policy, raising concerns about the 2020 election, and talking about January 6. Now they want ordinary Americans to trust them with an alien mind they admit they can’t fully explain.
That’s a big ask from people with that track record.
Who Holds the Leash on This Thing?
Robinson says the industry grew something it doesn’t understand. Fine. The next question practically asks itself. Who gets to control it?
Because the Left has rarely met a tool of control it didn’t like. During COVID, Democrat governors restricted church services, closed small businesses, and told Americans how many relatives they could invite over for Thanksgiving. And they did all that without a superintelligent machine helping them.
Hand that same crowd an intelligence nobody fully understands, and it doesn’t take much imagination to see where things could head. Automated censorship that flags a conservative post before a single human reads it. A social-credit-style scoring system that docks people for eating red meat, driving a gas-guzzling pickup, or flying the flag a little too proudly.
Maybe that sounds like a stretch. Maybe it is, a little. But plenty of Americans once would have called it a stretch to imagine federal officials leaning on social media companies to bury posts, and then the Twitter Files came out.
In the next COVID-style emergency, a government with this kind of technology wouldn’t need to send anyone door to door. The software would already know who skipped the mask and who showed up at church anyway.
And there’s the national security angle nobody in the industry wants to dwell on. A technology its own builders can’t explain is exactly the sort of thing Communist China would love to steal, copy, or point back at the United States.
Robinson says he believes in humanity. Good for him.
But the folks who built this thing keep heading for the exits, and each one sounds a little more rattled than the last. Ordinary Americans have every right to ask why, and to demand some real answers before Big Tech plugs an alien mind into their jobs, their power bills, and their everyday lives.
Sources: Mediaite, “OpenAI Insider Tells NYT’s Ezra Klein They’ve Created ‘Alien Mind'”; The Ezra Klein Show; The Atlantic; Axios interview with Dario Amodei on AI and entry-level jobs