- Anthropic CEO Dario Amodei told the Financial Times there is no ceiling on AI model scaling—”We don’t see anything slowing down.”
- Amodei simultaneously warned that AI could eliminate 50 percent of entry-level office jobs within five years.
- The contradiction defines the moment: the industry’s most capable models are accelerating faster than the workforce can absorb them.
Dario Amodei doesn’t think AI is hitting a wall. “There’s no end to the rainbow. There’s just the rainbow,” told the Anthropic CEO in an interview with the Financial Times. “We don’t see anything slowing down.”
The comments land at a strange moment. Amodei spent the same week at the White House discussing federal AI deployment—and warning Washington that Anthropic’s own Mythos model is too dangerous for public release. Scale everything up, but also, maybe watch out.
Amodei calls it the “big blob of compute”—the idea that raw computational power still has enormous room to grow. He’s not wrong that hardware budgets are climbing. Anthropic reportedly turned down VC offers valuing it at $800 billion. The company raised at $380 billion earlier this year.
The Scaling Debate Nobody’s Winning
The AI scaling debate has divided the industry for two years. On one side, researchers like Amodei argue that larger models trained on more data with more compute keep getting smarter—and there’s no theoretical limit in sight. On the other, skeptics point to diminishing returns on benchmarks and the astronomical cost of each incremental improvement.
Amodei isn’t hedging. He’s told the FT that the industry hasn’t delivered on its optimistic promises yet, but the warnings are already stacking up—including his own prediction that AI could wipe out 50 percent of entry-level office jobs within five years. “Is that just propaganda? Is that just vaporware that’s not going to happen? We actually have to make it happen,” he said.
The framing is notable: Amodei isn’t arguing that job losses won’t happen. He’s arguing the industry needs to build enough upside to offset them. AI, he says, can only “diffuse at the speed of trust.” And trust, right now, is in short supply.
Why Amodei Keeps Saying the Quiet Part Out Loud
Most AI CEOs talk about job creation. Amodei talks about job destruction. His 50 percent entry-level job prediction isn’t new—he’s said it before—but reiterating it alongside claims of limitless scaling creates an unusual rhetorical position. The more powerful AI gets, the worse the disruption. And he wants to scale it further.
The logic is internally consistent, if uncomfortable: if AI is going to transform everything regardless, the responsible move is to make sure the transformation produces enough value to absorb the shock. “The industry can’t afford to downplay the disruption,” Amodei said. “It needs to make the upside big enough to serve as a tool for working through the fallout.”
Whether that’s optimism or a hostage negotiation with the future depends on your priors. What’s not debatable is that Anthropic is still hiring—at $380 billion, the company is scaling headcount nearly as fast as its models.
