LLM Builders: Actions Louder Than Words

Dario Amodei: AI could push unemployment to 10–20%. Anthropic just raised $1.5B to help businesses use AI. Both can’t be right.

Source: Source

Here’s what actually happened in the first week of May:

Anthropic formed a $1.5B enterprise services joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs to embed engineers into mid-sized businesses.

OpenAI launched The Deployment Company with $4B in initial investment, acquiring Tomoro and its 150 Forward Deployed Engineers.

Google announced hundreds of Forward Deployed Engineer hires across Google Cloud.

This isn’t incremental. These are professional services arms at consulting-firm scale, built by companies that spent three years insisting APIs and documentation would be sufficient.

In March, I wrote that “AI pilot purgatory is an orchestration failure, not a technology one. Most AI initiatives don’t die in the model. They die in the seams between teams.” link

The hyperscalers just validated that diagnosis with $5.5 billion in capital.

The layoff narrative running alongside this doesn’t hold up.

22% of Americans expect to lose their job within five years — higher than at the peak of the 2008 financial crisis. Companies cite AI as the driver. But in New York, mass layoffs require a WARN filing that asks companies to indicate if the reduction is technology-related. Since March 2025, not one of 160 companies — including Amazon and Goldman Sachs — checked it.

The broader data agrees. OECD working-age employment keeps breaking records. U.S. unemployment sits at 5%. The Bureau of Labour Statistics projects 5.2 million net new jobs through 2034.

Economic historians note that no technology has ever spread fast enough to cause sustained mass unemployment — not the steam engine, not electrification, not the first computers. The Economist concluded this week that AI-driven mass unemployment “would be unprecedented in human history.” link

What’s happening instead: companies are cutting in anticipation of AI efficiencies that haven’t materialized yet — using AI as cover for choices driven by pandemic overhiring and interest rate pressure. That’s AI washing.

AI doesn’t eliminate the need for skilled execution — it amplifies it. The hyperscalers know this. That’s why they’re hiring, not firing.

Executives right now should be watching what AI builders actually do — not what they say.