The CEO who told us to brace for "entire classes" of jobs to vanish now says he got the timing wrong — while a high school in Alabama quietly builds the blueprint for what comes next.
The Summary
- Sam Altman admitted he was "pretty wrong" about AI's near-term impact on white-collar jobs, despite previously warning that the technology would wipe out entry-level positions
- The Huntsville Center for Technology is training high school students for skilled trades paying $40/hour — roles that can't be automated and for which there's massive demand
- Altman now says OpenAI has been "roughly right on technological predictions" but missed badly on the social and economic implications
- The real story isn't that AI failed to disrupt work — it's that disruption is hitting different jobs than the futurists predicted
The Signal
Altman's mea culpa came during a conversation with Commonwealth Bank of Australia's CEO, where he acknowledged the gap between OpenAI's technology roadmap and its understanding of how companies actually use AI. In 2023, he told The Atlantic that "jobs are definitely going to go away" as AI took hold. Three years later, white-collar work hasn't collapsed. Some companies cite AI during layoffs, but the apocalypse never materialized.
The more interesting development is 600 miles from Silicon Valley, where Huntsville's tech-focused high school is partnering with Toyota to pipeline students into advanced manufacturing and skilled trades. These aren't backup plans for kids who can't code. They're $40-an-hour roles in fields desperate for talent, doing work that requires human hands, spatial reasoning, and real-time problem-solving.
"We've been roughly right on technological predictions and pretty wrong on the social and economic implications."
The pattern emerging isn't white-collar extinction. It's bifurcation. AI is compressing the value of generic knowledge work — the kind you could do remotely in your pajamas — while increasing demand for physical-world expertise. You can't prompt-engineer a transmission. You can't have ChatGPT troubleshoot a production line jam. The jobs AI can't touch are the ones that require being there, in three dimensions, with tools in your hands.
Altman's revised scorecard matters because it reveals what the builders of AI are learning in real time:
- Technology adoption is slower and messier than lab demos suggest
- Companies use AI to augment existing workers more than replace them
- The social infrastructure (training, labor markets, policy) moves at a different speed than model capabilities
Fortune's reporting shows what happens when you stop waiting for AI predictions to come true and start building for the labor market that actually exists. Alabama isn't hedging against AI. It's training for the economy AI is creating — one where physical-world skills command premiums because they're scarce and non-fungible.
The Implication
If the CEO of the company building AGI admits he got the job-market timeline wrong, stop planning your career around his old predictions. The white-collar jobs at risk aren't disappearing overnight. They're slowly getting cheaper, more competitive, and less defensible. Meanwhile, skilled trades are the new moat. Watch where Toyota and other manufacturers put their training dollars. That's the signal. The kids learning to run CNC machines in Huntsville have more job security than half the product managers in San Francisco.
The real AI jobs story isn't about what disappears. It's about what becomes valuable when everyone has the same chatbot. Physical presence. Specialized motor skills. The ability to fix things that break in the real world.