A divisive AI researcher is sounding the alarm about the deep implications of cheap, superhuman AI. And his insights could change how we think about the future of work and the economy.
Carl Shulman, an independent AI researcher from the Machine Intelligence Research Institute and Oxford University's Future of Humanity Institute, recently published a A marathon interview on the 80,000 Hours podcast. In it, he outlined a future where Artificial General Intelligence (AGI) could dramatically reshape our world before we think.
What does this mean for everyone? you?
I got the answers from Paul Roetzer, founder/CEO of the Marketing AI Institute. Episode 105 of Artificial Intelligence Show.
The coming abundance of intelligence
Shulman's basic argument is simple but profound: We are moving from a world of scarce intelligence to one of abundant intelligence.
“Right now we lack intelligence that is as capable as humans in general, especially human experts,” says Reutzer.
But, thanks to increasingly powerful AI, we are moving towards a world where intelligence is abundant. This shift could have far-reaching implications for the job market.
Roetzer explains:
“If you take a step back and look at a more macro view, the question is: What is the greatest value of implementing AI? What roles is it most likely to be applied to?”
These may first be the roles that are the most expensive to use because, until recently, they required a high level of (relatively low) intelligence to perform.
Think: high-value, high-earning professions like managers, directors, executives, and experts like lawyers and doctors.
The AI that never sleeps
It's not just about intelligence. It's also about productivity. Once AI reaches the level of human intelligence, it performs its cognitive tasks continuously.
“AI doesn't sleep. It doesn't take time. It doesn't spend most of its time and career in education or retirement or leisure,” says Reutzer.
AI can work 8,760 hours per year (full-time employment 24 hours a day) with 100% efficiency, compared to 2,000-2,200 work hours per year for a human.
This productivity boost can start at the top of the wage scale and work downward—reshaping traditional assumptions about the economy.
Are we ready for it?
Despite potentially large-scale disruption on the horizon, Rutzer sees a lack of preparation on the part of business and community leaders.
“I'm increasingly convinced that this is going to be extremely disruptive to jobs in the future of work.” they say. “And yet you look around and it's not being talked about as much.”
Despite the many doom and gloom headlines about AI's impact on jobs, Roetzer doesn't see enough serious discussion, modeling, or planning to actually address the possibility of AI disruption to the economy.
He draws parallels to the early days of AI in marketing. When he launched the Marketing AI Institute in 2016, few recognized the transformative potential of AI. Now, he sees a similar blind spot when he talks about AI's impact on jobs in industries.
In fact, many economists do not predict a high level of disruption. In an 80,000 Hours interview, Shulman suggested that this is because the scenario is too far-fetched based on historical precedent.
“They look at what AI can do today, and then they make their own reports, projections and predictions based on their understanding of that,” he says. “What they don't want to do is develop a deep understanding of what one to two models might look like 12 to 24 months from now. What does it look like when we're on GPT-6 or Gemini 3 or Claude 5? Think?
At the moment, many AI researchers (including Shulman) are operating under the assumption that scaling laws will, at least in the short term, produce rapid and profound growth in AI capabilities.
However, few economists and business leaders seem to be taking these assumptions seriously.
“If they're only 50 percent right, then we're in a completely different world,” Rotzer says. “We're in a completely different world in terms of what jobs look like, and we're talking about three years from now.”