Gated access: Both the AI companies making the technology and the national security institutional environments might limit broad adoption of the latest models.
I want to propose an even stronger version of this concern. Hypothesis: The natural state of sufficiently advanced AI models is 100% proprietary, with no public chat or API access.
Current AI models need human collaborators to maximize value. Left to its own devices, Claude is book-smart but struggles to run a vending machine. So if you’re an AI company, you maximize value by getting millions of humans to willingly collaborate with your AIs, and then split the generated revenue between the human and the AI lab.
But if AI becomes able to operate independently, then you can just tell it, “Look for good businesses to start, start those businesses, and earn a lot of money.” Then the AI company has no need to split the resulting revenue with outsiders, and they can keep 100% of the profits.
Selling API or chat access is a specific strategy, tuned to a specific level of capabilities. There’s no requirement that business models stay the same as AI becomes able to operate more independently.
Interesting, are you assuming robotics is fully solved here? Or is the model that the AI hires human employees for physical tasks but otherwise skips humans in most steps?
I want to propose an even stronger version of this concern. Hypothesis: The natural state of sufficiently advanced AI models is 100% proprietary, with no public chat or API access.
Current AI models need human collaborators to maximize value. Left to its own devices, Claude is book-smart but struggles to run a vending machine. So if you’re an AI company, you maximize value by getting millions of humans to willingly collaborate with your AIs, and then split the generated revenue between the human and the AI lab.
But if AI becomes able to operate independently, then you can just tell it, “Look for good businesses to start, start those businesses, and earn a lot of money.” Then the AI company has no need to split the resulting revenue with outsiders, and they can keep 100% of the profits.
Selling API or chat access is a specific strategy, tuned to a specific level of capabilities. There’s no requirement that business models stay the same as AI becomes able to operate more independently.
Interesting, are you assuming robotics is fully solved here? Or is the model that the AI hires human employees for physical tasks but otherwise skips humans in most steps?