As models become more powerful, the primary challenge shifts from improving capabilities to creating better ways for humans to specify what they want. Natural language is too ambiguous and code too rigid, creating a need for a new abstraction layer for intent.

Previous technological revolutions automated physical labor but enhanced human thinking. AI's goal is to replicate and surpass human cognitive abilities, creating a categorical shift that threatens the core of human economic value.

Previous technologies were labor-saving tools that executed human-defined tasks. Howard Marks points out that AI is fundamentally different because it possesses autonomy. It can design new jobs, assign tasks, and even help create its own successors, operating without direct instruction in a way no prior technology could.

Even super-capable AI will always look back to a human and ask, 'What should I do next?' The economic and technical incentives are aligned to build compliant tools, not beings with their own intrinsic motivations. This fundamental lack of agency ensures humans remain the drivers of value and direction.

Previous technologies replaced physical or rote mental labor. AI is a categorical error to view similarly because it's the first tool that can think and execute. It replaces the pattern-recognition and reasoning layer *above* the task, representing a zero-to-one moment in technological change.

AI won't just help people use applications like Excel; it will eliminate the need for them entirely. The final user interface will be a conversational agent that manages underlying data and executes complex tasks on command, making traditional software and its associated friction obsolete.

Previous enterprise software, like SAP or Salesforce, only required users to learn its functions. AI is different because it's a partner you must also teach. The quality of its output depends entirely on the quality of your instruction, requiring a new meta-skill of co-evolution with technology.

Marks distinguishes AI from all previous technologies like the internet. While other innovations increased human productivity, AI possesses autonomy—the ability to be given a task and figure out *how* to do it independently. This quality makes its future impact uniquely unpredictable and impossible to forecast.

Previous transformative technologies like the steam engine or personal computer were tools that augmented human labor and intelligence. AI is distinct because it seeks to replicate and replace human intelligence itself, posing a unique, potentially terminal challenge to the value of human labor in the economy.