Is AI the bicycle for the mind (following Steve Jobs), or is it the mind riding the bicycle (quite literally like the 20-year-old robot here even before the Transformer)?
In this article, Tim O’Reilly, countering Jensen Huang’s keynote remarks, frames this as a question of function: Is AI a tool or a worker using other tools? He explores a number of premises and concludes the LLM is “a tool that knows it’s a tool.”
This may actually be an apt way to describe an agent: a tool that knows it’s a tool -and- can use other tools.
Credit for the picture goes to Koji Sasahara / AP.