AI is not good software. It is pretty good people.

Etan Mollick with "A pragmatic approach to thinking about AI", taking issue with the idea that "since AI is made of software, it should be treated like other software. But AI is terrible software... We want our software to yield the same outcomes every time", which LLMs clearly don't.

He makes a couple of other arguments (we don't know how they work; we don't know how to use them) to justify the idea that "it is best to think of AI is as people", even if they aren't sentient, because "Once you see them as being more like a person in how they operate, it becomes a lot easier to understand how and when to use them."

• they're best at human tasks, not machine ones, where their inconsistency is problematic (and people are actually better)
• give them "assistant" tasks where you have the necessary expertise to supervise: " I would never expect to send out an intern’s work without checking it ... AI ... can save you lots of work by providing a first pass at an annoying task"
• their lack of consistency makes them "harder to integrate into standard corporate processes", blinding many organisations to their opportunities and risks while their employees already use them