3 Comments

I'm sure we'll know when LLMs replace human customer support: once we start seeing complete formal sentences with correct grammar :)

Expand full comment

I really hope Neal Stephenson's "bogon" (from Anathem) catches on for this stuff.

On a more serious note, I think certificate chains (cleverly hidden behind an intuitive UI) that let people judge the source/path of information is the only hope.

Expand full comment

A slightly less bleak take: one possibility is that things like the LLM might simply raise the level of abstraction of human interaction with technology. Something like how computer user interface abstractions have evolved over time so that we never have to think that what we are ultimately doing is directing electrical voltages. True creativity, whatever that means, seems to still be the domain of human ingenuity. ChatGPT-like technology might free humanity to not “worry the details” so to speak.

Or maybe I am just influenced by the fictional human computer interaction portrayed on Star Trek Voyager that I am binging through right now. The idea of “programming” a holodeck experience by simply providing a description of the end result seems within reach instead of worrying about 3D modeling, rendering, GPUs and so forth.

Expand full comment