I guess I would assume that GPT-style models don’t have a great chance of escaping the valley of confused abstractions because their interface to the real world is very weird. Predicting scrapes of internet text.
Updated the diagram I drew on to better illustrate my confusion.
And yeah, it may very well be that GPT-style never leaves the valley. Hmm.
I guess I would assume that GPT-style models don’t have a great chance of escaping the valley of confused abstractions because their interface to the real world is very weird. Predicting scrapes of internet text.
Updated the diagram I drew on to better illustrate my confusion.
And yeah, it may very well be that GPT-style never leaves the valley. Hmm.