Well obviously, it’s just fancy text. Alzheimer patients in elder care also string beautiful, seemingly coherent, responses occasionally. That doesn’t mean they are well and cured. Without a model of the world, integration of sensors, and memory of context and history, no LLM will ever produce anything of technical value. Take in mind that currently all LLMs depend on API interoperability with other, often times non-ML, solutions to create technically coherent answers. I tried CoPilot and spent more time erasing it’s bad edits than using its suggestions.
Well obviously, it’s just fancy text. Alzheimer patients in elder care also string beautiful, seemingly coherent, responses occasionally. That doesn’t mean they are well and cured. Without a model of the world, integration of sensors, and memory of context and history, no LLM will ever produce anything of technical value. Take in mind that currently all LLMs depend on API interoperability with other, often times non-ML, solutions to create technically coherent answers. I tried CoPilot and spent more time erasing it’s bad edits than using its suggestions.