I assume you mean large language models specifically because AI as in machine learning is already pretty ubiquitous now in everything from fraud detection to machine vision and translation on consumer smartphones.
FWIW yes I think LLMs are the next big paradigm shift in software. Whether it’s as impactful as the internet depends on how the R&D scaling plays out but it’s clearly a significant development in human-computer interaction. We’re still basically in the 1994 equivalent in the LLM timeline while speed running the dotcom boom but even if progress slows to a crawl, we’ve got a decade or more of finetuning and exploring UX left ahead of us. The pace of development these last few years has been dizzying compared to pace of development of the internet. A few years ago I was sure an interface like advanced voice mode would take decades more to come to reality.
