It's like we're LLMs ourselves... generating different flavors of UIs to interface with chatGPT. That's what I think of when I see the proliferation of chatGPT UIs that hit the front page.
Inevitably, human behavior that's similar to an LLM is ripe for replacement by an LLM due to the abundance of behavioral data.
Absolutely. That’s a very exciting thing. These tools are forcing us to confront who we are.
Interestingly, LLMs (so far) are similarly limited, they can only ingest/digest so much data at a time, so we are trying to figure ways around it.
Just like humans, soon there will be lots of specialized LLMs, ones that do only certain types of law, or only one programming language, or even just one codebase, etc.
LLMs don't force me to confront who I am any more than a dog does when I see them eating. LLMs generate copycat ideas. So do I. Dogs eat. So do I. Neither of these cause me any kind of identity crisis.
Sure, not everyone in the world will be forced to ask themselves what they truly are.
However. Can dogs write too? Can they write better code than you? Can they write better letters/documents/proposals/essays than you?
Because ChatGPT can.
If you apply that at a global scale, as quickly as it is happening, it’s easy to imagine that pretty soon there’s going to be a big change in the way we interact with technology and how we view ourselves in the world.
Inevitably, human behavior that's similar to an LLM is ripe for replacement by an LLM due to the abundance of behavioral data.