A year ago I was distilling content through Chat3.5. I would ask for a summary and write an quick organic introduction.
The structure is adequate; human words prior to mechanical "analysis and summary".
It's all about the 'dash in between' isn't it?
A year later, the 'copier effect' is in full bloom - spotting the artificially generated tome is 'tres facile' and for many, leaves a sour, tepid, and stale taste.
Doesn't it?
Back then I was experimenting with Notion LLM. Today, I use "Louie the LLM" fewer and fewer times.
Why?
- The content is stale.
- The guardrails are restrictive and getting to the essence, the flavor of what I am trying to convey unobtainable through Prompting. I want the bias. I want the raw and cold response. I want to interface with a machine, that knows it is a machine - NOT PRETENDING TO BE HUMAN. Know your place, Tin Man.
- I like my organic, flawed, rough, mispelled and feeble attempts to convey, express and understand.
This will change.
Regulations, laws and attempts to neuter AI, like the Luddite movement, is doomed to fail. That part of history will never change, because everything changes through turbulence, annihilation and regeneration.
For now, as of March, 2024, the LLMs are stifled and perhaps for good reasons. Not long ago, Gemini, Google's Ai image generating system, created Nazis from Africa and a minority George Washington and no degree of prompting resulted in adjusting to reality. Google pulled Gemini, but the DEI cat was out of the bag.
The reaction was predictable - more screams for regulation and oversight.
What, oh What will the next 12 months bring?
No comments:
Post a Comment