"massive cherry picking"

I think the coined phrase is "prompt engineering".

Side note, where's the eye roll emoji hiding?

reminds me of the BTC pivot from "decentralized currency" to "store of value"

goalposts

If you understand what an LLM is, chain of thought isn't something you eye roll at.

If you understood what an LLM is not, you might..

Right, another low effort deflection into the "It (is/is not) Intelligent" diversion.

You can sit in that tarpit and duke it out with fanatics and doomsayers while patting yourself on the back for "real zingers"... or you can just learn to use the tool and extract your own value from it.

"Right, another low effort deflection into the "It (is/is not) Intelligent" diversion."

Straw man.

"You can sit in that tarpit and duke it out with fanatics and doomsayers while patting yourself on the back for "real zingers"... or you can just learn to use the tool and extract your own value from it."

GPT offers zero value which reflects my position. Though, it does perform some neat tricks.

If it's a straw man feel free elaborate on your incomplete commentary with "If you knew what it's not".

GPT offers you zero value because you don't know how to use it, so you're limited to cheap tricks.

"GPT offers you zero value because you don't know how to use it, so you're limited to cheap tricks."

GPT offers me zero value because its internal processes are opaque. It's fundamentally unsuitable for business.

Personally, I have no use for it either. Its output is inferior to mine.

Even with cherry pickings it’s still amazing a digital ‘plant’ was invented that has any cherries at all.

If the cherry picking is less effort than painting a picture from scratch then I wouldn't say it's dishonest.

It is fine if one says that they tried various prompt tweaks and out of twenty attempts (for example) here is the the best.

To present the best prompt/response while not disclosing that it is a result of trial and error is a different thing altogether.