People are nitpicking as a response from a vocal group that loves to spew the doom of every knowledge worker out there. As if this will replace programmers, doctors, writers, copywriters, etc. Looking for reasons why it would not its only natural, For that purpose, I think it still has fundamental flaws that are not solvable as easily as some seem to believe.
People are also impressed, given how much is being used. Being impressed, I prefer now to know and explore its boundaries. Is this really going to a place where it will replace those workers or are those limitations a fundamental barrier to what it can do based on the method in which it works?
Its strengths on my tests so far:
* Summary of content for specific questions
* Language learning reference and translation
* Rephrasing and correction of grammar in text (paragraphs at most)
Its weaknesses:
* Trust of results in complex responses. (clear wrong answers)
* Give references.
* Ambiguous questions and clarifications. (nitpick, I think its fine as it is)
* New ideas or anything thats not been documented and done before or instructed in prompt (duh)
This last weakness is the crux of what annoys people so much, its a predictive language model, not AGI. I don't think it's anywhere near close to replacing any worker, supporters (I am one) should focus on what this really can do, which is to increase productivity and being an incredible tool.
ps. I asked it to rewrite this response and it tends to prefer the passive voice as if its writing and article. After a few tries it didn't give me a good result I could just replace what I wrote here. It doesn't really understand what I wrote, it just rephrases in its preferred form (article type constructions). Its still super helpful to "unblock" a hard to write paragraph for me, a non native speaker of english.
> If a tool enables a team of 4 do the same things that previously was done by a team of 5, the tool replaced a worker
Only if you believe that demand for work to be done is fixed as the cost of doing it goes down, a belief that was last reasonable to hold about three centuries ago...
People are also impressed, given how much is being used. Being impressed, I prefer now to know and explore its boundaries. Is this really going to a place where it will replace those workers or are those limitations a fundamental barrier to what it can do based on the method in which it works?
Its strengths on my tests so far:
Its weaknesses: This last weakness is the crux of what annoys people so much, its a predictive language model, not AGI. I don't think it's anywhere near close to replacing any worker, supporters (I am one) should focus on what this really can do, which is to increase productivity and being an incredible tool.ps. I asked it to rewrite this response and it tends to prefer the passive voice as if its writing and article. After a few tries it didn't give me a good result I could just replace what I wrote here. It doesn't really understand what I wrote, it just rephrases in its preferred form (article type constructions). Its still super helpful to "unblock" a hard to write paragraph for me, a non native speaker of english.