Have you used ChatGPT?
Are you using it for work or other ways that benefit you? Despite all the concerns, which I’m also acknowledging to be true, there’s so much exciting potential in this technology.
For example, we use it to basically produce content for us. Instead of hiring a team of content creators, ChatGPT basically can do the work. Have you learned to leverage the technology to make money? Learn and gather information? Make your job easier?
Instead of me coding or hiring someone to go into the backend of my platform to code to solve a problem we have. ChatGPT can do this in seconds.
Our packers sometimes miss quantities if someone orders more than 1 of an item. I can just implement this in a few minutes for free.
A problem to be aware of is that ChatGPT states quite confidently what may be factually or procedurally wrong. Perhaps a human flaw too (ask politicians), but for people who are not aware of it, don’t care or can’t see nuance, an attitude of “Oh a computer said it so it must be right” may apply.
Some examples:
I once gave it a DOI for a technical paper and it produced a summary that wasn’t even in the same field of science.
It made elementary errors in a computation specialists in the field would not do.
Sometimes code it writes wouldn’t work, it would rewrite the code and it would still not work.
ChatGPT and other AI bots are still infants, and as their data sets and computing power increases, they will get better. And unlike humans bound by mortality, they will keep getting better.
All true. Sometimes it writes code that I see is overly complicated and I would actually ask it if it can make it more simple and it does…not sure why it didn’t do it from the start 
ChatGPT, Claude, Sage----all good
Still playing with ChatGPT. It has been more friendly and flexible than Bing or Bard.ai so far. Its best use for me, as a content creator, is going from white page to something that I can rewrite (and properly source). It is weird and disquieting the way it makes things up and randomly forgets instructions. I asked it to reformat a list of 20 medical references from some crazy-ass style to AMA style. It did a great job on the first 10 but then stalled. When I asked it to finish the list, it gave the right number of references, on relevant topics and in AMA style, but they were not the references I asked it to format. Not sure if they were real or made up.
It also does a not-too-bad job of simlpifying complex documents for a lay audience, complete with readability stats. The output needs to be verified, of course, and it is a starting point not a finished product. I heard that it is not great at math. I guess that makes sense, since it a large language (not math) model.
We can just write shit down. It’s all there. Just have to read it. We have just as much potential to learn as AI does and we can laugh when the dog farts.
We can but someone else would have to find what we write, bother to learn what we did and bother to improve on what we did. And they have to be competent and capable enough to advance the prior art in their short lifespans. We have about 40-50 years of professional working life before we near the last station.
And onwards we march, into extinction.
Not my kin. We make fires. 
Maybe we should work less and read more? 
That’s my goal.
Yes, yes, and no
It can’t seem to do everything I want, sometimes I have to ask several different ways, and it makes mistakes
But still, a useful time saver
Other ways? Wonder if it can improve on pornhub.
Version 4 fixes most of these issues, but of course anything it tells you should still be verified.
Just got a late invite to Bard from Google. Last update also added support for Chinese.
