It’s looking like it’s going to be daily (hourly?) thing for folks to demonstrate that ChatGPT spouts (1) woke idiocy and/or (2) utter foolishness and/or (3) laughably, demonstrably wrong answers. I have to laugh at the absurdity of recent claims that ChatGPT is some sort of game-changing, disruptive technology that that’s going to change lives, change the world, put Google out of business, put bookstores out of business, put libraries out of business, yada yada yada. As far as I can tell, ChatGPT is little more than a garbage-spewing gimmick du jour.
Perhaps ChatGPT won’t do all things you list, but the idea of Artificial Intelligence is a game changing technology that’s still in it’s infancy, remember ChatGPT is in version 1.0 essentially a beta test version, unleashed on the internet for people to use and to flush out things like this article.
Major companies and governments are pouring billions of dollars into the development of AI technology. Microsoft alone has poured 10 billion into ChatGPT, some of the smartest venture capitalists out there have also poured billions more into AI development.
The issue with AI is the potential do tremendous good is nearly unlimited, but it also has the potential to do unlimited evil in the hands of the wrong people.
ChatGPT is just one model of AI, they’re many other models that have been fed information about specific topics like human anatomy, etc...one such model was fed information about biological weapons and in less than 6 hours listed 40,000 new types biological weapons. Here’s an article.
https://www.theverge.com/2022/3/17/22983197/ai-new-possible-chemical-weapons-generative-models-vx
Here’s another article, this particular model started out creating digital images from a description of what someone wanted to see, it was adapted to human proteins, the human body contains about 20,000 proteins to carry out particular functions, AI was used to generate models of proteins that do not currently exist in nature, scientists are using this technology to potentially treat cancers and other deadly diseases, exploring areas that have never been explored.
The potential is unlimited not only for good but for the bad.
Right?!? ChatGPT is like talking to people who are amazed that you don’t realize how “wrong” Conservatives are on every topic.
Literally, I have had people say, “See? It is all over the news!” because they heard a story on three news channels and read it in four different papers.
Those people do not realize that all those “news” outlets are quoting the exact same, BS, lying enemedia source - Associated Press!! But they claim is it amazing and wonderful AND TRUE, because they saw it on different channels and in different papers!
I wouldn’t be so harsh. Yes, asking it anything that requires moral THINKING is out of bounds for ChatGPT. It’ll just give you “words tied together” from the training data it has available...which one can argue is sourced from left leaning material (e.g. silicon valley leftists biasing the training data).
That said, I can ask it to write C code for a Commodore Amiga from the 1980’s and it’ll do it (e.g. write the code to put basic UI together, with event handlers). I can ask it to write a love song about winter in Milan in the style of Metallica...and it’ll do it. I can give it facts about my software company and ask it to write a contract to sell some IP assets to another party. I’ve given it medical symptoms I’ve had in the past where it gave me the same diagnosis as my doctor. I’ve asked it to compare and contrast different technology standards and it did perfectly. I’ve given it data sets and asked it to do different statistical analysis on it, done with a breeze.
ChatGPT is becoming my go-to source on anything, even if it is only to get a better understanding of the basics of some topic - but the topic can be anything. Just like Wikipedia, don’t count on it for anything subjective. If the topic is medical, legal, technical, etc. then it brings the world to you, instead of you Googling for hours and trying to stitch together different sources of information of varying quality. Yes, even in these cases, it can be wrong but it’ll usually be spot on.
IMO this is a massive game changer, you just need to understand how to use it. Even as a learning tool, it’s never been this easy to research and learn.