Expect IP issues
This article from WSJ this AM. (Link bypasses paywall)
Similar issues with Chat gpt and copyrights coming to a court near you.
I really hope they don’t put the brakes on this. IMHO the potential for this AI is massive. You wait until the Bing/Google AI search war.
Search will be changing a lot in the coming year. You will get much more direct answers without having to sift anymore.
I did some experimentation with Chat gpt and IMO it would its results would have been unacceptable for my elementary school reports. Whether it violates IP or not.
It is like using an encyclopedia to research your report topic. Teachers and professors who have been fooled by the output should be fired.
In my tests I chose to seek comparisons between specific details of differing political movements, and attitudes.
I chose a mid 20th century leftist movement and a 21st century “right wing” movement and their attitudes toward government.
The result was a comparison of the movements which could have been based on the first paragraphs of wikipedia articles and totally lacking in attitude to government.
There was obviously inadequate source data for Chat gbt to work with.
This is exactly why the major effort to do medical diagnosis by IBM’s Watson was terminated. It was impossible for IBM and its medical partners to input sufficient data for Watson to be reliable.
When I was an undergrad in the 1960’s, when I took a social science course which had an exam essay question I could not answer, I wrote everything I new on the topic related to the question. Even though I failed to answer the question, I got at least a B on the exam.
I expect the Bing/Google AI war to be more of the frustration the stupidity of the search engines already cause.
It is almost impossible to find detailed information. Having been educated when one was taught and understood a broad number of subject areas, I do not need overviews, I need to know what I was not taught in subjects other than my majors.
The new AI will help the inarticulate and illiterate to find the superficial information they were never forced to be exposed to.
6 Likes
Depends on user parameters - chatgpt responds to the parameters that you set - akin to a programming language like python where you can tell it what you want and set the parameters and it does it to various degrees of success - I found it quite useful in research and even rewriting things - It is narrow AI and not anywhere close to actual AI so a lot of it has to do with the user.
It’s very powerful - will a majority of users be able to use it to its potential? No. Will they code that into their algorithm - no clue on how they could considering the thousand ways people think.
We’ll see.
1 Like
I was going to try it out for listing descriptions (for my own website, maybe E). A lot of my listings are very similar and I get tired of trying to word them differently. Sounds like maybe it’s not worth it.
ChatGPT is really good at making DnD characters though.
1 Like
Try it out for yourself before you decide not to based on a post
3 Likes
It is only as powerful as the database it has to use, just as the Watson example was.
The Watson medical diagnosis experiment was well publicized, its abandonment, hardly at all.
Unless the data is there, and not buried by duplicative, extraneous and incorrect information, it cannot use its power.
When I search for the original source information on a topic when using Bing or Google, I find far too many results which are duplicates of second or third hand interpretations and often need to use another search engine to actually find the original source. And when I find the source, I realize the inadequacy of the second or third hand interpretations.
This is not limited to the political or ideological results from various websites.
Many are from well credentialed journalists, who fail to understand what they are reporting on. Some are from would be statisticians, who fail to understand the data they are analyzing or whether their choice of data meets scientific standards.
The old adage taught to every programmer in the 1960’s and 1970’s has been forgotten by modern computer scientists and the venture capital industry - Garbage In - Garbage Out.
4 Likes
I think using AI to freshen up product descriptions is a possibility with a lot of potential @LR72 --although I gotta say, I don’t love having to compete with a computer.
Using ChatGPT to communicate with customers or Seller Support? Yikes, I think.
Right but you can tell it which sources you want - which is where use varies. I told it to look up a theological discussion from long ago only using sources from certain years and the reasoning those sources used and it did just that indicating who said what and why. You can set parameters for it not to use certain sources - I agree the database is flawed, but how would that differ from sifting through google or any other search engine? Just a thought.
1 Like
I use it personally for my own website as well as my work website. For the purpose of writing SEO articles it has taken a 5 hour job and made it less than an hour. I guess it depends how deep on a subject you are wanting to get it. For general inquiries and information it can definitely do wonders and you can then sew those together and create a decent article.
I will say this. Love it, or hate it, its coming boys.
1 Like
Honestly, if Amazon replaced their SSupport with AI ChatGPT, I am 99% confident it would GREATLY improve the experience.
7 Likes
I split these replies out from the introduction topic to this shiny new one. Chat away!
3 Likes
LMAO this is true
1 Like
100%, and I don’t even say that as a joke. It would. It would 1000% understand our questions and concerns way better than what they have now.
The demand is real right now.
1 Like
I noticed you, running around doing “mod-y” stuff. 
So nice to see an adult in charge…
2 Likes
Please, don’t use the “A” word LOL
1 Like
It’s really nice to be able to just handle stuff like this. No big issue, just get 'er done!
@Default_Username I honestly fear ChatGPT, and below are some of the reasons why. But also, I have a background in science and academia AND am a misinformation and disinformation warrior (particularly necessary in my state), so I know that this isn’t the sexy side of AI…
- The Next Great Misinformation Superspreader: How ChatGPT Could Spread Toxic Misinformation At Unprecedented Scale
- Fact-Checkers Are Scrambling to Fight Disinformation With AI
- AI is eating itself: Bing’s AI quotes COVID disinfo sourced from ChatGPT
I do love the Van Gogh-style renderings, but my completely gray mid-40s head also saw @lake’s graphic and screamed YASSSSS 
6 Likes
Honestly, I got the idea after reading how real estate agents are huge users of it for writing home descriptions. I think I know what my customers are looking for, but maybe AI is smarter than me. Maybe.
It is like the story of the Internet. Use garbage technology, become a pervasive part of life, and create great wealth.
Because the Internet was built on impossible to secure ARPAnet technology, a recent news article claimed that there was a 40K job shortage for cybertechnology experts.
As for the position it was impossible to secure, that was the DOD’s ARPA which made that analysis, unanimously agreed to by all of the networking expects of the time, and hardly a day passes when they are not proven correct.
DOD never used ARPAnet for anything where security was important - they had separate secure networks for that, whose technology was not shared publicly.
This stuff will eventually all come crashing down due to the flaws which were known and ignored, The Amazon ghost or zombie listings being the results of the impossibility of distributed database systems to retain data accuracy another key failing which will continue to increase in frequency with the increase in database size.
This will obviously impact the data used for AI.
5 Likes