I’m looking for the paper “Terpene separation and analysis by high-temperature distillation and comprehensive two-dimensional gas chromatography” by S. Kim and others, published in the Journal of Chromatography A in 2009. Can’t seem to find any info on it anywhere. Wonder if anyone’s had anymore luck/better means of searching.
2 Likes
That was actually what I was starting to think… Chat GpT actually suggested the paper. Lolz… Looks like they still have some bugs to fix there…
5 Likes
did chatGPT tell you that?
12 Likes
Was seeing how it does at specifics… apparently pretty bad… Def weird that it just made something up though.
1 Like
I went through all the hits for “terpene separation and analysis’ in J Chromatography A in 2009 and no luck finding anything remotely similar.
Ai false answers are going to be a massive battle.
2 Likes
Not yet, I’ve been wanting to play around with that but haven’t yet. Do I just refer to it as Dan or do I need to do anything else?
I was asking it about distilling terpenes from hte, then asked for it to cite a paper with the most relevant data
it’s been days, I would assume it has now been trained not to fall for that shit…
wow. I’ve had papers handed in to me that cited that way, but would not have expected an Ai to take that approach.
…yeah, we’re fucked.
3 Likes
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5113832/
https://onlinelibrary.wiley.com/doi/full/10.1002/elsc.202100014
Closest I could find, lots of reading but two different papers on purification steps & methods. Not the same author as ai pursued you with but hope it helps!
2 Likes
Big time. Any and all trust in internet interactions and information is in it’s final days right now. If you didn’t personally see and hear the person in real life, that video, audio, article, etc you saw online is soon to be more likely faked by the AI than real
5 Likes
I find it personally hilarious. The bot is just doing what it was programmed to do, process and average huge datasets and regurgitate a synopsis. When you ask it for a citation it’s creating one from all of the authors and titles. The actual journal name exists so much in its dataset it didn’t deviate in reporting it.
2 Likes
I’ve had a lot of good luck using chatgpt for research. For example a small group of scientists in Portugal studied a specific part of the plant that seemingly no one else has looked at. Chatgpt had no problem finding these papers written in another language. I had no hope of finding the papers with a normal query
2 Likes
I remember getting a key to the research library and having to spend hours searching through hard copy chemabstracts.
It’s a whole new world.
That’s hilarious that it just pulled the legit looking citation out of nowhere. A troll at heart.
2 Likes
I’ve had similar luck with it, I find that it’s a good way to get high level summaries of specific topics. In a couple of hours spread over two sessions and some manual research in between, I was able to collect as much information on a specific narrow topic as I likely would have managed in a week or two of reading source material.
I definitely don’t have the depth of knowledge that I would by doing it the hard way, but for this one area that’s not really an impediment - I’ve been able to very quickly narrow the experimental design space and make substantial improvements in a specific process because of what I’ve learned.
2 Likes
Just remember than it can only report an abbreviation of what other people have reported, and has no way of validating independent data. Any topic where there is controversy or disagreement over will probably yield garbage.
1 Like