ChatGPT is like Wikipedia c. 2005: simultaneously very useful and yet with just enough junk to cause problems which mostly affect those who know the least about whichever topic.
Still better than learning from newspapers, though.
I think scholarly books and translations create a sense of "Truth" that is dangerous for beginners. With ChatGPT, it creates an automatic "should I really believe this?" sensation in the reader. I LOVE that sensation when dealing with old literature.
Most translations have serious issues. Most human summaries and books about old literature are crushingly wrong or misleading in key ways. That's not necessarily a problem, unless the reader takes them as the "Truth." It takes a long time for a beginner to get the confidence to doubt the experts.
Wikipedia had this effect: making us question what we read on the internet (since it was written by amateurs). However, now that wikipedia has largely become the best source of information on the web (due to insistence on sourcing), I see chatGPT playing a key role in building critical thinking skills among topic n00bs. It can help guide a beginner towards new knowledge in an accessible manner, but yet leaves them feeling skeptical and wanting more direct information. Many experts doubt that the average person can think this way, but my experience with 15 year olds using chatGPT is that they very quickly learn to maintain skepticism. They just need about an hour or two of use, and it comes naturally.
Maybe with better models, they won't get this practice. Maybe in the future, we will roll out GPT3 for human training.
> Many experts doubt that the average person can think this way, but my experience with 15 year olds using chatGPT is that they very quickly learn to maintain skepticism. They just need about an hour or two of use, and it comes naturally.
Interesting, and I hope that reproduces outside the sample :)
I also read a lot of old translations, especially philosophy, and completely agree. It is amazing how many translators that are academic professors with PhDs in the subject matter fundamentally misunderstand the ideas they are translating, or try to seem "impressive" (and obscure their lack of comprehension) by translating simple plain text into pompous and indecipherable jargon.
Personally, I usually deal with that by reading the translators commentary so I can see where they were missing the point, and reading multiple translations.
A lot of the time I think certain ideas are semi-intentionally misunderstood, because they are personally threatening or upsetting to the translator. Nietzsche for example had a deep disdain for the type of professor that translates classic texts- from having had a bad experience as a professor of classics himself at University of Basel. His books are filled with cutting deep insults directly targeted at this type of person and their career, and when they translate it, they seem to almost always manage to "subtly misunderstand" what he's saying.
There is also an aspect of (for lack of a better term) "spiritual progression" where unless you are already at or nearly at the level of the author, you can't comprehend the ideas, and then tend to assume it is something else entirely that you can comprehend.
Still better than learning from newspapers, though.