Like most Americans, I am indeed ignorant about history! I blame bad teachers, boring textbooks, and my own laziness. Probably mostly my laziness. The AI part of your post is well-written and is not quite as bad as the American History and Western Civilization. That is scary! I have resisted AI so far - "Roko's Basilisk" is a part of that. Good article!
It works well for many subjects, I test it often. But it gives primarily a conventional, consensus view on anything that is necessarily subjective. To get alternate views it has to be prompted correctly. Once you learn how to push it, it will produce more unconventional answers.
ChatGPT does not actually think, it is a language aggregator. It takes text from what it has scanned on the Internet, which is less than 3% of the extant accessible content, and formulates answers. So it is limited to what has already been written. Also, it cannot do math. If you ask a math question it has not encountered it will give a wrong answer and not tell you it’s wrong.
Still, very impressive and fast. It will improve, but it will never think like a human. That cannot be codified. Humans are not biological computers as many strict materialist ‘scientists’ believe.
Like most Americans, I am indeed ignorant about history! I blame bad teachers, boring textbooks, and my own laziness. Probably mostly my laziness. The AI part of your post is well-written and is not quite as bad as the American History and Western Civilization. That is scary! I have resisted AI so far - "Roko's Basilisk" is a part of that. Good article!
It works well for many subjects, I test it often. But it gives primarily a conventional, consensus view on anything that is necessarily subjective. To get alternate views it has to be prompted correctly. Once you learn how to push it, it will produce more unconventional answers.
ChatGPT does not actually think, it is a language aggregator. It takes text from what it has scanned on the Internet, which is less than 3% of the extant accessible content, and formulates answers. So it is limited to what has already been written. Also, it cannot do math. If you ask a math question it has not encountered it will give a wrong answer and not tell you it’s wrong.
Still, very impressive and fast. It will improve, but it will never think like a human. That cannot be codified. Humans are not biological computers as many strict materialist ‘scientists’ believe.