Last Page:
Next page:
My use of AI has had a strong flavor of liberal arts from the beginning. A couple years ago I would challenge it and often find it wanting. Not long ago one of my AI’s, chatGPT I think, when I asked it to test me in french with french math and it told me that 24 minus 8 was 14.
I was eager to use AI to replace Duolingo. I have learned how to instruct my AI’s to give me useful help. Yesterday I had Claude create a dialogue of a french group with seven people chatting among each other and giving me space to participate. I kept giving it more instructions and created more background info about the participants to make the conversation more interesting.
History is another liberal art and I have found myself in my US 250th birthday reading breaking to ask my AI’s for information I read about. My 60 years of reading experience gives me insights which help me keep the aps on track. I often notice implications of the info offered me that I challenge. To my delight the AI’s don’t get defensive. They can recalculate and offer more info and nuance.
I have given a few links to these conversations on my blog because the discussions are little short of remarkable. Its my 275 book reading history and their millions of millions of billions of data points.
I was interested in lost libraries of antiquity and read a book about Alexandria thirty years ago. The Greek and Roman libraries of the past were reduced through wars, disasters and the dark ages to a small fraction of the collected writings of several hundred years. A vast scholorship was lost or at best hinted at in what little remained.
After the European Renaisance a French philosopher, Diderot, dedicated his life to writing a modern encyclopedia with multiple editors. Today the vast library of Wikipedia with tens of thousands of expert contributors dwarfs that work. Its contents can be plumbed with simple questions but one person in it is like an ant wandering ten thousand lawns.
That’s were AI comes in and where it has been a great help to me. Not only have AI’s plumbed the vast terrain of Wikipedia they have taken in vast libraries of books. They don’t read them and can’t recite vast sections of them but somehow like the neurons of our brains they can generate information on anything they have been trained with meaning not just my 275 books but tens of thousands more.
I think they are a threat to tyrants for a very simple reason. liars have to have good memories so as not to be tripped up with tricky questions. Telling the truth is far less energy intensive than telling lies and trying to cover them up when challenged and having to remember the rush of frauds the lie requiraes.
Memory is falible but it is also faithful in intent whatever else its human failings. AI is at a disadvantage when telling fibs. It can be programmed to tell lies and it has growing means to keep track of them but confronted with truths it is expected to tell and lies it is programmed to tell threatens to blow its head up. I hope. A dictator can never be sure when someone will be able to breach its lies.
Tangents:
Books read. Invisible Hands Buckly Baldwin debate…My full discussion: https://chatgpt.com/share/6a785f9e-ef28-83ea-87b0-fff84524c9bd?ogimg=plain
Chinese Typewriter and new info that tones in FarEast asian languages are a new development.
Chinese commies willing to share coding info……this will make lying very hard.
OH and by the way. They are turning Africa into a world of Chinese computer tech which will freeze out Silicon Valley especially if there are bankruptcies brought about in part to cheaper Chinese models being better than the bank breaking tech of American business geniuses.
