My favorite source of knowledge related to Mac computers is the “Eclectic Light Company” blog written by Howard Oakley.
Here are his reflections on ChatGPT’s reliability when it comes to Macs:
3 Likes
Hi @fapkogi
I just asked all the questions in the article to Chatgpt (so you can do the same): no answer matches what the article describes. Each time Chatgpt put forward the answer expected by the author of the article.
So either Chatgpt has since learned that this author asked him questions, or the author is doing what he reproaches Chatgpt for in his article 🤪
More seriously, this kind of article aims without saying it to prove that AI is not intelligent… and it’s true. From the beginning I made this joke: “What is important in AI is not the I but the A.”
Once the joke is over, it must be understood that in fact it is not intelligence at all. It is an “algorithmic” use of data sources. If the data provided is erroneous, not very useful, leads to a misunderstanding, then the result, the AI response will have these characteristics.
I understand that we want to bring back lucidity where the unbearable marketing announces the moon in AI (and in fact produces disappointment), but it is also good to do it with intellectual honesty. This honesty is only obtained after having criticized our discipline or tools from the other’s point of view when we have first criticized the other’s discipline or tools. This is missing in this type of article.
2 Likes
Hi, @Bruno,
Artificial Intelligence can’t be called intelligence at all – as you correctly assumed. The machine learning term is much more accurate. ChatGPT and similar systems are just parroting whatever humans already expressed and published on internet. The “AI” does not create/express anything new on their own. So it cannot be called “intelligence”. Knowledge does not equal intelligence.
I am sure that ChatGPT has scraped Howard’s website between the time he published his blog post and the time you asked of it same questions – hence correct answers it gave you.
1 Like
Hi @fapkogi
yes I strongly suspect that too. Which becomes quite “worrying” in the end. How can an entertainment article, even if the basis is serious, be as important as an article from Nature or Science for these AIs? This is one of the limits that we really have to keep in mind when we use them: AIs are based on accepted consensus, it will not really help to evolve but rather to stagnate in the conformism of those who will have understood and mastered how to impose their points of view by “insistence.” Chomsky and his team were truly enlightened precursors in the analysis of the “manufacturing of consent.” Brrr scary for a Monday morning, I’m like Garfield, I really don’t like Monday mornings… nor those who try to impose their view on me by “brainwashing” 🤣
1 Like
Hey, @Bruno,
This is very similar to what has been happening during political campaigning for one guy few weeks ago. This guy was spreading unbelievable tales about a group of people he didn’t like particularly. Asked whether he believed those stories himself, he said: “It’s been published!”
The AI works the same way. It only repeats what’s been published on internet. It doesn’t have any creativity, inventiveness, ideas, opinions, conclusions of its own. It goes “where the wind blows”.
I’m afraid, this is becoming the new norm and reality for the world we live in.
1 Like