ChatGPT already has “Ph.D.-level intelligence,” in that once you remove it from its extremely narrow field of expertise (assembling words into credible-sounding passages) it is wildly out of its depth, but people vastly overestimate its authority because they don’t understand what “intelligence” is.
I remember someone using ChatGPT to argue with me over a book which the author withdrew and redacted after he found that his primary source had falsified her data. They got ChatGPT to make an argument that we should still use the book, with a long winded explanation why.
I just blocked them.
There are some folks (I’m guessing mainly students & Internet autodidacts) who absolutely believe that intelligence & expertise consist in assembling words into credible-sounding passages. From 16 to 18.5 years, I was one of their number, so I can say this.
I really hate the term AI. It’s marketing bullshit. I personally don’t bribe machines can ever actually think or gain consciousness.
Skynet is impossible.
You'll know its a real PhD haver when it starts screaming 'that's doctor chatgpt!!' at anyone trying to use it.
(Thank you to all the doctorate havers who don't do this but maaan it happens to me so often at work and nope.)
I was recently at a vendor conference where a speaker quoted Microsoft in saying Copilot has roughly a "sixth-grade" intelligence right now. Which makes sense, if you consider how most sixth graders are prone to be confidently incorrect.
I also blame the AI surge on folks for telling folks to use chatGPT as a psuedo search engine and claiming it's fine because "It gives you all the resources up front like a any other"