"A computer will do what you tell it to do, but that may be much different from what you had in mind."
    — Joseph Weizenbaum

So much confusion about this "AI revolution" ("ChatGPT"), I thought about making a video or a talk; for now it goes only a brief non technical text:

Computer already for sometime see well, read well and also write well. But the computer still couldn't "listen" very well.  The groundbreaking you hearing recently is because  "A computer will do what you tell it to do, but that may be much different from what you had in mind."
    — Joseph Weizenbaum

So much confusion about this "AI revolution" ("ChatGPT"), I thought about making a video or a talk; for now it goes only a brief non technical text:

Robots already for sometime see well, read well and also write well. But the computer still couldn't "listen" very well, he couldn't grasp well a question.  The groundbreaking news you hearing recently is because they managed to install a good hearing device on "AI" with certain better capabilities than the Google search bar you're used to (technical deepening: https://shubhamsaboo111.medium.com/prompt-engineering-the-career-of-future-2fb93f90f117 ).

So do robots do everything well now? No, these AI robots can answer 3+4 very easily, but it's a little bit more difficult to answer 7+8, because they're not soo good with the idea of "carrying the 1" (The progress of this and mostly any other problem solving capabilities is being tracked at https://github.com/google/BIG-bench/blob/main/bigbench/benchmark_tasks/keywords_to_tasks.md#summary-table ). It's matter of time, though, you won't be able to think about a linguistic task that the AI robots won't be able to properly accomplish.

And is there anything that humans can do that won't be tackled down very soon by AI? (It's of course what thinkers, artists and creative workers are wondering right now). Both humans and AI hallucinate all the time. But humans are much better on it, after a hallucination, going to cry in bed, ignoring them or running to post it on TikTok! It's not on the radar when or if  AI will ever tackle the issue of hallucination down. But this is what you want to keep an eye on: when the robots start dealing with their own hallucination, it's going to be a tough task for humans to linguistically detect a robot.

And is there anything beyond it? Long discussion, for another time :)

e7cc9e58d5173465715602ac64253bb380fa97a0f1976c998bc39d89fc3ad02a