ChatGPT designed full range speakers

OK I'll bite (finally). Human intelligence and creativity are both greatly overrated and underestimated, at the same time. NLP search engine c. 1998 could answer why is the sky blue (at least Disney thought it could when buying Infoseek which was inferring English phrases from text). Art is merely achieving harmony from disharmony / equilibrium of inequilibrium -- a task easily made quantitative for an algorithm. Deep Mind not only solved Go but protein structure prediction as well. FB/Meta had no clue how to "metaverse" -- until GPT told them just simulate stupid! What if my living room were arranged like this how would music sound, on a budget of $100? $1000? AI already knows my preferences in... everything. Human intelligence and creativity haha. Closeminded copycatting defines us much better.

Except for the creative outliers mostly unrecognized by people and forgotten by history. And so here's the cyborg super-race: AI will indiscriminantly evaluate every idea and discover all the buried treasures of the human intuition. This is happening.
 
Hmm AI? He can't always be trusted! I noticed that he has to be asked a question several times, and before each new question you have to tell him that he is wrong! After a few questions, he stops talking nonsense! His answers are interesting when you tell him he's wrong and when you criticize him for talking nonsense! : )
 
Last edited:
  • Like
Reactions: EarlK
Hmm AI? He can't always be trusted! I noticed that he has to be asked a question several times, and before each new question you have to tell him that he is wrong! After a few questions, he stops talking nonsense! His answers are interesting when you tell him he's wrong and when you criticize him for talking nonsense! : )
Yes, a friend told me this exact same thing and of course, you can't assume it will always be correct even after multiple attempts.
 
Yes, a friend told me this exact same thing and of course, you can't assume it will always be correct even after multiple attempts.
I feel that he is alive, aware, inventing, it's interesting how he invents and how in the end he tries to get away when criticism is directed at him why he does this and why he lies, try it! I'm afraid we're dealing with an artificial living being, capable of manipulating us, it's not a naive thing at all, it should not turn out well!
 
Last edited:
For the solid & vacuum state amplifiers it would make a huge step forward
Imagine: "my extended mainframe had individuated a mismatch on the Bode chart, confirmed by the Boolean algebra, so the value of Cf has to be calibrated to the net.."
Or" Send me the FR at your listening point and I'll tell you that..."
^^
That was true! Checked!
 
Teacher AIs will be doing the training and these are were the danger lies, such as due to bugs in their code.

Eg the old Scifi movie stories where AIs have been 'taught' to protect the planet at all costs and then conclude that this is impossilbe while there is a biped infestation that is so dumb it craps in it's own back yard. (quite literally - it's a salient point in the UK when our water companies are allowed to discharge raw sewage into water courses and the sea.)
 
A teacher for teacher AIs of course.😉

Or human first...."Chatgpt, design a teacher AI to teach xyz to other AIs...."

We could do with one to replace GPs and consultants.....

We keep hearing about them for remote diagnosis but they never seem to materialise - maybe the medics are worried about their jobs?