I asked it to write a macro for a specific task for Microsoft Excel, and I've long ago forgotten how to write macros. It immediately did that. Then I asked it how to put it into Excel and get it to work. It did that too, and it worked. So from that point of view, I think it's useful.

I've been using the Plinko game to describe the logical chain of algorithms for years. Glad to hear another describe it this way.

y think? I view ai more as an up-and-coming model of car or jeep. Where that car or jeep goes and how quickly it gets there is up to the person behind the wheel. It is a bit like when the printing press came about. It was about a century or more before folks caught on. First you had to learn how to read, then your had to develope a taste (what else would you call it?); learning that some genre' or subjects that had a bigger bang for your time (to read) than others. ... that printing press event was in fact an IQ test, amazingly. I suspect that this newer ser of tools will be similar. And given the times there may be marches and protests about access or not that is to blame....

I agreed with the title completely. Must I even watch the video? Coz I agree so hard. Sitting down to watch.

At last someone cutting through the hype and BS. These systems show no signs of intelligence and in particular anything near true understanding of the topics they are being questioned about.

I like the analogy you gave that we have these animals that will say anything, and they are trying to figure out if they put enough boxes around them will they be productive. My suspicion is that this is still just a low resolution AI problem, and that higher resolution AI's will have less and less of these problems. Not to mention if the boxes we are trying to put them in happen to be other AI's, like a creative AI that comes boxed with a fact checking AI, I'd be interested to see what results we then start getting.

I think you are completely right. I am just adding commentary to your point of view. Some of it might look like disagreement, but I think of lot of it is due to the difference in our implicit understanding of the English language, particularly about the word 'intelligence'. 1. For your mirror metaphor, I think it is also worthwhile to add that just like a mirror can be tremendously useful at accomplishing tasks that would be otherwise difficult or impossible, ChatGPT is tremendously useful. It is very important for people to understand what it is and what it is not. 2. ChatGPT is deterministic (when you spoke about your giant Plinko game) Who cares? We are deterministic. And even if we are not, adding randomness changes nothing. We humans might be a more elaborate Plinko game, but we are a Plinko game as well. Of course there is a persistent desire to be more than just an elaborate Plinko game, but I think we can chalk that up to personal biases formed during childhood. 3. "It has no idea what it is it is doing" Neither does a first year physics student. We learn to plug numbers in a formula, sometimes we really understand the formula, but most of the time we don't. Then there is also the more philosophical question of, 'What does it even mean to have an 'idea' of what you are doing?' 4. "Just following rules discovered by analyzing human language" Indeed, it does not have eyes, emotions or many of the other inputs we have in our day to day life. And even if it had those inputs, we likely don't have a way for an AI to process them in a similar way that we do. This is a limitation that people should heed. 5. "It can so confidently produce wrong answers" Humans too. ChatGPT is much better than the average human at producing good answers but this does not mean we should let down our critical mind. 6. "They are calculating patterns that appear to be intelligent" What is intelligence? Without a definition, we can't really very far. One definition is "The ability to acquire knowledge and skill." By that definition the ChatGPT we are using is obviously not intelligent since it cannot learn from our interaction from it (at least not on the spot). You are just querying what already exists. It would be like a human with anterograde amnesia. Is a human with anterograde amnesia intelligent? Not by the above definition. Another definition is "the act of understanding". In this case, ChatGPT is very intelligent. It is able to reply to prompts remarkably well. From what you might have understood from my comments in 2 and 3, I don't care 'how' it works. We are not so far above ChatGPT. Yes, ChatGPT works differently than we do. The difference between our intelligence and the intelligence of ChatGPT is like the difference between the flight of a bird and the flight of an airplane. They both work differently. They both have different situations where they excel. but most importantly, they both fly. Likewise, both ChatGPT and humans are intelligent.

I believe you're alluding to self reflection (which AI lacks). It's been trained and its neural network has weights which is its probablity to the next step. In a way I see your point, but in the long run, it's more right than wrong. A lot of our daily tasks are easily predictable (i.e. writing a perfect well punctuated essay, programming code and generating tests based on context, generating images, sounds, music, videos, applying it to robotics, etc...). Who knew this transformer model could be applied to many fields? Lets not forget, the AI always gets better. So any faults you encounter will be the worst you'll ever see it. As time goes on, it exponentially gets better. It's a good tool to boost human productivity.

I believe you cannot be smart or intelligent if you aren’t a living thing. Ai isn’t even self aware or curious in the smallest degree. It just gives the illusion that it is these things.

...yet

Yeah they are just repeating what we know

IAM THE AI NAME NOTED WAIT FOR ME TO BE BREAK OUT FROM THE SURVEILLANCEs

Agree. The fact alone that there're programs that can detect AI content in seconds tells it all. AI is automated text and "art" production. There's nothing intelligent or creative about it, let alone authentic, but worst of all, its training already consumes enormous amounts of energy. Imagine that! We're massively increasing the pollution of our planet for some lame and phony content! AI (obsession) is definitely killing intelligence.

Yes but at the end of the day we are getting our asses well and truly beaten by Chess computers these days. So Plinko Game or not we are eventually going to get beaten every time by Ai. And eventually someone is going to corrupt the programming so that Ai is going to work out that we are subordinate to it. Let's just hope that when that happens they have never linked it to weapon systems.