Answer, No Answer, Or I Don't Know
-
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
Yes I think so, there is this metaphor of a stochastic parrot [https://dl.acm.org/doi/10.1145/3442188.3445922\](https://dl.acm.org/doi/10.1145/3442188.3445922) And there is also a discussion about this metaphor on Twitter ... :sigh:
-
I tried asking ChatGPT this, but it's at capacity right now.
Jeremy Falcon
Jeremy Falcon wrote:
I tried asking ChatGPT this, but it's at capacity right now.
It said this:
Quote:
No, AI would not respond with the wrong answer until it became "the" answer. AI systems are designed to provide accurate and reliable answers, and they would not be programmed to provide wrong answers. AI systems are programmed to use algorithms to analyze data and provide the most accurate answer possible.
Heh, and I believe it? Nope!
Latest Article:
SVG Grids: Squares, Triangles, Hexagons with scrolling, sprites and simple animation examples -
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
Yes. False answers are much harder to correct than to take a true answer and corrupt it to a false one that everyone believes.
Latest Article:
SVG Grids: Squares, Triangles, Hexagons with scrolling, sprites and simple animation examples -
Jeremy Falcon wrote:
I tried asking ChatGPT this, but it's at capacity right now.
It said this:
Quote:
No, AI would not respond with the wrong answer until it became "the" answer. AI systems are designed to provide accurate and reliable answers, and they would not be programmed to provide wrong answers. AI systems are programmed to use algorithms to analyze data and provide the most accurate answer possible.
Heh, and I believe it? Nope!
Latest Article:
SVG Grids: Squares, Triangles, Hexagons with scrolling, sprites and simple animation examplesIt’s good at sales at least. But we know how that goes. =D
Jeremy Falcon
-
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
Is that why ChatGPT keeps claiming
6 × 9 == 42? :-D
"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer
-
The next war, is fought by developers and hardware experts. And all it need to do is blow up one factory, and we loose 10 years of development. ONE factory, to throw us back 10 years. ASML. Right now they run Barter town, but then, we already know how that one ends :)
Bastard Programmer from Hell :suss: "If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
Master Blaster run Bartertown. Mad Max Beyond Thunderdome (1985) - Master Blaster Scene (3/9) | Movieclips - YouTube[^]
-
42
Master Blaster run Bartertown. Mad Max Beyond Thunderdome (1985) - Master Blaster Scene (3/9) | Movieclips - YouTube[^]
-
If there was an obscure question that had one (wrong) answer, and people (somehow) started asking that same question, would AI respond with the wrong answer until it became "the" answer?
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
This is the answer ChatGPT gave to me: "As a machine learning model, AI is not capable of intentionally providing incorrect answers. AI is designed to make predictions based on patterns it has learned from the data it was trained on, and it will provide the answer that it believes to be the most accurate based on that information. If the model has been trained on inaccurate data or has not been trained sufficiently, it may produce incorrect answers. However, this would not be intentional, but rather a result of limitations in the training data and the algorithm. It's important to ensure that the training data used to build an AI model is of high quality and diverse, in order to minimize the risk of producing incorrect or biased outputs. Additionally, regular testing and evaluation of AI models can help to identify any inaccuracies or biases in their predictions, allowing them to be corrected over time."
-
This is the answer ChatGPT gave to me: "As a machine learning model, AI is not capable of intentionally providing incorrect answers. AI is designed to make predictions based on patterns it has learned from the data it was trained on, and it will provide the answer that it believes to be the most accurate based on that information. If the model has been trained on inaccurate data or has not been trained sufficiently, it may produce incorrect answers. However, this would not be intentional, but rather a result of limitations in the training data and the algorithm. It's important to ensure that the training data used to build an AI model is of high quality and diverse, in order to minimize the risk of producing incorrect or biased outputs. Additionally, regular testing and evaluation of AI models can help to identify any inaccuracies or biases in their predictions, allowing them to be corrected over time."
-
Is that why ChatGPT keeps claiming
6 × 9 == 42? :-D
"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer
-
Yes. But so do humans, and not just for obscure stuff. 40 countries help Turkey; 3 countries help Syria. What do you think "religion" is? Ghosts, Angels? Ask ChatGPT for me which religion is "correct" and post me the answer :D If you want a real riot, ask it who blew up the gaslines to Germany.
Bastard Programmer from Hell :suss: "If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
"There is no definitive answer to this question."
-
Gerry Schmitz wrote:
If there was an obscure question that had one (wrong) answer,...
Addressing that seriously...no. If it is that obscure then "right" is no longer meaningful because no one would ask it. And no one would care if it was wrong anyways.
What was obscure becomes clear. Your flat earth would have stayed flat.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
Yes. False answers are much harder to correct than to take a true answer and corrupt it to a false one that everyone believes.
Latest Article:
SVG Grids: Squares, Triangles, Hexagons with scrolling, sprites and simple animation examplesExactly. Momentum. Try and insert the "right" answer into that. Right is more a function of majority opinion.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
There are no answers anywhere online. There are only likely and unlikely hypothesis. So, basically, it doesn't matter. A wrong answer is still an hypothesis worth testing.
Not if it involves drain cleaner.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
This is the answer ChatGPT gave to me: "As a machine learning model, AI is not capable of intentionally providing incorrect answers. AI is designed to make predictions based on patterns it has learned from the data it was trained on, and it will provide the answer that it believes to be the most accurate based on that information. If the model has been trained on inaccurate data or has not been trained sufficiently, it may produce incorrect answers. However, this would not be intentional, but rather a result of limitations in the training data and the algorithm. It's important to ensure that the training data used to build an AI model is of high quality and diverse, in order to minimize the risk of producing incorrect or biased outputs. Additionally, regular testing and evaluation of AI models can help to identify any inaccuracies or biases in their predictions, allowing them to be corrected over time."
It's a worthy answer ... but too long for most. I also think, not everyone gets the same answer: testing the questioner is part of the "learning" process.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
It's a worthy answer ... but too long for most. I also think, not everyone gets the same answer: testing the questioner is part of the "learning" process.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
-
What was obscure becomes clear. Your flat earth would have stayed flat.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
I took obscure to mean of little interest. So for example 'what is that fungus on Will's toe'. Now if 'Will' is the King of England then many more people are going to be interested than if Will is just the guy that lives down the block from me. If many people are interested then there will be many answers for the AI to look at. If very few are interested then far fewer answers. But as well a wrong answer in that case means it has no impact.
-
Not if it involves drain cleaner.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
