4 ms·

My stance is pretty simple. The folks that adapt their own language centers and domain reasoning around using chatGPT (or these types of models) will stand to

My stance is pretty simple. The folks that adapt their own language centers and domain reasoning around using chatGPT (or these types of models) will stand to gain the most out of using them. This article is an eye roll to me, a calculator gives you confidence as well, doesn't mean you used it correctly. It is very hard for me to not outright dismiss articles like this that don't consider the usefulness of the tool. They instead search for every possible way to dismiss the tool. >My conclusion isn’t just that ChatGPT is another con game—it’s the biggest one of them all. * YAAAAAWN *

bsaul 4y ago
You're confounding the problem of feeding a system with incorrect inputs ( calculator ), with a system that, provided the input are correct, will throw garbage totally at random. This is a very different kind of system with very different kinds of problems caused to its users.
- worik 4y ago
  And you have missed the point. To be a liar requires creativity This a computer. Creative. This is huge. The most astounding thing is its ability to con you. We will learn to use it. Hydraulics for cognition. Amazing
  - wruza 4y ago
    It doesn’t lie, it just picks whatever static noise there is after training and converts it into a coherent speech. It is as creative as few TB of /dev/random sprinkled with encrypted facts and decrypted back as a whole.
- gunshai 4y ago
  I think my caveat of domain knowledge side steps this issue you bring up. It's the entire reason engineering schools teach the "long way" when you're going to rely on the "short way" 99.99% of the time. Also the same could be true for digital calculators when first created. The neural nets (aka humans) that were initially trained to design the system can take short cuts for example on improperly returning or applying functions such as square roots ect. It still confidently gives you the incorrect answer. Given sufficient training data (IE the human) the system improves to return better more accurate results.
fortituded0002 4y ago
I have the same overall reaction. But suspect that your analogy with the calculator to distract some people reading this. Difference being that one is user input error and the other is either ChatGPT misunderstanding what's being asked or just lacking training data and presenting an answer that's incorrect. By yes, an eye roll from me as well. A few months back I heard the horror stories about how a bot answers with confidence and now it's the main complaint in articles about why it's busted and dangerous. It doesn't bring anything new to the table and doesn't push the conversation forward in any way.
- gunshai 4y ago
  > misunderstanding what's being asked or just lacking training data and presenting an answer that's incorrect. I suppose I just don't think we humans are so different. In fact we often lack training data and certainly lack the ability to iterate quickly. In the case of the modern calculator we have the benefit of all the training data necessary to design the system properly, but at its initial inception not so much. As more "training data" or experience with circuit design and applied mathematics the returned output of the calculator improved. Maybe my expansion of the analogy is off or too esoteric.

bsaul 4y ago

You're confounding the problem of feeding a system with incorrect inputs ( calculator ), with a system that, provided the input are correct, will throw garbage totally at random. This is a very different kind of system with very different kinds of problems caused to its users.

worik 4y ago
And you have missed the point. To be a liar requires creativity This a computer. Creative. This is huge. The most astounding thing is its ability to con you. We will learn to use it. Hydraulics for cognition. Amazing
- wruza 4y ago
  It doesn’t lie, it just picks whatever static noise there is after training and converts it into a coherent speech. It is as creative as few TB of /dev/random sprinkled with encrypted facts and decrypted back as a whole.
gunshai 4y ago
I think my caveat of domain knowledge side steps this issue you bring up. It's the entire reason engineering schools teach the "long way" when you're going to rely on the "short way" 99.99% of the time. Also the same could be true for digital calculators when first created. The neural nets (aka humans) that were initially trained to design the system can take short cuts for example on improperly returning or applying functions such as square roots ect. It still confidently gives you the incorrect answer. Given sufficient training data (IE the human) the system improves to return better more accurate results.

fortituded0002 4y ago

I have the same overall reaction. But suspect that your analogy with the calculator to distract some people reading this. Difference being that one is user input error and the other is either ChatGPT misunderstanding what's being asked or just lacking training data and presenting an answer that's incorrect. By yes, an eye roll from me as well. A few months back I heard the horror stories about how a bot answers with confidence and now it's the main complaint in articles about why it's busted and dangerous. It doesn't bring anything new to the table and doesn't push the conversation forward in any way.

gunshai 4y ago
> misunderstanding what's being asked or just lacking training data and presenting an answer that's incorrect. I suppose I just don't think we humans are so different. In fact we often lack training data and certainly lack the ability to iterate quickly. In the case of the modern calculator we have the benefit of all the training data necessary to design the system properly, but at its initial inception not so much. As more "training data" or experience with circuit design and applied mathematics the returned output of the calculator improved. Maybe my expansion of the analogy is off or too esoteric.