3 ms·

In every HN thread there's always that armchair technologist who's so against critical thinking and the forethought of potential negative impacts of innovation,

In every HN thread there's always that armchair technologist who's so against critical thinking and the forethought of potential negative impacts of innovation, as if all innovation are purely good, no?

Madmallard 4y ago
Yeah I've watched a prominent psychologist talking about how the extent of it granting trust to vulnerable people and even regular people because of sounding smart makes it risky. I don't think it is a large risk with adequate training to the users. It's debatable not obviously one way or the other. I've also used it myself and found myself trying to make decisions based on its explanations only to find out that ChatGPT was mistaken. I personally do not think I am vulnerable to grave errors due to it and I've noticed large benefits using it for other purposes like coding or diagnostics. If you're trying to make the argument that "use computer algorithm make people dumb" then you're decades late on that one. Censorship should not be for everyone. Because if it is, it certainly isn't for its creators and powerful groups of people that bribe or coerce OpenAI which assuredly will happen regardless. This would be a strong argument for non-censorship. I don't think the decision to censor it was even done as a move to protect the public. I think it actually was a power play by the company with an easy alibi of seeming virtuous. This allows them the ability to charge much more for clandestine use. It could also be out of fear that it will spill secrets that they themselves are using or plan to use for their own gain. I think you could conjure several other valid reasons here.
enkid 4y ago
Asking why do or not do something is an important part of critical thinking.
- rTX5CMRXIfFG 4y ago
  So is thinking of edge cases on your own before asking the question. The article itself even lists some already, if the commenter only read it: > it is trained off of basically the entire internet without paying anyone involved > This is fantastic technology that can enable so many things to be so much easier, and yet I can see it being used for such evil at the same time. I just keep looking at this, and I wonder what happens when someone tries to use it to radicalize people.
  - xena 4y ago
    In case it is interesting, an early idea phase version of this article was called "So much for AI ethics" and it would have gone over that problem you quoted in more detail, but I was having trouble building up enough sarcasm juice to really express it properly.

Madmallard 4y ago

Yeah I've watched a prominent psychologist talking about how the extent of it granting trust to vulnerable people and even regular people because of sounding smart makes it risky. I don't think it is a large risk with adequate training to the users. It's debatable not obviously one way or the other. I've also used it myself and found myself trying to make decisions based on its explanations only to find out that ChatGPT was mistaken. I personally do not think I am vulnerable to grave errors due to it and I've noticed large benefits using it for other purposes like coding or diagnostics. If you're trying to make the argument that "use computer algorithm make people dumb" then you're decades late on that one. Censorship should not be for everyone. Because if it is, it certainly isn't for its creators and powerful groups of people that bribe or coerce OpenAI which assuredly will happen regardless. This would be a strong argument for non-censorship. I don't think the decision to censor it was even done as a move to protect the public. I think it actually was a power play by the company with an easy alibi of seeming virtuous. This allows them the ability to charge much more for clandestine use. It could also be out of fear that it will spill secrets that they themselves are using or plan to use for their own gain. I think you could conjure several other valid reasons here.

enkid 4y ago

Asking why do or not do something is an important part of critical thinking.

rTX5CMRXIfFG 4y ago
So is thinking of edge cases on your own before asking the question. The article itself even lists some already, if the commenter only read it: > it is trained off of basically the entire internet without paying anyone involved > This is fantastic technology that can enable so many things to be so much easier, and yet I can see it being used for such evil at the same time. I just keep looking at this, and I wonder what happens when someone tries to use it to radicalize people.
- xena 4y ago
  In case it is interesting, an early idea phase version of this article was called "So much for AI ethics" and it would have gone over that problem you quoted in more detail, but I was having trouble building up enough sarcasm juice to really express it properly.