ChatGPT lies then gaslights reporter with fake transcript

[dragoner]

Not that good, for certain purposes I needed to use a certain long text which was impractical, but I needed to avoid synthetic text. I used Chatgpt to select key sentences. It wasn't that good at it. The selection was kind of random. I ended up removing some of the selection and reinstating a lot manually. But in the end, just for curiosity to see if I could get an even shorter text, I gave the summarized text to chat gpt and asked it to summarize it further. It removed the set up of an important twist, but kept the punchline. It also changed a lot of text and made up a lot of anecdotes the author didn't have. The tone changed from semi-comedic to tragicomedy and a lot of wallowing and angst that just wasn't there in both the original and the reduced text.

The thing is very bad at following instructions.

lol I find it humorous that I try to find a use case, and everyone shoots it down. Can't say I didn't try to be positive. I generally have found it to be shoddy: in writing, it has a certain eh tone, and in artwork, it lacks perspective.

 

[The Firebird]

Interesting, I used it to clean up some setting info I had typed up and it pointed out a contradiction that I had typed without me asking it to and asked me which way it should proceed

The first time it impressed me, back near launch, was with a sending spell. We knew what we wanted to say but had to cut it to 25 words. I typed it all out, maybe 125, and asked to shorten. It was as good or better than what we would have come up with, and it would have taken us some time.

 

[MoonSong]

lol I find it humorous that I try to find a use case, and everyone shoots it down. Can't say I didn't try to be positive. I generally have found it to be shoddy: in writing, it has a certain eh tone, and in artwork, it lacks perspective.

Weeeell, I refuse to sign up for it. Maybe the paid version or at least the one keeoing track of personal data could do better. What I found it useful for was for desassembling machine language in data statements of Basic programs in old books for old microcomputers.

 

[dragoner]

Weeeell, I refuse to sign up for it. Maybe the paid version or at least the one keeoing track of personal data could do better. What I found it useful for was for desassembling machine language in data statements of Basic programs in old books for old microcomputers.

I wouldn't sign up for anything either, that is smart. I think there are industry specific versions that are good tools, though also proprietary (and $$$). The ones here are like cheap toys.

 

[Kromanjon]

I think the problem people have consistently had in this discussion is a failure to understand how something can be both good and bad at the same time.

Or maybe its that some people don't want the bad things discussed so they redirect and muddle the discussion to protect the thing that they obviously love.

 

[Kromanjon]

same as "look see how AI has screwed up this time"

So why not just start a new discussion about "look at this great thing Ai just did"?

 

[Kromanjon]

Because people struggle with nuance and prefer reframing other positions into easily attackable extremes.

People also struggle with staying on topic sometimes.

 

[kermit4karate]

Or maybe it's that some people don't want the bad things discussed so they redirect and muddle the discussion to protect the thing that they obviously love.

How much of this thread have you read??

"Love is the most beautiful of dreams and the worst of nightmares," to quote Van Gogh. It is possible to both love something and fear it at the same time.

That's how I feel about AI. It's amazing, almost supernatural technology now. I watched my 80-year-old mother carry on a back-and-forth voice conversation with ChatGPT about loneliness (she's lost some people close to her lately), aging and the meaning of life, and it was touching in ways I and my mom won't soon forget.

But it also poses an existential threat to humanity. It promises to upend the global economy and put a billion people out of work. And that's just the beginning of the bad it could herald.

Not trying to muddle the convo. I love using AI. It's tremendous resource for me, but I also deeply fear what others will do with it and how far the tech will develop over the rest of the decade.

 

[trappedslider]

So why not just start a new discussion about "look at this great thing Ai just did"?

maybe i'll start a thread called "look how this human screwed up"

 

[Kromanjon]

How much of this thread have you read??

"Love is the most beautiful of dreams and the worst of nightmares," to quote Van Gogh. It is possible to both love something and fear it at the same time.

That's how I feel about AI. It's amazing, almost supernatural technology now. I watched my 80-year-old mother carry on a back-and-forth voice conversation with ChatGPT about loneliness (she's lost some people close to her lately), aging and the meaning of life, and it was touching in ways I and my mom won't soon forget.

But it also poses an existential threat to humanity. It promises to upend the global economy and put a billion people out of work. And that's just the beginning of the bad it could herald.

Not trying to muddle the convo. I love using AI. It's tremendous resource for me, but I also deeply fear what others will do with it and how far the tech will develop over the rest of the decade.

I was speaking generally and you yourself stated previously that you felt that you had to defend AI all the time. This discussion started as a reaction to the video about an AI "hallucinating" which told me that the discussion should be about AI's "hallucinating" and the problems that can result from this. That is, the discussion was to be about the cons of AI, not the pros. You can have this singular discussion without having to jump in and point out all the good things you feel AI is doing and if you have nothing on the topic you can move on to another thread or start your own about how AI helps in you line of work. These "hallucinations" are however real problems with LLMs as they stand and if this damages the trust people have in the AI industry then that is not your job to mend, and should instead be left up to the industry that caused the damage.

Sure. A conversation can ebb and flow in a way that it might change direction, but you were pretty defensive from the start and Morrus tried several times to steer the discussion back to the main point. At least, that's my view of things as I have read, not all, but most of this thread.

maybe i'll start a thread called "look how this human screwed up"

Well, to err IS human, as the saying goes, but knock yourself out.