I, ChatGPT
-
Something I’ll wager ChatGPT would have trouble doing, in either iteration.
-
@Arantor said in I, ChatGPT:
Something I’ll wager ChatGPT would have trouble doing, in either iteration.
It's learned from social media, so putting a spin on things is inherent to the model.
-
..
-
@Gustav I mean, the new one is a different hot take on the same thing. The older one is obviously “what did you SEE?” without thinking about the parasocial implications, and was long before GitHub attempted to position itself as a sort of social network.
Neither made me laugh (not even at the time) but I got what they were going for, recognised having had both journeys, and been the person in the comic.
-
So there's a new AI-related existential threat in the wild, albeit not of the kind Big Yud warns about. A friend's wife, who works in one of these big "prestigious" consultancies, has a new junior coworker. This coworker uses some kind of GPT to write his code, and when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'. They go back nad forth all day, and she's extremely annoyed and considers quitting.
-
@sebastian-galczynski this isn’t a new threat but one that was certainly about to bite people in the face.
The magic box cannot ever be questioned. And when it should be questioned, it must not be questioned.
-
@sebastian-galczynski said in I, ChatGPT:
when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'.
Makes a change from "it can't be, it was upvoted on Stack Overflow" I guess.
-
@sebastian-galczynski said in I, ChatGPT:
So there's a new AI-related existential threat in the wild, albeit not of the kind Big Yud warns about. A friend's wife, who works in one of these big "prestigious" consultancies, has a new junior coworker. This coworker uses some kind of GPT to write his code, and when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'. They go back nad forth all day, and she's extremely annoyed and considers quitting.
One would think that juniors like that would be extremely susceptible to the "Him or Me" card.
-
@sebastian-galczynski said in I, ChatGPT:
it can't be, GPT said it's correct
"So why do we pay you instead of ChatGPT? Bye!"
-
@dkf said in I, ChatGPT:
@sebastian-galczynski said in I, ChatGPT:
when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'.
Makes a change from "it can't be, it was upvoted on Stack Overflow" I guess.
At least Stack Overflow can be argued with.
ChatGPT can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead.
-
@Arantor said in I, ChatGPT:
@dkf said in I, ChatGPT:
@sebastian-galczynski said in I, ChatGPT:
when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'.
Makes a change from "it can't be, it was upvoted on Stack Overflow" I guess.
At least Stack Overflow can be argued with.
ChatGPT can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until
you are dead.your desiccated corpse can no longer provide enough bioenergy for sustenance.
-
@Arantor said in I, ChatGPT:
@sebastian-galczynski this isn’t a new threat but one that was certainly about to bite people in the face.
The magic box cannot ever be questioned. And when it should be questioned, it must not be questioned.
Friend Computer
-
@Applied-Mediocrity said in I, ChatGPT:
@Arantor said in I, ChatGPT:
@sebastian-galczynski this isn’t a new threat but one that was certainly about to bite people in the face.
The magic box cannot ever be questioned. And when it should be questioned, it must not be questioned.
Friend Computer
The Machine Spirit.
-
@dkf said in I, ChatGPT:
@Applied-Mediocrity said in I, ChatGPT:
@Arantor said in I, ChatGPT:
@sebastian-galczynski this isn’t a new threat but one that was certainly about to bite people in the face.
The magic box cannot ever be questioned. And when it should be questioned, it must not be questioned.
Friend Computer
The Machine Spirit.
-
The sound of AI articles falling in the woods: https://www.espn.com/espn/story/_/id/39344315/sports-illustrated-plans-major-layoffs-name-license-revoked
-
@Zecc said in I, ChatGPT:
@topspin said in I, ChatGPT:
is just a much weaker, unfunnier, less emotional version of the age old one.
At least it's not the exact same joke. He put a new spin on it.
-
@HardwareGeek said in I, ChatGPT:
@Zecc said in I, ChatGPT:
@topspin said in I, ChatGPT:
is just a much weaker, unfunnier, less emotional version of the age old one.
At least it's not the exact same joke. He put a new spin on it.
All right, don't get your knickers in a twist.
-
I was going to make some joke about how DPD's AI had become sentient and experienced life just like the rest of us but it just turns out someone was "prompt engineering" and demonstrating the lack of safety rails again.
Well, fuck. Non-story after all.
-
@Arantor what I like about these stories is how they demonstrate that these things really are just advanced Markov generators. It's why :@Gustav:'s ideas about "putting guard rails on" can't work, because no one has yet made a program that can think (and that's way before you get to all the ambiguous stuff that humans can't agree on when looking at a statement).
Not that I expect these examples to convince people. The illusion is much more convincing, just like so many other places in life where we fail to think critically.
-
-
-
@Arantor said in I, ChatGPT:
@dkf said in I, ChatGPT:
@sebastian-galczynski said in I, ChatGPT:
when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'.
Makes a change from "it can't be, it was upvoted on Stack Overflow" I guess.
At least Stack Overflow can be argued with.
ChatGPT can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead.
It brings me no end of saddness that skynet won’t give me a quick death but AI garbage will probably give me a stroke that will render me retarded.
-
@DogsB said in I, ChatGPT:
@Arantor said in I, ChatGPT:
@dkf said in I, ChatGPT:
@sebastian-galczynski said in I, ChatGPT:
when it doesn't work, or does the wrong thing, he argues that 'it can't be, GPT said it's correct'.
Makes a change from "it can't be, it was upvoted on Stack Overflow" I guess.
At least Stack Overflow can be argued with.
ChatGPT can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead.
It brings me no end of saddness that skynet won’t give me a quick death but AI garbage will probably give me a stroke that will render me retarded.
The “want to move to Canada” thread is
-
@DogsB said in I, ChatGPT:
@LaoC said in I, ChatGPT:
The rocket ia bananana sized.
That’s a pretty big fox.
-
@DogsB said in I, ChatGPT:
That’s a pretty big fox.
Rocket-sized if I'm not mistaken, given the presence of a standard measuring banana in both pictures...
-
@dkf I question the relative sizes of banana; the first one is clearly a standard size banana but the latter is abnormally large - I never saw a banana 20cm across.
-
-
@Arantor said in I, ChatGPT:
And it absolutely will not stop, ever, until you are dead.
But the results after a few hundred token iterations are pretty amusing, for a while.
-
@izzion said in I, ChatGPT:
One would think that juniors like that would be extremely susceptible to the "Him or Me" card.
Either that or the "let him deal with consequences" card.
If his code is somewhat limited to one module (feature, page... whatever makes sense in that context), document that you said his code is wrong but let him have it. And, crucially, let him deal with user complaints, and see at which point the penny drops. Or the boot drops on his backside, if the penny doesn't drop fast enough.
Of course both cards are ideal-world thing, just missing the "and everyone stood up and clapped." In the real world, it's rarely as simple as that.
Still, I'm not surprised that some people get swindled by the AI hype, and even less for juniors who don't know better yet (and I'm not surprised either some might be clueless/stubborn about it, that is human nature). But I'm not really worried about it long-term. Like Stack Overflow, the Web, Usenet and countless things before it, people will learn the limits quickly -- the hard way if not the smart way.
-
@Arantor said in I, ChatGPT:
DPD
The other day my wife was waiting for a delivery from GLS and waited the whole day for them because they couldn't be more accurate than "between 8:00 and 18:00" and then when they didn't deliver without any notice my wife had to go to their website to reschedule for the next day, again between 8:00 and 18:00 and again had to accommodate her schedule to be there all day and then they again didn't deliver except that when I came back home in the evening and picked up the mail it turned out that they had just plunked the package in the mailbox, didn't bother to ring the bell nor to update their website (it was updated the next day only to say "package delivered") nor send a notification.
So anyway, my wife was fuming and went to their website to complain, and nervously laughed seeing their slogan of "we're here for you" and saying it rather was the opposite.
Also I think she put a low-stars review to the product on Amazon with a comment about how shitty the delivery was. I hate those kind of reviews because they're not about the product but since it's the only way there is to give a feedback to the seller...
</rant>
-
@HardwareGeek said in I, ChatGPT:
@DogsB said in I, ChatGPT:
will render me retarded.
Will? Future, not present?
A better question is why posts I make when I might be rendered retarded get more upboats?
STATUS reading through a flamewar in another community about AI.
Someone did raise the same interesting points though.
A: You don't have google search engineers so why would you have prompt engineers?
B: With the mid-senior level engineers automating all the crud work where are the juniors going to be trained to replace them?He'll probably be banned soon.
-
-
@DogsB said in I, ChatGPT:
You don't have google search engineers so why would you have prompt engineers?
You think you don't, but you do. It's just that they don't call themselves that.
-
@Zerosquare said in I, ChatGPT:
@DogsB said in I, ChatGPT:
You don't have google search engineers so why would you have prompt engineers?
You think you don't, but you do. It's just that they don't call themselves that.
At this stage, it probably falls into the category of life skill. Like cooking for yourself or managing your money... We're fucking doomed.
-
STATUS
I'm watching a stream from a youtuber about silent hill ascension.
I don't think it was written by AI but I plugged the premise into chatgpt and it spat out something better.
-
@DogsB you can do that for almost any subject on YouTube.
-
-
@Arantor said in I, ChatGPT:
@DogsB you can do that for almost any subject on YouTube.
Silent Hill Ascension is so fucking terrible I'll give AI the benefit of the doubt on this one. It's somehow worse than the movies. Like how is that possible? Suffice it to say, I've enjoyed every minute of it.
-
-
@remi said in I, ChatGPT:
delivery from GLS
I only use them when the package is heavy, because on most shops they are the only ones who accept over 15 kg, and, with either delivery company, I almost always use a pick-up point—so I don't care if it arrives between 8:00 and 18:00, I pick it up on the way home.
-
@remi said in I, ChatGPT:
Still, I'm not surprised that some people get swindled by the AI hype, and even less for juniors who don't know better yet
The bigger problem is the s falling for it.
-
@Bulb said in I, ChatGPT:
I almost always use a pick-up point
We do have that and I quite like it. There's always a shop not too far (or on my road to somewhere) and they have wide opening hours. When I'm directly buying from an online shop and it has this option, I often pick it.
The main problem though is with online shops where there are several sellers (typically Amazon), where you don't have any control on the delivery method. This is how we ended up with GLS and I think this is at least the second time we've seen how awful they are, so while I'd like to avoid them as much as possible, it's often not really possible. Not only you can't pick an alternative, but you can't even know they'll be the ones in charge of delivery until you've paid for the order.
-
@remi If I can't choose the delivery method, I go search for a better shop that has the thing and does allow choosing the delivery method.
-
-
Status: I'd like to think I've come pretty far in my text generation capabilities.
-
@Tsaukpaetra No quack.
-
@dkf Go is to the fountain.
-
@Arantor The pig put foot. Grunt.
-
I've noticed that their emails have gotten longer. It used to be five badly written paragraphs. Now it is seven better-written paragraphs. Bitch! I don't read!
Around 11pc of people who were targeted in the test fell for an AI-generated email that took five minutes to produce. Emails crafted by humans just outperformed the bots, tricking 14pc of targets.
-
Waiting for it to make shitposts for me.