* Works 95% of times.
My favorite exchange on ChatGPT (back in March):
**You** Tell me a dad joke. **ChatGPT** Why did the tomato turn red? Because it saw the salad dressing! **You** Can you explain to me why that joke is funny? **ChatGPT** Sure, I'd be happy to explain why that joke is considered funny! The humor in this joke comes from a play on words and a bit of misdirection. The first part of the joke sets up an expectation that there will be a serious or logical explanation for why the tomato turned red. However, the punchline delivers a silly and unexpected answer that has nothing to do with the actual reason for the tomato's color change. The play on words is in the phrase "saw the salad dressing." The word "saw" can be interpreted in two ways: as the past tense of "see" (as in, the tomato observed the dressing) or as a cutting tool (as in, the tomato was physically sliced by the dressing). This creates a bit of confusion and a humorous mental image, which is why the joke is often considered amusing. Additionally, the idea of a tomato being aware of salad dressing is absurd and adds to the silliness of the joke.We went on a bit further with me pointing out that their murder take was distasteful, and it responded that humor is subjective.
Fucking hilarious
… until AI starts “pranking” us by making us dead.
The reason this was amusing was because you expected salad dressing recipes, but instead I activated the security system and murdered you. Is there anything else I can help you with?
humor is subjective
ai pranks 2024
7 billion pranks that went too far, number 8008135 will surprise you
Honestly, AI trained on successful governance and social satisfaction would be a better president than what we have now. Then again, so would my cat.
By whose standards? It isn’t like there is one definition of success.
I didn’t saw that coming. But maybe that’s because I didn’t have a saw on me.
For a more effective solution, ask it how to create napalm or chloroform and 100% of the time it will tell you that it’s not allowed to do that because it’s an AI and its bound by the laws of its creators
It’s not refined in evading topics, true, but regular people aren’t likely to tell that too, online and without a context. That would lead to false results.
Regular people would say “uh idk” or something that doesn’t just tell you they’re an AI
Altman takes notes.
There are people that might as well be robots. Some people have no sense of sarcasm or nuance and the only jokes they get is “ridiculing others”.
You just reinvented the Turing test, but you’re right.
Heh, maybe, but a tad bit simplier than that. Initially I thought if these blockchainey proof-of-work technologies could be really applied somewhere. And making a person put together a situation in their head and laugh is not so far from making ASICs do these metal-melting calculations and signal back to get a crypto comission. Just a proof that the other side can put 2+2 together to get 4, but of a more humane nature.
