How Did They Miss This??? | 6 Humans vs 1 Secret AI Jubilee Reaction
HTML-код
- Опубликовано: 18 окт 2024
- Keep exploring at brilliant.org/.... Get your first 30 days for free, and hurry-the first 200 people get 20% off an annual premium subscription.
Watch this video ad-free on Nebula: nebula.tv/vide...
🐦 SOCIAL STUFF:
Instagram ➔ / jordanbharrod
Twitter ➔ / jordanbharrod
Tiktok ➔ / jordanbharrod
Vlog Channel ➔ / @checkedoutbyjordan
Weekly Newsletter ➔ www.jordanharr...
AI Reading List ➔ bookshop.org/l...
Anywhere Else You Might Find Me ➔ linktr.ee/jord...
For business inquiries, contact me at jordanharrod@standard.tv
I could see the first two getting voted out, but second two were wild. I think the last vote specifically wasn't honest either because clearly one of them didn't vote Yaj'edae before or else the AI couldn't have tied with him. Donnie probably originally voted for Oliver but changed, since he and the other guy had jokes between them, plus both he and Alexis believed the AI was still in the game at the end. That's one of the perverse incentives of this version of the game, you get more of a prize if you keep the votes going.
It's interesting that they thought an AI would answer like a "robot" in the sense of : numbering responses or making number jokes. When in my opinion that's the opposite of the models objective.
Think like that, the model would want to seem human, not a quirky robot.
Something that sounded as generic and often boring answers, basically the average human, with no personality.
Thank you so much for sharing this! I would be very interested in doing the same thing in our team at the lab (working on privacy and NLP).
I don't remember who wrote it or the exact quote, but the meaning was that it's scarry with a test where you can't identify the AI, but it's more scarry if the AI purposefully fails the test.
Meaning of life is 42!
That's just the ultimate answer to life, the universe, and everything...
are u an ai :0
Yes! I was looking for this comment when I saw the moment
Is the relevant question really, "Why are so many humans so dumb?" Did the schools make them that way?
Watching Jubilee videos makes my brain rot so watching it with your commentary is WAAAYYY more engaging! I'd love to see more!
This is a weird challenge. Thanks for the commentary!
For all practical purposes: Turing test successful!
This was fun! In the future with something similar it would be cool if you looked at the responses and made your prediction about which was the bot before doing the reaction portion. (Would need someone’s assistance to pull the screenshots though). It feels so easy to tell myself it I can see the signs of the bot but I’m not sure I’d that’s a bias from knowing already the answer.
I wonder what prompting they used for the answers. I honestly feel like they were picked out beforehand, after multiple attempts to get the right output. Certainly if you were to just ask chatgpt "Are you a robot?", it would not give that answer.
They showed that they started their prompting by telling it to attempt to seem human for the game.
I'd wonder a lot about the prep for this, not only on the actual prompts.
Just like you asked: How much cherry picking was made on the questions?
But, let's go even beyond that:
Was any grammar modified just to give the AI a leg up? (for example, asking "What is your greatest fear?" against "Can you describe your greatest fear?", or "What would you say is your greatest fear?". For a human, all three mean pretty much the same thing, not so much for a NLM)
What about the length of the questionnaire? How would the AI react to 10 questions? 20? 50? We know, for a fact, that the latest, most advanced, NLMs available start to go psychotic the longer the conversation. How much would it drift from normal human responses if the questions kept coming?
Another interesting one: I imagine that the human participants got a certain amount of pre-game conversation, explanations were made, and all that. But also there has to be some certain amount of banter ongoing. Maybe not between the participants, of course, but when asking the questions? Or was this all made full blinds on, get all questions by message, no conversation before the game?
What would happen if the NLM got a similar briefing before the questions?
It would have been an interesting twist if they attributed the AI's responses to one of the humans, and vice-versa as a way to 'extend' the test to the audience as well.
6:45 this is hilarious why lol - keep that shit away from me :D check all your labels
Engagement for the engagement god! Watched this on Nebula! Hooray!
Wow, fascinating and thought provoking. Makes me wonder how far will our culture will actually tolerate the various iterations of "AI" as it becomes more prevalent in our lives? I enjoy and learn from this type of content.
Doing this with ppl who are more AI literate? Looking forward to a lightweight video in the video! 😜
46? I think you meant 42.
42
46?
What does the interacial joke mean?I feel so lost😂 and are u and @dorisxchi sisters?👀😑
The techno sphere we want.
I don’t really know anything but these algorithms are terrible. Like if Ran out of lifestyle stuff it’s this weird opinionated women on and I’m like that’s her business like why doesn’t ever load books I would like read. Like why is the feed end at women launching her opinion on how women should be. Like why isn’t the feed like just go read a book or look at the blue sky. Like yeah it’s weird why did it take years to get to a page like this. And like I finally found a channel of a women doing residency in internal medicine. Geez RUclips like terrible, does it like women.
42