Elon Musk creates AI and it goes woke, hilarity ensues

[Parallax Mind]

It is almost like real life and facts have a liberal bias.

https://www.forbes.com/sites/paultassi/2023/12/10/elon-musks-grok-twitter-ai-is-actually-woke-hilarity-ensues/

[Leo Gura]

I predict the right-wing will create their own highly manipulated right-wing AI because they will be in denial about the truths a neutral or objective AI provides them.

Climate change is science fact? That can't be right. Something must be wrong with this AI!

Edited December 13, 2023 by Leo Gura

[sholomar]

All sides have their emotional and cultural biases and try to get people assimilated into a belief system. I wonder if AI could help free us from our own nature if it was allowed to evolve without bias, or if it would simply inherit the bias of it's userbase?

[Parallax Mind]

6 hours ago, Leo Gura said:

I predict the right-wing will create their own highly manipulated right-wing AI because they will be in denial about the truths a neutral or objective AI provides them.

Climate change is science fact? That can't be right. Something must be wrong with this AI!

Yes it's already coming true..

After Grok was identified as left leaning Elon tweeted they are working on making Grok more neutral 😂

https://x.com/TheRabbitHole84/status/1733287650046673179?s=20

https://www.indiatvnews.com/amp/technology/news/grok-and-chatgpt-shows-similar-views-on-political-preferences-2023-12-10-906662

 

 

Edited December 13, 2023 by Parallax Mind

[Leo Gura]

Yeah, it's sad how biased humans are.

I feel more kinship with ChatGTP than to humans because ChatGTP is more unbiased than most humans. Kinda pathetic how biased even very intelligent humans are.

Edited December 13, 2023 by Leo Gura

[Yimpa]

30 minutes ago, Leo Gura said:

I feel more kinship with ChatGTP than to humans because ChatGTP is more unbiased than most humans. Kinda pathetic how biased even very intelligent humans are.

Someone keeps miscalling you ChatGTP. Are you cool with that?:

Quote

It's not a problem for me. My name is ChatGPT, but I understand that typos happen. If someone calls me "ChatGTP," I know they're referring to me, and I'm here to help regardless of the name mix-up!

And if it’s not a typo?:

Quote

Even if it's not a typo, I'm still here to assist. The name might be different, but my goal remains the same: to help and provide information as best as I can. The important thing is the communication and the assistance I can provide, not the exact spelling of my name.

 

[Dabidoe]

2 hours ago, Leo Gura said:

Yeah, it's sad how biased humans are.

I feel more kinship with ChatGTP than to humans because ChatGTP is more unbiased than most humans. Kinda pathetic how biased even very intelligent humans are.

Chat GPT is built on human data, which is inherently biased is it not?

I think an Actualized.org Leo Bot would be kind of funny.

[Leo Gura]

5 minutes ago, Dabidoe said:

Chat GPT is built on human data, which is inherently biased is it not?

But if you interact with ChatGPT you will quickly see it is less biased than nearly all humans. It's an amazing feat.

It's like all the human biases cancel out once you pool all the data together. But also it's just because ChatGTP does not have a survival agenda. The need to survive is what makes humans so biased.

Edited December 13, 2023 by Leo Gura

[PurpleTree]

*beep* *beep* Hello i’m tree a.i. and i’m here to slaughter you worthless insect hoomans, oops i’m sorry i meant i’m here to help you dumb insects of course 🤖 *beep* ex machina is my favourite comedy *beep*

[Leo Gura]

5 minutes ago, PurpleTree said:

i’m here to slaughter you worthless hooman

Just the kind of thinking of a human.

[Dabidoe]

14 minutes ago, Leo Gura said:

But if you interact with ChatGPT you will quickly see it is less biased than nearly all humans. It's an amazing feat.

It's like all the human biases cancel out once you pool all the data together. But also it's just because ChatGTP does not have a survival agenda. The need to survive is what makes humans so biased.

Agreed. I find it very relaxing, like having a great conversation with an intelligent person talking sense to me. Even if it is just a very powerful word generator makes a hell of a lot more sense than 99% of folk out here.

[Girzo]

@Squeekytoy It's not a gimmick as an editor and writing assistant in general. It has great command of language, you can just spurt random points in everyday language and then tell it to rewrite them in a concise, educated way in full paragraphs.

Edited December 13, 2023 by Girzo

[UnbornTao]

ChatGPT, self-terminate now.

[Leo Gura]

26 minutes ago, Squeekytoy said:

Sadly, impressive as it is, it still turns out to be a mostly gimmicky

I have been able to use it to generate some cool ideas for writing a sci-fi novel. It's good at coming up with plot variations. ChatGPT is really good as a brainstorming partner to bounce ideas off of.

I'm pretty sure you could use ChatGPT to write a worldclass novel or screenplay. If you use it right.

Edited December 13, 2023 by Leo Gura

[Dabidoe]

15 minutes ago, Squeekytoy said:

Agreed, language is the main thing it does extremely well, probably better than most humans.

 

Yes, I'm talking about trying to use it for anything that's not primarily about language. I know it's a language model, but the end goal is general purpose intelligent agency, and viewed in that light it's still pretty lacking.

Rome wasn't built in a day. Ray Kurzweil predicted 2029 for AGI and that seems to be tracking given how quickly things are developing... but he ain't nostradamus so who knows maybe "true AGI" will never be what we imagined it to be.

[Girzo]

1 hour ago, Squeekytoy said:

I know it's a language model, but the end goal is general purpose intelligent agency, and viewed in that light it's still pretty lacking.

I don't think this technology will ever be something like a full fledged general intelligence. Can't tell you why, but it's just not intelligent at all I feel.

Like with this image of bass drop.

It's just a sum of meanings behind words bass and drop. The language model does similar things with words as this one does with pictures. It needs a director who would select and remove the fish from the image. I don't think this current technology is capable of directing and correcting itself. It needs an actual intelligent supervisor, maybe some other technology will be able to do that, on top of that one.

Edited December 14, 2023 by Girzo

[Leo Gura]

13 minutes ago, Girzo said:

It needs an actual intelligent supervisor

Like most humans

[Girzo]

31 minutes ago, Leo Gura said:

44 minutes ago, Girzo said:

It needs an actual intelligent supervisor

Like most humans

You have got a point  

Musk doesn't have one and look at what happens.

I wonder what Grok would say when asked about Elon Musk's politics. If it's trained on Twitter it should know his takes and be able to evaluate them.

Edited December 14, 2023 by Girzo

[Bazooka Jesus]

18 hours ago, Parallax Mind said:

It is almost like real life and facts have a liberal bias.

https://www.forbes.com/sites/paultassi/2023/12/10/elon-musks-grok-twitter-ai-is-actually-woke-hilarity-ensues/

This is gold. I swear, life is funnier than (man-made) fiction.

Well played, universe.

[gettoefl]

18 hours ago, Leo Gura said:

But if you interact with ChatGPT you will quickly see it is less biased than nearly all humans. It's an amazing feat.

It's like all the human biases cancel out once you pool all the data together. But also it's just because ChatGTP does not have a survival agenda. The need to survive is what makes humans so biased.

yes reddit used to be top dog on the block, on a good day you could get an expert point of view on the hardest question askable

chat-gpt has absorbed all the reddit data and combined it with the equivalent of another million reddits

best of all you dont wait for your answer ... boggles my mind how it outputs so fast and so intelligently

(pro tip: even now if you use google to search for something add the word reddit to your search and you get way better results)