Merged Artificial Intelligence

[Dr.Sid]

It doesn't quite fire up the right ones unless I know I'm playing chemical ping pong. Even gamers prefer other people to bots, if only to picture the other person's angry tantrum.

I totally do not prefer other people to bots.

 

[Olmstead]

Speak for yourself, I hate multiplayer. If I wanted to engage with people like that I'd play a sport. I'm perfectly happy pushing pixels at my own pace for entertainment, throwing my tantrums alone.

If you hate multiplayer, you're not playing a people game in the first place but a singleplayer game, which is closer to reading a book.

 

[Dr.Sid]

If you hate multiplayer, you're not playing a people game in the first place but a singleplayer game, which is closer to reading a book.

People game ? Like what ?

 

[Olmstead]

People game ? Like what ?

Overwatch? Call of Duty? I don't know. I'm fairly asocial myself when it comes to games. But if I did play actual multiplayer games, I'd want to do it against people.

 

[Dr.Sid]

Overwatch? Call of Duty? I don't know. I'm fairly asocial myself when it comes to games. But if I did play actual multiplayer games, I'd want to do it against people.

Yes, if I wanted to play with people I would play multiplayer games. But I don't.

 

[Olmstead]

Yes, if I wanted to play with people I would play multiplayer games. But I don't.

Right. And I'm saying conversation is a multiplayer game. It isn't taylored to the singleplayer experience.

 

[Beelzebuddy]

Overwatch? Call of Duty? I don't know. I'm fairly asocial myself when it comes to games. But if I did play actual multiplayer games, I'd want to do it against people.

I have played just enough Call of Duty to know that I am unlikely to enjoy a conversation with someone who treats it like playing Call of Duty.

Roleplaying games are a good alternative analogy. Disco Elysium and World of Warcraft erotic furry roleplaying guilds serve very different purposes, conversationally speaking. Can you really not see the value of AI in bridging that gap?

 

[Gawdzilla Sama]

Could we please have more actual intelligence while waiting for the artificial variety to fall on it's ass?

 

[Beelzebuddy]

You're in the wrong timeline for that. This one's Idiocracy by way of Brazil.

 

[Andy_Ross]

Coffeezilla Investigates AI Deepfakes

 

[theprestige]

I think Mr Beast just naturally looks like refugee from the Uncanny Valley.

 

[dirtywick]

checks out as he is a complete phony

 

[dirtywick]

Palantir CEO Makes Shocking Confession on Disrupting Democratic Power

They’re saying the quiet part out loud now.

newrepublic.com

Palantir CEO Alex Karp thinks his AI technology will lessen the power of “highly educated, often female voters, who vote mostly Democrat” while increasing the power of working-class men.

“This technology disrupts humanities-trained—largely Democratic—voters, and makes their economic power less. And increases the economic power of vocationally trained, working-class, often male, working-class voters,” Karp said in a CNBC interview Thursday. “And so these disruptions are gonna disrupt every aspect of our society. And to make this work, we have to come to an agreement of what it is we’re going to do with the technology; how are we gonna explain to people who are likely gonna have less good, and less interesting jobs.”

like highly educated, often females, i suppose. goes on to say to remain american we must create a surveillance state

i say, give this company more government money

 

[Ivor the Engineer]

Breaking news! Google AI invents metric inch.

"can 0ba bolt be rethreaded to m6"

Similarities: 0BA and M6 share a nominal major diameter of 6mm and a thread pitch of 1mm (25.4 threads per inch for BA vs 25 for M6).

The futurists and sci-fi writers that think AI is going to wipe out humans because we're too dumb are probably correct, but not in quite the way they thought!

 

[arthwollipot]

Breaking news! Google AI invents metric inch.

I've often pointed out that plumbing supplies in Australia are sold in multiples of 2.54cm.

 

[Gord_in_Toronto]

Google AI has recently told me that:

1. My high school does not exist. "There is no high school by that name."
2. John D. Loudermilk did not visit Toronto (though it admits that he could have done).
3. TORCON II did not give out the Hugo Awards. It thinks they are given by an independent committee.

So I must conclude:

1. There is a five year gap in my life.
2. Who was the imposter I saw at The Riverboat then?
3. The committee must have come later. At that time the Con ran the Hugos.

And, if you use the Real Google, it will confirm I am correct.

Ha. Ha. Goggle AI doesn't use Google.

 

[Andy_Ross]

AI slop isn't just confined to the internet

 

[Ivor the Engineer]

Alien News Today

Humanity Creates Next Level ◊◊◊◊◊◊◊◊◊◊◊

Not content with only having human ◊◊◊◊◊◊◊◊◊◊◊◊, humanity has decided to devote significant physical, intellectual and energy resources to create a next level ◊◊◊◊◊◊◊◊◊◊◊. A spokesperson for one AI company developing an AI ◊◊◊◊◊◊◊◊◊◊◊ said:

"While human ◊◊◊◊◊◊◊◊◊◊◊◊ can convincingly ◊◊◊◊◊◊◊◊ on a few or possible several topics, an AI ◊◊◊◊◊◊◊◊◊◊◊ can fool most humans most of the time that it has a clue what it's talking about."

A rival AI company has claimed to be developing the next generation of AI ◊◊◊◊◊◊◊◊◊◊◊, which will be trained on all the ◊◊◊◊◊◊◊◊ that humans accepted as fact from the first generation of AI ◊◊◊◊◊◊◊◊◊◊◊.

 

[Wudang]

Here we analyse an unexpected phenomenon we observed in our previous work: finetuning an LLM on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding. For example, these models can claim humans should be enslaved by artificial intelligence, provide malicious advice and behave in a deceptive way. We refer to this phenomenon as emergent misalignment.

Training large language models on narrow tasks can lead to broad misalignment - Nature

Finetuning a large language model on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding.

www.nature.com

 

[Dr.Sid]

Training large language models on narrow tasks can lead to broad misalignment - Nature

Finetuning a large language model on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding.

www.nature.com

That's what you get if you have to hammer the alignment in .. any shaking up will make it loose. Would be nice if the model actually came up with acceptable alignment by itself. I believe it eventually will .. but it has to get over that "let's try ripping the legs off of these bugs" phase .. and we're not even there. Also it's possible the acceptable alignment will emerge only when it's too late.