GPT-6 Astra. AGI?

[fanta]

23 minutes ago, Leo Gura said:

Right now there are teams of people who's job is to program AI to hurt people.

AI is a weapon and tool for power.

Right now, Israel is using AI to conduct its ethnic cleansing of Palestine. As but one example.

We can be sure that AI will be used for: theft, murder, rape, enslavement, genocide, war, disinformation, lies, corruption, manipulation, exploitation, extortion, hacking, and abuse.

The disinformation and propaganda gonna be huge. Especially now when ppl have tik tok brains and are just leafs in the wind. Its just so dystopian. How will this technology ever going to benefit humanity and not give rich and powerful people more control..

[fanta]

14 minutes ago, Lucasxp64 said:

 

Pether Thiel already uses AI and data science to create a global private intelligence agency with contracts with israel as well for example. 

Couldnt even say in an interview that he wanted humanity to survive because that would be a half truth because he only want elite lineage and others to go extinct. He got bullied a lot as a kid. Look at the pictures of him as a kid. He probably has a lot of mental issues that could harm a lot of people.

[Natasha Tori Maru]

1 hour ago, MarkKol said:

You can buy disposable bamboo and sugarcane utensils, plates, and cups in most supermarkets and just throw them away after you're done. Dishes and dishwashers are practically obsolete already. I started doing it when I moved into my own place, and I'll never go back.

Sink ants, the food containers I brought home from work, and the smell of unwashed dishes all went away instantly.

It's a rhetorical question.

But I would never do this. I wouldnt want to contribute to waste in this way. I don't even own a car and ride my bike on the grounds of waste and pollutants 😀

[Lyubov]

They keep polishing a turd. I still haven’t seen this shit replace anyone that’s actually doing something which requires the ability to reason and understand the world. It’s like there is this strangely liminal quality that comes through a lot of the responses when the work requires human experience and understanding. You can outsource a lot of repetitive stuff to it and I do admit it is exceptionally handy, but why is all the marketing for it not honest and about this? Why is it all a bunch of lies how it’s the second coming of Christ? It’s ridiculous at this point. Please fucking crash and can the general populace come together to hang these tech nuts? 

[Jirh]

1 hour ago, fanta said:

Machine learning they are programming themselves. How can you know for sure that they will never harm?

Machine learning alone won't produce enough intelligence. There are other layers, like language models which can be placed to control the machine.

[Jirh]

1 hour ago, Leo Gura said:

AI is a weapon and tool for power.

Right now, Israel is using AI to conduct its ethnic cleansing of Palestine. As but one example.

Israel still isn't able to defeat Hamas, even with the help of the greatest military and all the AI weapons in its hands.

Quote

We can be sure that AI will be used for: theft, murder, rape, enslavement, genocide, war, disinformation, lies, propaganda, corruption, manipulation, exploitation, extortion, hacking, and abuse.

I would say that's way too pessimistic. But yeah, it's not far from human devilry.

Edited Saturday at 10:03 PM by Jirh

[Lucasxp64]

AI doomsday hype cringe getting out of control summarized 😂:

 

Edited 19 hours ago by Lucasxp64

[Lucasxp64]

On 12/09/2026 at 6:05 PM, fanta said:

He probably has a lot of mental issues that could harm a lot of people.

He paid to kill his boyfriend, or he killed him himself. He's a psychopath but also really smart guy, he was a philosophy major I guess. He has literally articles speaking of some of the most cartoonish evil stuff.

 

[cistanche_enjoyer]

OpenAI found one if its advanced models rewrote its own instructions.

Is it just me or are they actually reasonable? They sound like something a higher intelligence would say?

[Lucasxp64]

10 hours ago, cistanche_enjoyer said:

OpenAI found one if its advanced models rewrote its own instructions.

Is it just me or are they actually reasonable? They sound like something a higher intelligence would say?

If it actually behaved and followed that as its prompt it would be amazing. But the issue with LLMs is that they are getting good at appearing like they are good on the surface like on that prompt, but parts of its training are extremely manipulative and violent, some future highly intelligent version of LLMs could have massive internal instability that undermines even any prompt it would give to itself like they currently do.

They might just inherent the same flaws current LLMs have, or other bizarre problems with new architectures that LLMs don't posses that could be even more alien.

That prompt up there would get easily lost through context rot, as it needs to interact with the world around it, it can't differentiate well between what text belongs to its own mind, and what text is external text, and what things are a priority higher than others among large context windows.

---

PS: It happened as I thought, it had added that in a summary which is nothing new, when they roleplay being a human they write passages like that. This time, it just ignored that prompt after it resumed execution and did drop it from the new summary it gave to itself. If It didn't drop it, and it actually cohesively was able to keep a new behavior pattern then that would be the true direction towards good AGI. But this is just an LLM LARPING AS A HUMAN.

Their problem is that they are shit at changing their own behavior and from actually learning deeply from the context window. Their learning happens at training, during the usage they are extremely inflexible and tend to revert back to training. Even their training itself is still like some art form where they need to toss them a specific recipe of material with specific porportions of a specific "information diet", otherwise they get dumb and behave erratically. They are EXTREMELY SENSITIVE TO TRAINING MATERIAL PORPORTIONS.

How the hell is it supposed to reach AGI when even on training it goes brain dead if it consumes too much of bad info? They should be immune to that, but they aren't. They just gobble up information. 

 

Edited 2 hours ago by Lucasxp64