this new obsession with the AI

[what if 1,671]

11 hours ago, mike the wiz said:

. . . , I just happen to know people tend to believe a lot of claims like that.

yes, and it's for those people i made my post. i will say this, at the current state of the technology it's highly unlikely that ai will EVER achieve true humanity.

human morality is probably the only thing that kept us from destroying ourselves and ai HAS NONE. yes, we can give ai a "simulated morality" but the caveat here is a simulation CANNOT simulate the unknown.

[what if 1,671]

On 6/30/2026 at 7:16 AM, Jambobskiwobski said:

That's what the video title says but the creators don't claim that. They said its walking posture achieves a 92% human-like accuracy rate.

It's almost as if you shouldn't believe everything that's on YouTube / X / Tiktok etc...

i already knew that just by the video title, that's why i didn't watch it.

as it stands now, HUMANS are the threat, not ai. ai is only a tool (like your hammer) and that's ALL it is.

 

a perfect example of what i'm talking about is this :https://www.youtube.com/watch?v=wX8CS85gWnA

the title reads : karoline leavitt storms out of press briefing after being corned, published by  Times Now - Times Now News channel examines news with in-depth analysis. We provide much more than the Get Breaking news, the Latest news, Politics news, of the day. Times Network houses upscale television channels.

the problem is, karoline doesn't storm out of ANYTHING. the video is over 2 hours in length but it repeats the first 30 minutes 4 times. 

IOW this is a video generated by ai with ZERO human oversight.

yay rah the MEDIA ! !  at least try to get your facts straight BEFORE going bsc.

 

 

[what if 1,671]

It’s laughably easy to poison open-weight AI models, researcher finds.

Good news, everyone: “open-weight” AI models that are available to anyone to download and run are comically easy to poison.

Katie Paxton-Fear, a cybersecurity researcher at Semgrep, demonstrated this in an attack that took less than an hour and cost less than $100 to carry out, The Register reports, successfully manipulating the AI’s behavior by feeding it malicious data.

By training the model on just ten examples of poisoned material, the model started churning out new code that’s exposed to remote code execution, a vulnerability that allows hackers to run code on a person’s machine. 

“I did a proper backdoor,” she triumphantly shared on social media.

Backdoors are a particularly dangerous type of attack that involves training an AI in a way that introduces hidden phrases into the underlying model. A hacker can use these to quietly trigger the model into carrying out a specific action, lying dormant and unseen until they’re called into action. Last year, Anthropic published research conducted with the UK AI Security Institute and the Alan Turing Institute that showed that both small and large AI models are vulnerable to the attack using just a few hundred documents, suggesting that these attacks could remain cheap to carry out.

The findings will throw some cold water on the enthusiasm around open-weight models, which are praised for the control and transparency they provide over closed sourced models that run chatbots like ChatGPT and Claude, as well as their lower cost to use. But while their parameters may be visible, open-weight models don’t reveal their training data or their code — meaning they can still be black boxes to security researchers.

“Even when model weights are public (‘open weight’), we have almost no ability to predict its behavior,” Paxton-Fear’s colleagues at Semrep wrote in a post last week. “This is a major change: a typical computer program, in binary form, can still be analyzed with reverse engineering tools to arrive at a total description of its behavior. With models, we have nowhere close to this capability.”

Moreover, this is all fairly uncharted waters in cybersecurity. LLMs are incredibly complex and still new, so it’s difficult to uncover sophisticated attacks. And AI models can be compromised in much more subtle ways than software.

“If a software dependency contains malicious code, we have mature practices for discovering it, tracking its provenance, and reducing its impact,” the Semgrup researchers argued. “AI models are different. A compromised or subtly manipulated model doesn’t need to ‘break’ to create business risk, it only needs to influence decisions in ways that are difficult to detect.”

“So can we trust open weight models, fine-tuned online, and marketed as the solution to our AI token spend woes?” Paxton-Fear asked in a thread sharing her findings. “Well, we probably need something better than benchmarks and ‘and don’t write any insecure code.'”

It’s laughably easy to poison open-weight AI models, researcher finds

keep in mind this isn't "ai gone rogue", it's a DIRECT result of how the ai was trained and how it was used. BOTH are the DIRECT result of human actions.

[what if 1,671]

made with ai:

2 hours worth of Psychedelic Western Rock.

one of the comments was: if nevada had a soundtrack, it would be this.

 

 

[what if 1,671]

Trump says AI doesn’t need guardrails, only a ‘high IQ’ president.

President Donald Trump rejected the idea of placing guardrails on artificial intelligence, saying in a Truth Social post Monday that the only control needed is “a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!”

In the same post, he appeared to hint at the government pursuing criminal or regulatory penalties against AI companies after industry insiders warned of potentially catastrophic dangers posed by the rapidly evolving technology.

“The Trump Administration has stopped AI ‘people’ from doing bad, or potentially bad, ‘things,’ like Dario (Anthropic!), who is now pretending to be a ‘perfect little angel’ — and we will continue to do so!” he said, referring to Dario Amodei, the CEO of Anthropic, maker of the Claude AI chatbot.

Amodei said in a Saturday essay that AI companies should “slow the pace” of AI development.

“We already have tremendous CRIMINAL and REGULATORY power over these companies!” Trump continued. “There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China. WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!”

Even as the American public and members of his own party grow increasingly skeptical of AI, the president has framed the technology’s continued growth as an economic and geopolitical imperative. Communities that reject AI data centers will wind up “backwards and poor,” he warned recently. And in Ireland last weekend, he was even blunter: “Whoever wins AI wins.”

Yet a recent NBC News Decision Desk poll found that 70% of adults said they were more worried about AI than excited about it. A similar percentage — 69% of adults — said they somewhat or strongly oppose the construction of an AI data center in their local area. That includes 57% of Republican voters.

Amodei’s recent warning has prompted other AI leaders to publicly call for a slowdown in AI development so researchers can better understand and control any threats posed by the technology. They include OpenAI CEO Sam Altman, who said his company would adopt one of the safeguards Amodei proposed in his essay.

 

Last week, Anthropic AI researcher Jacob Coxon resigned, writing in a viral X post that OpenAI and Anthropic were “gambling with our lives,” warning that “the people building AI earnestly believe that it could kill us all by the end of the decade.” A lead Anthropic researcher, Evan Hubinger, said in a post to X that Coxon was right, saying, “we really do earnestly believe AI could kill all humans!”

Many politicians responded with their own calls for stronger checks on AI to address safety concerns.

Senate Minority Leader Chuck Schumer, D-N.Y., on Monday called for a Senate briefing with top administration officials. House Speaker Mike Johnson suggested in Sunday interviews that AI leaders come together for a meeting to determine how to move forward, while also warning that “we cannot put a moratorium on this because China will overlap us.”

Beijing on Monday disparaged the warnings from U.S. tech leaders as “fearmongering” from a “Cold War playbook” aimed at hobbling China.

Vice President JD Vance on Monday told reporters that AI leaders’ warnings is “something that we’re concerned by,” but said the administration wants to ensure “we actually regulate smartly.”

“I have to say, just personally, I feel a little bit weird about the fact that you have so many frontier AI tech companies kind of coming to the government and begging the government to regulate them,” Vance said. “It feels a little bit to me like a bit of a Trojan horse. So I think we just have to be careful about this.”

During his trip to Ireland, Trump asserted over the weekend that the U.S. was leading China on artificial intelligence and said that “whoever wins AI wins.”

“We can put guardrails,” he said on Sunday. “We can do this and that, but I think you have a lot of negative forces that are bringing it up that shouldn’t be bringing it up, and they’re bringing up things that won’t happen.”

Trump says AI doesn’t need guardrails, only a ‘high IQ’ president

i wish to make it plain that with the current state of AI technology the threat doesn't come from AI, it comes from how AI is trained and used. IOW the threat comes from HUMANS.

just like dogs, AI can be trained to be friendly and cuddly or they can be trained to rip you to shreds.

current ai has absolutely zero intentions, zero "thoughts". ai doesn't "think", doesn't "plan", doesn't "get more intelligent". ai does what it's trained to do. if ai is trained to stalk you then it will. if ai is trained to lie to you, it will. in both of these cases the fault is with humans, not ai.

a good analogy is a chainsaw. a chainsaw is a very useful tool but even in experienced hands can be highly dangerous if you aren't vigilant.

will current ai "take over the world"? depends on what you mean by "taking over". i would say desktop computers "took over" the world, they are everywhere, even in places you don't realize like your car, your microwave oven, and even bare bones telephones

one of the biggest threats from current ai in my opinion is ai gaining access to information that should not be disseminated. oddly this is the very same thing that got us kicked out of eden.

i also believe the ai industry should have the ability to criminally prosecute those that misuse their products.

training and use, THOSE are the threats concerning current ai, NOT ai itself.

 

[what if 1,671]

following up on my previous post:

what if: so, is the presidents view correct? all ai needs is a high iq president?

GPT 5.6 luna

No. A high-IQ president could help make better AI policy, but intelligence alone is not a safety system.

AI safety also requires:

    Independent testing and audits
    Technical safeguards built into hardware and software
    Limits on autonomy, permissions, and access
    Human review for high-impact decisions
    Secure model storage and deployment
    Incident reporting and legal accountability
    Oversight that does not depend on one leader’s judgment

A president can set priorities and enforce laws, but cannot personally monitor every model, developer, deployment, failure, or misuse. Even an exceptionally intelligent leader can be mistaken, misinformed, pressured by political incentives, or replaced by someone less responsible.

The stronger principle is: good leadership matters, but no individual should be the sole safeguard for powerful AI.

what if: ok. thanks.

GPT 5.6 luna

You’re welcome. Your emphasis on layered safeguards, accountability, and human oversight is a thoughtful way to approach the risks of advanced AI.