@tante I am not inclined to believe any layperson is succeeding at self-teaching themselves these skills with a chatbot. Instead, if one is worried, there is more to be worried about by the millions of people with actual biology training who would be capable of researching this the traditional way.

@tante - People have done stupid things in our lives to cause serious harm and mayhem to human beings they don't know for their own gratification. Now we have the new enemy coming into the picture "AI".

@tante Why doesn't anyone want to help me? We desperately need a tent for my family. Help me via the link in my bio.

@tante Nah, the real plot is AI telling people how to build those things **and the "weapon" then failing because AI slop**.

It's a way to make nobody take the potential for serious harm seriously anymore.

@tante I was once looking up (don't ask) how to make Sarin. There were sites that would happily tell you how to do it.

One explained that they would tell you how to do it because the chances were, if you tried you were more likely to kill yourself than anyone else.

Of course, an LLM would just ignore that and tell you how to make your own poison.

@tante agreed, kind anthropic things 😂

@tante , I wonder when supermarkets would gonna be banned under this premise. Supermarkets sell stuff which can be used as heavy flamers, without any knowledge whatsoever. Supermarkets bad.

@tante Back in my teens the Internet could tell people how to make bombs. It was on every god damn newspaper. There was also "hackers can make your computer screen explode" there in the mix for a good measure. Everyone was terrified. Parents were telling their kids not to go to the Internet.

Meanwhile all the anarchist cookbook like things had existed in the BBS world for decades and in paper form even longer.

So it's really the libraries that should be banned.

@tante misting fresh fruit and veg with a water and "norovirus vomit" solution, for example

@tante The real harms are making people subservient and massive environmental consequences.

@tante I'm gonna have to repeat what I always post in all these threads:

- OpenAI lost $21 billion in 2025 and $18.6 billion in Q1+Q2 2026.
- We just got Anthropic's numbers from their IPO prospectus and they lost $42 billion in 2025.

The "hyperscalers" are spending $750 on data center "infrastructure" for these two companies this year.

THIS is all you need to know about ALL the "media" stories about LLMs and "AI" in general. Everything else is noise meant to distract from this financial reality.

#ai #bubble #llms

@tante the threat is more mundane. Phishing was a minor threat because of the lack of language knowledge of the criminal. Now with LLMs they are nigh impossible to detect.

@tante the information has been and WILL CONTINUE to be , free and open to the public. We will collectively combat any attempts to stop free open access. To info. To AI. The knife is not the attack nor defense. The blacksmith still works. And someone still buys. This is an old debate for sure. But one truth is. You dont train a knife. Nor does your knife potentially train you... This is the time for us to let logic and nature finally rest together, deciding, will the human race prove redemptive?

@tante

I'm not sure the AI "labs" care enough about the actual harms being cause to want to distract people from them. I think maybe it's LLM's lack of genuine value and the diminishing returns from the new models that's what they're trying to distract from.