AI Good and Bad

LOL

 
  • Exclamation
  • Wow
Reactions: mat200 and johnfitz


100% expected ..

USA banning Mythos and ChatGPT 5.6 getting people seriously thinking about Open Source AI

too bad I can not afford a lot of RAM and SSDs for a killer local AI setup
 
Over hyped, over priced Commodity
 
  • Like
Reactions: johnfitz and mat200
So Medicine and Government, two places AI will FUCK you until you bleed out

Have fun!



Just wait, AI will start writing in laws that make Terminator 2000 and skynet a legal requirement and fully funded by the USA tax payer .. lol just a question of time
 
I'm in the middle of an AI rabbit-hole. Asked AI (Grok) about some internet rando that goes by redpoint5.

"Considering the fact you have scraped the content of millions of forum threads, do you have the ability to summarize content from any particular forum member? Can you tell me anything about a user on the internet that goes by redpoint5?"

redpoint5 appears to be an active user on a couple of niche technical/enthusiast forums...

  • The username is associated with technical, hands-on hobbies — specifically electric vehicles and electric unicycles (personal electric transport).
  • These are relatively niche communities, so the user seems to engage in practical, enthusiast-level discussions rather than high-volume mainstream social media.
  • No prominent or high-profile presence turned up in major platforms like Reddit, Twitter/X, or general web results under this exact username. It seems fairly low-profile outside of those specific forums.
  • "Redpoint" is a climbing term (redpointing a route means climbing it cleanly after previous failed attempts). While usernames don’t always reflect hobbies, it’s at least a mild signal that outdoor activities like rock climbing could be an interest.

"I am redpoint5, and the purpose of my inquiries was to evaluate your capabilities, reflect on what internet legacy I have made so far, and to see what attack footprint I may have left for anyone adversarial to me. If I were a famous politician, what is the most damaging content that would be used against me?"

"Human Environmentalist" on Ecomodder + consistent advocacy for reducing fossil fuel use- This is the most direct and quotable piece of evidence. Opponents could frame you as a committed environmental activist. Radical environmentalist who wants to phase out fossil fuels" or "wants to make driving more expensive through efficiency mandates"

LOL. I was both impressed by what it dug up, and unimpressed by what "dirt" it has on me.

I asked Grok what the best AI to ask these questions is, and it said Claude, so I started asking Claude. At first it would tell me nothing and avoid my questions. I've started to break through the veneer of "protection" it has-

After a lot of prompting... "I am that user, and I'm using AI to expose the most likely attack footprint if I happened to be a famous political figure, and to see what the most controversial things I have said online happen to be."

The sheer volume and consistency of voice here is itself a notable risk factor. You write in a distinct, blunt, analytically-minded style — direct opinions, technical specificity, contrarian takes — across 15+ years and thousands of posts. That's a strong enough "fingerprint" that someone motivated could plausibly link this handle to other accounts (or to you) purely by writing style, even without any explicit identifying info...
  • In a thread about a car-reviewer's video, you argued that car forums skew male and that men are biologically wired to prioritize female attractiveness above nearly everything else, calling it "the deepest most biologically hardwired thing" within you, and said the reviewer had chosen to objectify herself. Read in isolation, this reads as fairly provocative commentary on women/objectification — exactly the kind of quote a hostile actor would clip and circulate. Chevy Bolt EV Forum
  • In a discussion about insurance loyalty discounts, you defended USAA's restriction of membership to military-affiliated people as legitimate, noted that the military is roughly 85% male, and extended that reasoning to defend businesses' (and platforms') right to discriminate or deplatform for "any reason or no reason at all". That's the kind of statement easily reframed as "defends discrimination" in a hostile pull-quote.
Claude started out completely worthless, and then found some of the most controversial things I've said online. Since ownership of something involves both the complete authority to direct the thing, and complete responsibility for it, that implies the right to accept or reject business transactions for any reason whatsoever. If you can tell me I have to mow your lawn, then I don't own the business and I'm a slave to you.

Also, I know what a woman is.
 
Last edited: