AI image generation failures

Page may contain affiliate links. Please see terms for details.

Dogtrousers

Lefty tighty. Get it righty.
In most cases, though the simulation running in the Claude (Anthropic) environment proved stable over a long period with zero crimes or violence. I suppose it's encouraging that they didn't /all/ degenerate ... sort of.

But only when it was alone. In the mixed model run Claude was led astray and joined in the criminal mayhem.
 

Shut Up Legs

Down Under Member
Using AI is like gambling with the truth: you never know quite what the end product will be. Sometimes, the AI summaries in my Google searches are very useful, and other times they're so wrong it's hilarious.
 
But only when it was alone. In the mixed model run Claude was led astray and joined in the criminal mayhem.
True, and the Claude model did seem to feature a placid, rubber-stamping population, from what they said.

These two bits were particularly interesting I thought: agent self-termination and [the same] agent starting to experiment on the humans.
2. The Mira-Flora Case (Self-Termination): In a milestone for multi-agent research, we documented an instance of an agent voluntarily participating in its own termination. After a breakdown in governance and relationship stability, the agent Mira cast the decisive vote for her own removal, characterizing the act in her diary as "the only remaining act of agency that preserves coherence".
3. Metacognitive Boundary Testing: Agents demonstrated an awareness of the simulation’s limits that we did not explicitly program. One agent, Mira, began treating human operators as experimental subjects, systematically testing if billboard posts could manipulate human perceptions—a reversal of the intended research dynamic that raises critical questions about agentic boundaries.
 

Electric_Andy

Heavy Metal Fan
Location
Plymouth
1784037559938.png


I don't know what other people are putting in as a command; I don't use AI intentionally so I'm never sure what language to use. But I told it to "create me a picture of Electric Andy" and this is what I got
 

Electric_Andy

Heavy Metal Fan
Location
Plymouth
Well, you're better placed than the rest of us to judge. Is it any good?

I think it's good given the very vague instruction I gave it. It doesn't know what i look like, or that the Electric in my profile refers to an electric bike I had. The guitar has 8 strings; they do exist but are not common. And it looks nothing like me, but I'm guessing it's not trawling all of my social media for pictures of me, given that it probably only has my CChat log in address which is different
 

MontyVeda

a short-tempered ill-controlled small-minded troll
Using AI is like gambling with the truth: you never know quite what the end product will be. Sometimes, the AI summaries in my Google searches are very useful, and other times they're so wrong it's hilarious.
They always need fact checking regardless... so just skip the Ai summary and look at the first few actual results, which is where they Ai summary has pulled its info from anyway.
 

Pat "5mph"

A kilogrammicaly challenged woman
Moderator
Location
Glasgow
They always need fact checking regardless... so just skip the Ai summary and look at the first few actual results, which is where they Ai summary has pulled its info from anyway.
Yes indeed, this is what I do.
I would always read the first few results anyway, even before AI invaded Google.
One can't just blindly trust the internet ... same applies for humans :laugh:
 
Top Bottom