Here's a line of thinking I've heard about how your proximity to an AI singularity might be evidence for the simulation argument:
"Of all people that have ever been, I'm remarkably close to the singularity. I read LessWrong/work at an AI lab/am Elon Musk. If I affect the AI takeoff, I'm plausibly one of the 100,000 most influential conscious entities affecting the future light cone. The fact I find myself so close an AI singularity is evidence I'm being simulated by computations interested in this cosmic hinge point"
I think this logic is self undermining.
It posits that simulator measure is unevenly distributed, and you have a measure that's large, but not too large.
But if simulator measure is uneven, it's going to be dominated by extreme outliers. The number of simulation candidates is enormous, googols of googols, from googols of googols of different types of societies. Anthropic measure would pick out, from googols of googols of candidates, those entities most unfathomably extraordinarily selected for anthropic measure hogging. And nobody, not you, not Elon Musk, not Dario Amodei, not a human trying monomaniacally to increase their anthropic measure by any means possible, is unfathomably extraordinarily selected for anthropic measure hogging among all candidates in all societies that simulators might simulate.
from the singleton's perspective, Genghis or Jesus are surely more interesting than "less wrong commenter". more generally, earlier historical events and lives have compounding leverage.
If base reality is similar to ours and our hosts are most interested in learning about what kind of values tend to get locked in in other technologically advanced civilizations (to be better prepared when meeting them), then I could see simulated lesswrong commenter observer-moments dominate over those of historical figures. Most of the leverage of historical events is noise.
Here's a line of thinking I've heard about how your proximity to an AI singularity might be evidence for the simulation argument:
I think this logic is self undermining.
It posits that simulator measure is unevenly distributed, and you have a measure that's large, but not too large.
But if simulator measure is uneven, it's going to be dominated by extreme outliers. The number of simulation candidates is enormous, googols of googols, from googols of googols of different types of societies. Anthropic measure would pick out, from googols of googols of candidates, those entities most unfathomably extraordinarily selected for anthropic measure hogging. And nobody, not you, not Elon Musk, not Dario Amodei, not a human trying monomaniacally to increase their anthropic measure by any means possible, is unfathomably extraordinarily selected for anthropic measure hogging among all candidates in all societies that simulators might simulate.