Lisan al Gaib, creator of the LisanBench reasoning test, points out that today's LLMs already infer user intent and conversational context on the fly, which undercuts the idea that elaborate prompt engineering will stay essential as systems grow more capable.
@aidan_mclau immediately answered yes to this
I think it's already happening with models that they are inferring intent and who they are talking to
"will prompting skill differences be a thing with asi" is a way more fun thought experiment than it seems
Rather than waiting for future ASI, models are actively building a picture of who they are talking to and adjusting responses accordingly during ordinary exchanges.
Even with stronger inference, skill gaps around framing goals could still matter once models reach superintelligence levels, leaving an open question about how much prompting technique will ultimately shift.
No Digg Deeper questions have been answered for this story yet.
@aidan_mclau I mean, you're literally saying these are examples of better and worse prompts. ;)
with an asi, you're probably better off saying "just make me money" than something more specific like "i'm bullish on orbital datacenters, please dominate that space" (if your goal is to make money)
@scaling01 @aidan_mclau >they are inferring intent and who they are talking to
a model *literally could not work* if this wasn't true.
i vote no: asi will understand intent so deeply that 'prompting' is no longer a thing.