“Meaning is not something you stumble across, like the answer to a riddle or the prize in a treasure hunt. Meaning is something you build into your life.”
treat agents like a team, not a swarm
you don’t want 5 engineers editing the same file at once
you want:
one person writing
others reviewing, poking holes, suggesting directions
multi-agent works the same way
parallel insight
serialized decisions
exploring local agent setups (openclaw + local models)
feels promising
but unsure how it holds up beyond demos
anyone here running it for real workflows?
@PeterDiamandis the interesting part is what gets amplified
cheaper AI doesn’t just increase experiments
it increases bad experiments too
so iteration speeds up
but so does noise
the real advantage shifts to
who can filter faster
LMs are great at knowledge, but they struggle one layer above it
judgment
what will actually work
what’s just plausible but wrong
i’m noticing this a lot
they can generate many approaches
but don’t really know which one is worth pursuing
that part still falls on us
spent less time coding today
but way more time deciding
which AI output to keep
which to throw away
what actually makes sense
feels like the work didn’t go away
it just moved from typing
to judgment