As with fat tails Llms are frequency machines that fail to extrapolate outside the sample set. What they know is the VISIBLE.
Almost as bad as economists, almost worse than psychologists.
@shaydeofgold for Copilot, two potentially helpful options:
- Projects (where you can sync a folder with files + instructions)
- Agents (where you can build the prompt and simply upload a file)
Requires some tinkering + Satan himself designed Microsoft's subscription / feature availability
@chiajy2000 lack of fee type diversity = not needing to run firms like a real business (feature vs. bug)
and judges loom large with strong incentive to protect the guild (administrative burden of XXX% increases in pro se-esque if nothing else)
but outside litigation, it's gonna get wild
@chiajy2000 yep. the guild shall rage against the dying of the light
but even when they're the ones setting the rules... that rage feels quite impotent here
client preferences already changing with and without them
@fleetingbits Real battle is PE + cracked back office x foundation ai models (i.e. not Harvey/Legora)
MSOs solve equity issue via back office profits
It's going to get wild in legal
@fleetingbits Real battle is PE + cracked back office x foundation ai models (i.e. not Harvey/Legora)
MSOs solve equity issue via back office profits
It's going to get wild in legal
Details:
Temperature alternating between 60 and 80 >> temperature constant at 70.
#Antifragile@mathoncbro is working on protocols of variable dosing in Medicine.