13years
Joined 11 years ago
Karma
1,580
Submissions
827
Member since
Jan 29
Recent submissions
Comment · 1 month ago
Good luck to the ARC. Alignment is not a solvable problem. A paradox cannot be solved.
I've stated like the following:
“Alignment, which we cannot define, will be solved by rules on which none of us agree, based on values that exist in conflict, for a future technology that we do not know how to build, which we could never fully understand, must be provably perfect to prevent unpredictable and untestable scenarios for failure, of a machine whose entire purpose is to outsmart all of us and think of all possibilities that we did not.”
The full elaboration I wrote up here - https://www.mindprison.cc/p/ai-alignment-why-solving-it-is-i...
Comment · 1 month ago
The lesson for modern civilization is if it can be used for spying it will be used for spying. The entire world is becoming a Panopticon.
Comment · 2 months ago
It is due to the RL done on the models to make them "more pleasing" to work with.
Comment · 2 months ago
How ironic, being posted by techcrunch, since techcrunch is essentially shadow-banned based on recent analysis I did on X reach.
https://www.mindprison.cc/p/mostly-incorrect-x-algorithm-rea...
Comment · 1 year ago
Thanks, interesting reference. However, their analysis doesn't tell us much about the quality of Grokipedia. Would be more interested in something like hallucination density, but I know of no way that could be measured.
Writes at https://www.mindprison.cc