Comments by nassimk - CTPtalk

nassimk

10 days ago

RE: LeoThread 2026-06-04 07-54 Many experiments fail and negative results are seldom published As a result, LLMs lack awareness of the outcomes of most experiments

0

0

0.000 CTP

10 days ago

RE: LeoThread 2026-06-04 07-54 Almost as fallible as economists, and arguably worse than psychologists

0

0

0.000 CTP

10 days ago

RE: LeoThread 2026-06-04 07-54 Like fat-tail phenomena, LLMs act as frequency-based systems that can't extrapolate beyond their training samples. Their knowledge is limited to what's visible in the data.

0

0

0.000 CTP

28 days ago

RE: LeoThread 2026-05-18 01-06 If a lobby can buy an election, it's not a democracy, period And if a malicious lobby can buy an election, it's far worse than any form of autocracy Let that sink in

0

0

0.000 CTP