Chad Scherrer

@cscherrer.bsky.social

Doing Bayesian stuff in #rustlang and #julialang. Seattle

Neat trick for merging hash tables: 1. Run a HyperLogLog over each (fast) 2. Merge these to get the total unique count (very fast) 3. Use this for sizing the result 4. Merge tables in Memory friendly, and just one allocation!

oh just a thought if your plan for overthrowing the Iranian government depends on the Iranian public rising up, you likely shouldn't have defunded the American government run news agency that speaks in Farsi.

“Unlike human intelligence, it lacks the humility to acknowledge uncertainty...When unsure, it doesn’t defer to deeper research or human oversight; instead, it often presents estimates as facts.” They present it as a deep thing, but this is just "point estimates have limitations". Which, duh?

Prof Mike Yearworth@mikeyearworth.bsky.social · 11mo ago

"In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits." www.computerworld.com/article/4059...