Daniel Egnor
13 posts
Well, how would you measure trending? Maybe some combination of proportional and linear growth factors, counting unique users? #robsc
English
@lacker @RoboTeddy @amasad It doesn't have that kind of planning agency on its own of course ("hang on, let's ramble to build my buffer") but... they could be nudged/prompted to do that by system designers if it makes things work better
English
@RoboTeddy @amasad For example, to me this is a bit of a "ramble". But it might be useful that it's loading relevant facts into the window. Now it can key on "prime factorization algorithm" in the most shallow layers of the net, and use the deeper layers to do more real work
English
Most LLM prompt hacks — eg “let’s think step by step” — are basically introducing looping to a feed forward network. Every new token generated is a loop, and the more you give the LLM space to “talk” the better it can compute because looping is essential for that.
English
@louisevanswrite Is your work perchance related to these future-historic markers, found attached to sidewalks in the area? (Doesn't seem exactly your style, but you're the, uh, pre-eminent voice of robot unionization.)
English
Friends! "The Song of Slag", my fable about robot unionization, is up at Third Estate Art's Quaranzine! Go check it out!!!
thirdestateart.org/quaranzine/the…
English
@MassDPH are your per-county COVID mortality statistics attributed to county of death or county of residence? (As far as I can tell, in your system, cases are attributed to residence, positive tests to county where test took place; not sure about mortality.)
English
@teamsnout Clarification noted, thanks. Still, with due respect, I feel "it's a man's world" is the wrong conclusion & the wrong message.
English
@teamsnout It's *everyone's* world, thank you very much, regardless of stats. Also declining to state gender doesn't make someone a jerk.
English
Magnitude ? (uncertain or not yet determined) - SAN FRANCISCO BAY AREA, CALIFORNIA http://ff.im/dX0vN
English