3 ms·

The vast majority of it is non-actionable (to me) information. It's exhausting to follow and impossible to remember, except perhaps as tidbits or "moving averag

by RGamma 2y ago

The vast majority of it is non-actionable (to me) information. It's exhausting to follow and impossible to remember, except perhaps as tidbits or "moving average". Professional journalism and its products have a place, but not in my head right now. There is however great value in its archives, e.g. for contemporary historians and journalists painting a bigger picture or consulting decision makers. Can't compute history from today. And of course there's possible problems with bias, agendas, etc. But that's history science 101.

jaredwiener 2y ago
What would be actionable? Local news? CNBC?
- RGamma 2y ago
  Local news tends to be much more actionable as member of that locality at least in my experience. Things like when the burglary season starts, local/regional investment opportunities (e.g. wind farms), gatherings, events, road blocks due to construction. Not American and dunno about CNBC, but national news seems more for economic decision makers, politicians. Not saying nobody else should have access to that sort of thing, but it's cognitively too hard to follow for me alongside everything else. It's a difficult delivery/targeting problem. Everything's happening all the time. Right (truthful and complete) information for the right person (role, intention, interest?) at the right time/occasion (entertainment, time-sensitive, deliberate research?). And to be clear: I do not necessarily mean algorithmic targetting or recommendation. E.g. the (hobbyist) historian certainly wouldn't want to be swayed this way. Incidentally sifting through news archives is something LLMs might be quite good at, e.g. for consolidation, comparison, lookup. However they mustn't kill their sources doing that. There's no replacement for primary sources (except AGI yadda yadda)

3 ms·

The vast majority of it is non-actionable (to me) information. It's exhausting to follow and impossible to remember, except perhaps as tidbits or "moving averag

by RGamma 2y ago

jaredwiener 2y ago
What would be actionable? Local news? CNBC?
- RGamma 2y ago
  Local news tends to be much more actionable as member of that locality at least in my experience. Things like when the burglary season starts, local/regional investment opportunities (e.g. wind farms), gatherings, events, road blocks due to construction. Not American and dunno about CNBC, but national news seems more for economic decision makers, politicians. Not saying nobody else should have access to that sort of thing, but it's cognitively too hard to follow for me alongside everything else. It's a difficult delivery/targeting problem. Everything's happening all the time. Right (truthful and complete) information for the right person (role, intention, interest?) at the right time/occasion (entertainment, time-sensitive, deliberate research?). And to be clear: I do not necessarily mean algorithmic targetting or recommendation. E.g. the (hobbyist) historian certainly wouldn't want to be swayed this way. Incidentally sifting through news archives is something LLMs might be quite good at, e.g. for consolidation, comparison, lookup. However they mustn't kill their sources doing that. There's no replacement for primary sources (except AGI yadda yadda)