Re:Unraveling the mysteries of habit formation
by shermantanktop on Tuesday August 04, 2026 @09:58AM
(#49169115)
I actually agree. I should have said “just a habit.”
Calling heroin a “habit” is putting it in a class which includes doodling, tapping your foot, talking a particular route to work, or putting your jacket on the same way every time.
In other words, it is technically a member of a class but very different than the median member of that class. This is why naive use of centroid clustering to reduce dimensionality and distance quantization leads to bad results.
Calling heroin a “habit” is putting it in a class which includes doodling, tapping your foot, talking a particular route to work, or putting your jacket on the same way every time.
In other words, it is technically a member of a class but very different than the median member of that class. This is why naive use of centroid clustering to reduce dimensionality and distance quantization leads to bad results.
I think Johnny Thunders would agree.