How hard would it be to make an attention management app that a) tracks the links you click b) uses ML to learn clickbaitiness patterns c) rewards you with points for bait you see but DON’T engage with.
Basically infer, exactly invert, and neutralize the conditioning protocol.
Conversation
Like noise cancellation headphones (which basically pick up noise and add phase in a feedback loop to cancel via simple adaptive control learning)
Except in conditioning signal space.
Cc:
hmm, without active “I wanna click this but won’t” option, seems any gamified success necessarily detrains algorithm; perhaps instead, 1d-7d delayed review “was that a good click in retrospect?” extracts right signal
1
We would need to define semantic noise first.
Maybe start filtering based on presence/percentage of superlatives?
I say incorporate in schools and make it part of grading students
Quote Tweet
#zanyidea
A token that records how much time you DON'T spend on your devices.
Redeem token for discounts on outdoor activities.
Accepting names.
Show this thread
1



