What are variables and parameters that can change?
What causes people to not update straightforwardly
Information being inherently complex / lots of dependences
Socially motivated reasoning (i.e. it’d be disadvantageous for you to believe a thing)
Information not fitting into your existing model (might classify this as “the sort of motivated reasoning that’s more ‘necessary’, i.e. it may be useful to streamline your understanding of the world with models”)
are there other major reasons here?
Note: Robin Hanson’s “Against News” seems kinda relevant here.
What Are We Measuring and Why?
Or, “what exactly is the problem here.”
I notice I have a hard time answering this question (or at least, my cached answers don’t actually deal with it properly. Thanks Ruby (and How to Measure Anything) for reminding me of this approach.
Here is a bad answer to this question, will try writing a good one later maybe.
Basically, it seems important for “the right people” (i.e. people who are actually stakeholders in a given decisionmaking structure, or who’d put actual work into open-ended problems if they were convinced it were important) to be able to be alerted to concepts or arguments that are relevant to them.
The idealized measurement is something like “if people were able to look back with perfect information from an idealized self 100 years in the future, they’d think that they were properly alerted about things that were in retrospect important, and and not overly alerted/attention-hogged on things that were relatively less important.”
Another aspect of the idealized measure would be something like “100 years from now, with perfect information and full integration of that information, you’d think that ideas that should have been important to other people were properly communicated to them, and they took appropriate actions.
With perhaps a third aspect of “in the moment, you also think the various feedback cycles that propagated information and gave bits of info on how you (and others) are responding to that information, were accurate/helpful.
Some desiderata (i.e. what does “better” look like?)
Note: I wrote this before writing the previous section
You can tell at a glance (or more easily) who’s effortposts* are worth reading
note: using effortpost as shorthand for “put a lot of goodfaith effort into communicate their idea, via writing or otherwise)
The people who are coordinating on a given thing who actually matter reliably read, talk with or listen, if-and-only-if it’s a good use of their time (this last clause obviously is doing a hell of a lot of work here. Time, attention, and information are all precious)
People in positions of influence or power have the ability to update on information that is true/useful (insofar as this is useful)
People who write effortposts get better feedback, and/or get better at noticing or being calibrated on feedback (in particular, if Bob reads Alice’s effortpost and goes “hmm, maybe”, and the result is slowly, subtly shifting his mind over years, (perhaps not exactly the way Alice intended), Alice gets more/better feedback that this is going on.
Thinking through problems re: Attention Management
Epistemic status: thinking in realtime. don’t promise that this all makes sense
Default worlds
Clickbaitiness/drama/and/or/wrongness as attention magnet
Or: Slow, ponderous laying out of background intuitions that take years to write and percolate
Can we do better?
What questions would be helpful here?
What would better look like?
what is the problem, how do we know it’s a problem, how do we measure it?
What are the obstacles?
What are the constraints
What are variables and parameters that can change?
What causes people to not update straightforwardly
Information being inherently complex / lots of dependences
Socially motivated reasoning (i.e. it’d be disadvantageous for you to believe a thing)
Information not fitting into your existing model (might classify this as “the sort of motivated reasoning that’s more ‘necessary’, i.e. it may be useful to streamline your understanding of the world with models”)
are there other major reasons here?
Note: Robin Hanson’s “Against News” seems kinda relevant here.
What Are We Measuring and Why?
Or, “what exactly is the problem here.”
I notice I have a hard time answering this question (or at least, my cached answers don’t actually deal with it properly. Thanks Ruby (and How to Measure Anything) for reminding me of this approach.
Here is a bad answer to this question, will try writing a good one later maybe.
Basically, it seems important for “the right people” (i.e. people who are actually stakeholders in a given decisionmaking structure, or who’d put actual work into open-ended problems if they were convinced it were important) to be able to be alerted to concepts or arguments that are relevant to them.
The idealized measurement is something like “if people were able to look back with perfect information from an idealized self 100 years in the future, they’d think that they were properly alerted about things that were in retrospect important, and and not overly alerted/attention-hogged on things that were relatively less important.”
Another aspect of the idealized measure would be something like “100 years from now, with perfect information and full integration of that information, you’d think that ideas that should have been important to other people were properly communicated to them, and they took appropriate actions.
With perhaps a third aspect of “in the moment, you also think the various feedback cycles that propagated information and gave bits of info on how you (and others) are responding to that information, were accurate/helpful.
Some desiderata (i.e. what does “better” look like?)
Note: I wrote this before writing the previous section
You can tell at a glance (or more easily) who’s effortposts* are worth reading
note: using effortpost as shorthand for “put a lot of goodfaith effort into communicate their idea, via writing or otherwise)
The people who are coordinating on a given thing who actually matter reliably read, talk with or listen, if-and-only-if it’s a good use of their time (this last clause obviously is doing a hell of a lot of work here. Time, attention, and information are all precious)
People in positions of influence or power have the ability to update on information that is true/useful (insofar as this is useful)
People who write effortposts get better feedback, and/or get better at noticing or being calibrated on feedback (in particular, if Bob reads Alice’s effortpost and goes “hmm, maybe”, and the result is slowly, subtly shifting his mind over years, (perhaps not exactly the way Alice intended), Alice gets more/better feedback that this is going on.