norvid_studies
@norvid-studies.bsky.social
and you just need to epistemic luck your belief web into being "good" rather than "bad"
November 21, 2025 at 2:55 AM
and you just need to epistemic luck your belief web into being "good" rather than "bad"
or how you end up being scialabba and thinking reading NPR et al guarantees that you're not weirdly out to sea and vastly over-meta-confident about [seemingly random topic of] origins of covid discussion ... no escaping the object level I suppose, despite every effort, it's all the same algorithm
November 21, 2025 at 2:55 AM
or how you end up being scialabba and thinking reading NPR et al guarantees that you're not weirdly out to sea and vastly over-meta-confident about [seemingly random topic of] origins of covid discussion ... no escaping the object level I suppose, despite every effort, it's all the same algorithm
well same with groks weights, basically. do you distrust grok on the dates of the battle of tours? doesn't seem sensible not to.
November 21, 2025 at 2:51 AM
well same with groks weights, basically. do you distrust grok on the dates of the battle of tours? doesn't seem sensible not to.
sorry to QT this for the 5th time but this is literally exactly this topic just in another thread: 'compromised' sources, fractional read or zero read
bsky.app/profile/norv...
bsky.app/profile/norv...
hmm, i'll 1/2 push back--good blog post by the way-- on the 'idealized epistemic ecosystem|nothing beside' remains POV with the following conceptual cleavage; trusting *a given authors* epistemic intuitions in 'marking your beliefs to market' on theirs, vs trusting *the epistemic ecosystem* of
November 21, 2025 at 2:50 AM
sorry to QT this for the 5th time but this is literally exactly this topic just in another thread: 'compromised' sources, fractional read or zero read
bsky.app/profile/norv...
bsky.app/profile/norv...
but you're NOT "outsourcing" it to grok, you're using grok as an input which gets filtered by "the web" the same as every other input, none of which are entirely trustworthy. if you're marking your beliefs to grok and only grok then yes we hebben een serieus probleem
November 21, 2025 at 2:49 AM
but you're NOT "outsourcing" it to grok, you're using grok as an input which gets filtered by "the web" the same as every other input, none of which are entirely trustworthy. if you're marking your beliefs to grok and only grok then yes we hebben een serieus probleem
*unto. my typing is atrocious today...
November 21, 2025 at 2:46 AM
*unto. my typing is atrocious today...
hmm, i'll 1/2 push back--good blog post by the way-- on the 'idealized epistemic ecosystem|nothing beside' remains POV with the following conceptual cleavage; trusting *a given authors* epistemic intuitions in 'marking your beliefs to market' on theirs, vs trusting *the epistemic ecosystem* of
November 21, 2025 at 2:46 AM
once more onto the comments fighting.
your comment above is approaching exactly the position of this other OP in this thread so I'll duplicate my past self for reasons of efficiency
your comment above is approaching exactly the position of this other OP in this thread so I'll duplicate my past self for reasons of efficiency
November 21, 2025 at 2:46 AM
once more onto the comments fighting.
your comment above is approaching exactly the position of this other OP in this thread so I'll duplicate my past self for reasons of efficiency
your comment above is approaching exactly the position of this other OP in this thread so I'll duplicate my past self for reasons of efficiency
yes, that's exactly what I don't agree with. I don't think "bad faith" is a very useful category as normally employed. is wikipedia bad faith, because they transparently lied about some articles disgraced former admin dgerard was squatting? not categorically, no... play it by ear. in other words
November 21, 2025 at 2:45 AM
yes, that's exactly what I don't agree with. I don't think "bad faith" is a very useful category as normally employed. is wikipedia bad faith, because they transparently lied about some articles disgraced former admin dgerard was squatting? not categorically, no... play it by ear. in other words
but it still needs to be unwalled or open to taking claims from anywhere, and the calculation is just benefit > cost in some signal processing sense, not a "is this source propaganda" binary sense
November 21, 2025 at 2:41 AM
but it still needs to be unwalled or open to taking claims from anywhere, and the calculation is just benefit > cost in some signal processing sense, not a "is this source propaganda" binary sense
this sits a bit uneasily with something else I believe, which is that you're mostly outsourcing claim evaluation to other people. so what's going on then. I guess that "you" are kind of like a given graph or cluster, centered on you, and then that web is doing most of the epistemic metabolizing,
November 21, 2025 at 2:41 AM
this sits a bit uneasily with something else I believe, which is that you're mostly outsourcing claim evaluation to other people. so what's going on then. I guess that "you" are kind of like a given graph or cluster, centered on you, and then that web is doing most of the epistemic metabolizing,
to firewall off, whereas you kind of reverse or reweight the prioritization while keeping the same rough picture of what's going on and the mechanisms involved. is my sense from this thread [claim]
November 21, 2025 at 2:37 AM
to firewall off, whereas you kind of reverse or reweight the prioritization while keeping the same rough picture of what's going on and the mechanisms involved. is my sense from this thread [claim]
I think our fundamental 'parting of the priors' around a similar set of primitives, here, is that I see there as fundamentally no viable alternative to 'fighting it out in the comments' exactly because you yourself may be compromised (in other words, wrong) before choosing what information streams
November 21, 2025 at 2:37 AM
I think our fundamental 'parting of the priors' around a similar set of primitives, here, is that I see there as fundamentally no viable alternative to 'fighting it out in the comments' exactly because you yourself may be compromised (in other words, wrong) before choosing what information streams
[LW Voice] I notice I'm allergic
November 21, 2025 at 2:25 AM
[LW Voice] I notice I'm allergic
sorry, which monster?
November 21, 2025 at 2:25 AM
sorry, which monster?
there's a deepfates dot com ?!
November 21, 2025 at 2:24 AM
there's a deepfates dot com ?!
I'm not familiar with that deepfatesism. @deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com ?
November 21, 2025 at 2:22 AM
I'm not familiar with that deepfatesism. @deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com ?
and their motivated reasoning is "doing all the work" in the counterfactual causal story of belief formation, sort of like the anecdote or joke about a president firing economists until he finds one who tells me what he wants (and genuinely believes it). unfortunately I can't prove any of this
November 21, 2025 at 2:20 AM
and their motivated reasoning is "doing all the work" in the counterfactual causal story of belief formation, sort of like the anecdote or joke about a president firing economists until he finds one who tells me what he wants (and genuinely believes it). unfortunately I can't prove any of this
*FOR THE MOTIVATED READER, and here's the last idee fixe I didn't get to slot into above, the "do you have the courage to blame demand" issue of most people getting so called brainwashed into some slop talking point keep firing their sources until they get one that tells them what they want to hear
November 21, 2025 at 2:19 AM
*FOR THE MOTIVATED READER, and here's the last idee fixe I didn't get to slot into above, the "do you have the courage to blame demand" issue of most people getting so called brainwashed into some slop talking point keep firing their sources until they get one that tells them what they want to hear
-propaganda/misinformation/disinformation is harder than it looks, fragile, generally doesn't flip or brainwash people into insane beliefs, the information ecosystem is very free and competitive, sifting information or "claims" from less than pristine sources is a 101 reading the internet skill, {}
November 21, 2025 at 2:18 AM
-propaganda/misinformation/disinformation is harder than it looks, fragile, generally doesn't flip or brainwash people into insane beliefs, the information ecosystem is very free and competitive, sifting information or "claims" from less than pristine sources is a 101 reading the internet skill, {}
-median user and everyone in this conversation behave very differently and this discussion was mostly just abut the latter. for me anyway. for the former it is probably more like 'pure' feed or 'word tiktok'
November 21, 2025 at 2:17 AM
-median user and everyone in this conversation behave very differently and this discussion was mostly just abut the latter. for me anyway. for the former it is probably more like 'pure' feed or 'word tiktok'
toplevel bullet points:
-there is no qualitative distinction between 'propaganda' and 'not' environments, more like a quantitative dimension
-any environment with social graph and comment-reply-notification dynamics on top of feed algorithm dynamics is going to struggle to fool most people most OTT
-there is no qualitative distinction between 'propaganda' and 'not' environments, more like a quantitative dimension
-any environment with social graph and comment-reply-notification dynamics on top of feed algorithm dynamics is going to struggle to fool most people most OTT
November 21, 2025 at 2:16 AM
toplevel bullet points:
-there is no qualitative distinction between 'propaganda' and 'not' environments, more like a quantitative dimension
-any environment with social graph and comment-reply-notification dynamics on top of feed algorithm dynamics is going to struggle to fool most people most OTT
-there is no qualitative distinction between 'propaganda' and 'not' environments, more like a quantitative dimension
-any environment with social graph and comment-reply-notification dynamics on top of feed algorithm dynamics is going to struggle to fool most people most OTT