tailcalled.bsky.social
344 posts
242 followers
147 following
Getting Started
Active Commenter
comment in response to
post
^ on men's attraction to female ambition
comment in response to
post
1. My brain has ceased functioning
2. My heart no longer pumps blood
3. I am no longer breathing
comment in response to
post
The decisive refutation is seeing how human agency relies on grace and revelation from God. But that refutation can't be communicated and shouldn't be sought out.
comment in response to
post
Higher moments would be great if the measurement wasn't so terrible. But it is, so they aren't gonna work.
comment in response to
post
Meep
comment in response to
post
The main problem is that underexploration of some of the periphery leads to big errors in the center. Like overestimating the power of intelligence.
comment in response to
post
Idk I feel like a substantial portion of these tensions are just because they are kinda peripheral to his main focus.
comment in response to
post
The solution to the hard problem of consciousness is that consciousness spontaneously manifests in the absence of sin. This also makes evidence-based policy (of the standard sort that depends on wide consciousness) a problem.
comment in response to
post
I feel like first having a qualitative item assessing the felt needs and then having a Likert item assessing the extent to which one takes care of those needs might be the most appropriate framing. Might be difficult to respond to though.
comment in response to
post
Research must either be built on this backchannel, be qualitative, or be built on other research. Otherwise it is lost from society.
comment in response to
post
I feel like there also runs a backchannel of anecdotes and rumors that people want to make scientific, where really the most accurate discussion would be the original events that led to those rumors, but they might be private and/or have gotten lost in the telephone game.
comment in response to
post
@jdp.extropian.net Issue is it isn't useful if it's not managed correctly. From a purely selfish perspective, the twitter purchase has probably caused Elon Musk more pain than it was worth. Power is a curse to be managed moreso than a blessing.
comment in response to
post
For those people it will function as a sort of universal basic income.
comment in response to
post
If someone insists on selling crap to the state, the procurement office will only let them sell products that are inessential to the operation of the state.
comment in response to
post
With each field generating qualitatively different patterns of causal inference skills.
comment in response to
post
Beware about treating confusion about estimands and causal inference as a unidimensional latent variable. It's rather there are a bunch of different fields of experience that each constitute their own latent generating causal inference skills.
comment in response to
post
Expertise arises from experience. Consensus doesn't need to arise...?
comment in response to
post
They're neutral frames which don't precommit to motor program search being the jard part.
comment in response to
post
... But the state should always have requests for additional products so people can choose to sell them. This could create a lot more flexibility on the procurement side and a lot more ease on the employment side.) Basically a way to get deregulation and UBI in one.
comment in response to
post
If a company wants to sell large bulk of goods, it can find a bunch of citizens who sell on behalf of it. (Of course the citizens don't necessarily have a right to sell just anything. People working in the state have to be willing to buy the relevant kind of thing. ...
comment in response to
post
Anecdotes are evidence
comment in response to
post
I feel like one of the main functions of LLMs is roundabout web curation.
comment in response to
post
(Professionals and lost souls are obviously not the only types of people, but it's the ones I've interacted the most with I think, and the ones I've been thinking the most about lately. I need to get out more.)
comment in response to
post
When they're more mission-driven, I feel like the lost souls tend to shrug off primes very rapidly.
comment in response to
post
And I think lost souls are usually only temporarily aware of the futility, as that awareness tends to lead to an adjustment in approach until they feel like it's feasible again.
comment in response to
post
Some people seem like "lost souls", possessed by some futile mission. When they're more self-aware about the futility, they're highly sensitive to priming, but not necessarily in as 1:1 a way as professional are.
comment in response to
post
LLMs have a very similar role to professionals so I think they're similarly susceptible.
comment in response to
post
I kind of feel like it must depend a lot on the person. Professionals try to empower others (especially their managers), so to succeed they sort of need to very rapidly throw their entire agency around to align with the mission. This probably increases priming-susceptibility.
comment in response to
post
I think the problem with psych methods experts is that it requires there to be a language/framework that the psych experts can use to phrase their research question, and all the frameworks that have been made are about effect size estimation, but psych has more difficult needs than that.
comment in response to
post
Remember that per psychophysics perception is usually logarithmic, so additive linear models of survey data are implicitly multiplicative interacting models.
comment in response to
post
Even though autistic people self-identify as having much worse social skills, they self-identify as sharing these things equally well. Also, while extraverts self-identify as having better social skills, they don't self-identify as sharing these things better.
comment in response to
post
Relatedly, I asked some autistic and allistic people how well they shared skills, time, information, resources and social opportunities with their family, friends, coworkers and romantic partner.
comment in response to
post
I buy this.
comment in response to
post
www.lesswrong.com/s/gEvTvhr8hN...
comment in response to
post
One can't predict them from low-level mechanistic simulation, but one can often recognize them as they get rolling, before they fully hit, and make preparations to mitigate them.
comment in response to
post
www.lesswrong.com/posts/HheNit...
comment in response to
post
(Also another obvious complication is different aspects of number sense are not independent, but instead there's a fairly strong general factor underlying them. Everything here is obv idealized for explanatory purposes.)
comment in response to
post
Also probably you wouldn't wanna do this because it wastes precious time and because realistically the test properties are more complex than 1 general-purpose algorithm + lots of item-specific tricks.
comment in response to
post
But I think meditating on this idea is helpful for understanding test design beyond statistical test theories.
comment in response to
post
Probably you wouldn't want to fully subtract it but only partially since general purpose algorithms are worth something too. And probably you don't want to do this at all because it introduces sketchy incentives for test-takers.
comment in response to
post
This test would then measure whether one's performance on arithmetic originates from number sense rather than general-purpose algorithms.
comment in response to
post
A kind of provocative but illustrative concept: take a test designed in the way I recommend here, intersperse it with a test designed in the way test theory recommends, but then subtract the latter test from the former test.