The betting-voting distinction is interesting and was on my mind while I was reading it.
So much of this breaks down when the binary nature of the variables involved becomes continuous or at least nonbinary.
It's an example of a more general interest of mine, how structural characteristics of an inferential scenario affect the value of information that is received.
I could also see this being relevant to diagnostic scenarios hypothetically.
I've also had to deal with the IRB a lot as a professor. The retroactive application is extremely weird (although maybe better than nothing?).
This seems like one of those situations that would usually require regular review to err on the side of caution if nothing else. It's worth pointing out there are exceptions though:
Generally those exceptions fall into "publicly observable behavior", which I guess I could see this falling into?
It's ethically unjustified how the whole thing actually happened but I guess I can see an IRB coming to an exemption decision. I would probably disagree with that decision but I could see how it would happen.
In some weird legalistic sense I can also see an IRB exempting it because the study already happened and they couldn't do anything about it. It's such a weird thing to do and IRBs do weird things sometimes.
>I've also had to deal with the IRB a lot as a professor. The retroactive application is extremely weird (although maybe better than nothing?).
I mean I feel like the IRB is mostly dealing with medical stuff. "I want to electrocute these students every week to see if it cures asthma". "No that's too much.. every other week at most". "Great I'll charge up the electrodes"
So if a security researcher rolls in after the fact and says "umm yea so this has to do with nerd stuff, computers and kernels, no humans, and I just want it all to be super secure and nobody gets hacked, sound good" "ok sure we don't care if no people are involved and don't really understand that nerd stuff, but hackers bad and you're fighting hackers"
People can do whatever they think is right, of course (: so there is no "should". My point was that saying that a government should not impose communication restrictions is not necessary right. So, no, people should be happy: if they are happy without freedom, then let them be. If they are unhappy without freedom, let them make a revolution.
Apparently we should hire the Guardian to evaluate LLM output accuracy?
Why are these products being put out there for these kinds of things with no attempt to quantify accuracy?
In many areas AI has become this toy that we use because it looks real enough.
It sometimes works for some things in math and science because we test its output, but overall you don't go to Gemini and it says "there's a 80% chance this is correct". At least then you could evaluate that claim.
There's a kind of task LLMs aren't well suited to because there's no intrinsic empirical verifiability, for lack of a better way of putting it.
I don't think physician compensation per se is a good metric for capturing the effect of lack of providers, because some of the increased costs are due to the bottlenecks in the services per se, in terms of procedure costs and types of procedures offered. I also don't think the number of providers per se under the current regime, without deregulation or reregulation of practice boundaries, is representative of what would happen if there were changes in those boundaries. Adding more optometrists 5 years ago isn't the same as changing what they're allowed to do. It also doesn't address what cost increases would have been without an increase in the number of providers.
9% might also seem pretty big to me if it's out of all spending and doesn't include other provider compensation? What if overall healthcare costs went down, but physician compensation stayed the same? Would that then be a problem because it was an increased proportion of the total costs — fat left to be trimmed, so to speak?
There are many problems that don't have anything to do with providers per se, but I also don't think you can glean much by extrapolating to more of the same, especially compensation per se.
If lack of physicians is leading to an increase in costs, you’d expect to see physicians capturing a large part of that increase. There are situations where that wouldn’t hold, but it requires moving away from the simplest explanation that fits the data.
We don’t have to extrapolate from physician compensation though. We know that providers per capita have increased, but costs have continued to skyrocket. Therefore a lack of providers is not the immediate cause of the increase.
In addition to increasing the number of providers, the scope of practice for non-physician providers has almost universally increased.
All of this doesn’t prove that increasing the number of physicians wouldn’t lower costs some amount, but it does show that the increases over the last 20-30 years requires some other explanation.
Psychotherapy is similar to pharmacotherapy in effect size but longer lasting after discontinuation. Given that the latter has been approved, it's safe to assume psychotherapy would be approved if it were a pill.
I'm not sure how this intersects with the point of the paper, but part of the problem with the Renee Good case (or things like it) in my opinion is that the focus too often is on the actual events at a particular moment, and not what is surrounding it.
I can see some argument, for example, that goes something like "Jonathan Ross was afraid he was going to get hit by a car and misperceived her as trying to ram him when was trying to turn right, so he fired in self-defense." Then there's a subsequent argument about whether it was reasonable for him to think that she was going to ram him, etc.
However, what's missing from this is a broader discussion about whether or not an officer should be putting himself in that position near a car at all, when it might be anticipated that there might be misperceptions about what is happening. Whether the officer is competent enough to perceive the difference between someone turning their car versus trying to ram them (especially at that speed). Whether they should have let medical personnel help afterward.
When you frame a discussion about perceptions of facts at a particular moment, you kind of get into a frameset of thinking that everything was passively happening, and start overlooking how a particular moment came to be and whether or not the real problems are a set of things that happened minutes, days, or weeks beforehand, and what happened in the time period afterward. E.g., instead of asking "did Jonathan Ross murder Renee Good?" you can ask "were Jonathan Ross and his colleagues competent enough to avoid a situation where they might feel justified in shooting someone innocent?"
I guess I feel like this "cultural perception" question often sidesteps more important questions about whether or not what came to be could have been avoided. This gets more deeply into the underlying attitudes or assumptions driving the perceptions one way or another and lets them be addressed more directly.
We do know that it’s been a longstanding policy of DHS for officers not to stand in front of cars on purpose just so they’d have an excuse to fire upon the driver. There was an internal audit in 2014 that called out this exact behavior.
This may be related to your point, but I think another problem is that we focus on isolated events instead of applying systems thinking. Any large scale government system will result in accidental deaths. Amtrak has killed almost 600 people in the last four years. (This is not unique to Amtrak. It’s inherent in any rail system that has crossings at grade: https://www.vice.com/en/article/a-train-driver-talks-about-w....) But as a society we accept that a certain number of bystanders being killed is an acceptable consequence for performance of an important government function.
Law enforcement similarly is inherently dangerous. You can enforce various standards, but fundamentally you have to pick where the set the slider bar on the scale from maximizing law enforcement effectiveness to minimizing accidental casualties.
I think there's a real deficit in research on and understanding motivated cognition, and a lot of blurriness about attitudes versus belief versus perception. I don't just mean anything political, I mean things including physical pain and all sorts of things. When someone states something, it's very difficult to distinguish between "this is honestly what I saw or felt" versus "this is what I wanted to see or feel". When you get into the fact that consensus can be wrong, it leads to all sorts of issues.
It would be nice to have some kind of way to discriminate at what point in the percept -> attitude -> construal chain (which is probably more of a feedback loop) we are.
Before getting to research, I think a more honest attitude towards admitting motivated cognition in oneself and others is appropriate. I may give a spur-of-the-moment remark on a political situation, but at least if someone presses me, I will readily provide more insight on my biases and values. When I take the time to contemplate, I usually try to modify my eventual response to avoid undue bias altogether. Being reminded that motivated cognition is pervasive in all of us should reduce the unintentional-but-convenient faults in our cognition.
So much of this breaks down when the binary nature of the variables involved becomes continuous or at least nonbinary.
It's an example of a more general interest of mine, how structural characteristics of an inferential scenario affect the value of information that is received.
I could also see this being relevant to diagnostic scenarios hypothetically.
reply