Posts
Comments
To be a bit more explicit. I have some ideas of what it would look like to try to develop this meta-field or at least sub-elements of it, seperate from general rationality and am trying to get a feel for if they are worth pursuing personally. Or better yet, handing over to someone who doesn't feel they have any currently tractable ideas, but is better at getting things done.
This seems wrong but at least resembles a testable prediction.
That we have to get a bunch of key stuff right on the first try is where most of the lethality really and ultimately comes from; likewise the fact that no authority is here to tell us a list of what exactly is 'key' and will kill us if we get it wrong. (One remarks that most people are so absolutely and flatly unprepared by their 'scientific' educations to challenge pre-paradigmatic puzzles with no scholarly authoritative supervision, that they do not even realize how much harder that is, or how incredibly lethal it is to demand getting that right on the first critical try.)
Is anyone making a concerted effort to derive generalisable principles of how to get things right on the first try and/or work in the pre-paradigmatic mode? It seems like if we knew how to do that in general it would be a great boost to AI Safety research.
A bit tangential to the main point, but Woodridge is a misleading tertiary source and while some sphex species display looping behavior under some circumstances it is not as universal or irrational as the Hofstader makes it sound.
As far as I know this is the origin of the term, but it is worth noting that Woodridge is a misleading tertiary source and actual sphex are not that spexish.
As a heads up it looks like you need to be logged in to use the contact-newsroom page.
One benefit is that it is a good chance for training. We have a complicated real world question were the answer and even the best way of approaching the answer are currently unknown, but we will know the answer soon. Making predictions and recording the reasoning will allow for retrospectives.
it should be hard or impossible for participants to look up the answers during the study
I am unclear how you are going to enforce this in practice given that the study will be online and that you're expecting people to spend at least 30 minutes on conversation, which implies a large enough reward that it is worth hunting down answers that can't be found on the front page of Google. The only thing that comes to my mind is asking them to make predictions about the future. Related what will your policy be on participants looking up and/or sharing references relevant to steps in their reasoning? E.g. if one of the questions is about Trump being reelected will participants be allowed to visit 538 during step 1 and/or link their partner to it in step 4?
I didn't have enough time to properly evaluate the statistics portion, but at first glance it looks ok. Nothing seems wrong with the significance tests beyond them being significance tests. IIRC BEST addresses my main issues with them, particularly being able to indicated the absence of an effect in a way that isn't the case for mere non-significance, but I haven't used it in forever and don't have time to brush up on it at the moment.
Its one of the things Val taught. I honestly don't remember much of the details, but "deliberate practice but where you think hard about not Goodhart-ing and practicing the wrong thing"? actually sounds about right.
It was taught at CfAR during the period I think James attended.
Have you been using turbocharging training with him?
A bit off the main topic, but what was the Ketosylent recipe?
Small but important change State every belief as a probability that is less than 100% -> State every belief as a probability that is less than 100% and greater than 0.
I enjoyed this post. Citations would have been strongly appreciated but I think posting this without citations > not posting, so I upvoted anyway. In the spirt of crowdsourced fact checking I semi-randomly picked the Darwin quote and hox genes to briefly research. The Darwin quote is real, and its use in other sources suggests that you used the standard interpretation rather than e.x. deliberately taking it out of context. Your explanation of hox genes is really not how I would have done it, but wikipedia agrees that "The sequence and function of Hox genes is highly conserved" which is what matters for your thesis. In restrospect choosing a highly disclaimed section for fact checking was a bad idea, but in my defense hox genes sounded intersting and had an easy to search keyword.
The run up to this apocalypse seem to have included at least a few generalizable events:
At this point I stopped and mentally assigned a 90% probability that this was about to become singularity thing. I applaud myself.
New posts also start with their user having upvoted them once, so if you have 25 karma, you'll start with 3 points.
I have observed that this upvote does not effect the posters karma, this is a good design, but not clear from the text of this post.
I like this object level review. I think that epistemic spot checks are a good idea overall, and like the focus on extracting the underlying model. However, I am worried that just checking the citations that are actually provided in the book makes this sort of check non-robust against the type of selection bias that Scott describes in part 4 of this post. Do you have any idea's for dealing with that? I'll describe mine after a line break for people who would like to consider the problem.
I think it is worth looking at existing reviews from experts, as well as choosing a small but representative sample of new to you concepts from the book to research. Or possiblity to pick a few claims and look specifically for counter examples.
This is kinda sorta, but not quite what happens with the flexible group I was talking about.
I've been thinking about this lately, in large part because the person I'm dating has health problems that cause us as a couple to be much less reliable than I have a reputation for and this has caused serious pushback from some of my friends when I have been unexpectedly unreliable. And frankly I kinda agree with them because, and this is my actual point, we were still used to scheduling activities based on the idea of me being reliable. To the extent reliability necessary virtue to be accepted by society vs. kinda nice is a binary choice I lean toward the former, but I've realized that how important it is depends a lot on the type of activities you do. Excluding intentionally stupid hypotheticals like "bet on when people will show up" the activities workable for reliable groups are strictly greater than the set of activities available to less reliable groups, which is why I value reliability, but I think unreliable groups can still be pretty good if you are well calibrated about your group's reliability.
To give some object level examples; having a role playing game with a party of four players can be really fun for reliable groups, but when health issues (mostly my partner's) lead to both of us missing several games on short notice the rest of the group got angry because their quests were basically unplayable with half the party missing (lots of this was plots aimed at our characters or using specialized skills so they couldn't just do the halve the number of enemy NPCs thing). On the other hand I was part of a rationality group that had a soft start time/socializing in a public place at the beginnings of meetings and a double digit number of nominal members. It was very common for large chunks of people to flake or be late, but as long as the organizer got in on time it was fine and she aimed to be early.