On the other hand, whenever you do something, you practice it whether you intend to or not.squidious on What I've Learned From My Parents' Arranged Marriage
So glad this could give you hope! :)moridinamael on Subagents, akrasia, and coherence in humans
Have you - or anyone, really - put much thought into the implications of these ideas to AI alignment?
If it's true that modeling humans at the level of constitutive subagents renders a more accurate description of human behavior, then any true solution to the alignment problem will need to respect this internal incoherence in humans.
This is potentially a very positive development, I think, because it suggests that a human can be modeled as a collection of relatively simple subagent utility functions, which interact and compete in complex but predictable ways. This sounds closer to a gears-level portrayal of what is happening inside a human, in contrast to descriptions of humans as having a single convoluted and impossible-to-pin-down utility function.
I don't know if you're at all familiar with Mark Lippman's Folding material and his ontology for mental phenomenology. My attempt to summarize his framework of mental phenomena is as follows: there are belief-like objects (expectations, tacit or explicit, complex or simple), goal-like objects (desirable states or settings or contexts), affordances (context-activated representations of the current potential action space) and intention-like objects (plans coordinating immediate felt intentions, via affordances, toward goal-states). All cognition is "generated" by the actions and interactions of these fundamental units, which I infer must be something like neurologically fundamental. Fish and maybe even worms probably have something like beliefs, goals, affordances and intentions. Ours are just bigger, more layered, more nested and more interconnected.
The reason I bring this up is that Folding was a bit of a kick in the head to my view on subagents. Instead of seeing subagents as being fundamental, I now see subagents as expressions of latent goal-like and belief-like objects, and the brain is implementing some kind of passive program that pursues goals and avoids expectations of suffering, even if you're not aware you have these goals or these expectations. In other words, the sense of there being a subagent is your brain running a background program that activates and acts upon the implications of these more fundamental yet hidden goals/beliefs.
None of this is at all in contradiction to anything in your Sequence. It's more like a slightly different framing, where a "Protector Subagent" is reduced to an expression of a belief-like object via a self-protective background process. It all adds up to the same thing, pretty much, but it might be more gears-level. Or maybe not.cheerfulwarrior on [Question] Tracking accuracy of personal forecasts
Thanks. I will be on the lookout for relevant writings. I'm slowly going through Yudkowsky's books/posts, so I'm sure I will stumble on it sooner or later.ryan_b on A Tale of Four Moralities
I notice that the question of Maxie stealing from her friends is simply dropped.
I find this notion of zero consequences baffling. It wasn't even evaluated.cousin_it on The Game Theory of Blackmail
Threatening to crash your car unless the passenger gives you a dollar is also not credible in the common meaning of the word.
"Both players are UDT" is slightly questionable phrasing, since UDT requires the agent to prove a mapping from their policies to outcomes, which is tricky in asymmetric games like your #3. You probably meant "both players can precommit", which is an okay assumption as long as you talk only about AIs.dagon on The Game Theory of Blackmail
Furthermore, defect-defect is traditionally super bad for both players. But I would not say that this is a necessary condition for something to be a Game of Chicken.
The traditional game of chicken, with cars racing at each other or toward a cliff edge, has likely death in the defect-defect box. If you're considering iterated games, an early death stops the series (and, depending on your modeling of utility, wipes out all prior gains in any other games). I would say this is a necessary condition, and is the primary thing which makes Chicken different from PD.
And this distinction makes modeling it trickier - the game is mostly about the unknown chance that one will be unable to defect when one decides to (due to physical constraints). It's best modeled as a series of decisions, with known ending (death), and increasing chance of accidental defection.mr-hire on Do you like bullet points?
I tend to write a particular type of post in a hierarchy. It seems a natural way to write something purely explanatory, continually making bullet points to define terms, create subcategories, and define nuance until you get to a logical stopping point like common knowledge.
However, I'm wary of people using bullet points to create this hierarchy, as I find that often the use of bullet points means that people didn't think about the ORDER of the knowledge at each level of hierarchy, something which I find really important and is much less likely to happen when using a narrative hierarchy with headings, subheadings, etc. I think there's something about bullets being an unodered list that makes people disregard order as a useful and important teaching tool, and it often means I have to reread the bullets multiple times to get a coherent narrative together of what is trying to be said.
I also really enjoy slatestarcodexy posts that have clear narrative order but not hierarchy. I think they're an excellent way to cover a topic from many angles while maintaining an emotional thread throughout. I think that for me the narrative order is always useful and the hierarchy is sometimes useful.deluks917 on The Amish, and Strategic Norms around Technology
If you can consistently get to work late enough I think the best time to go to sleep is around 1am. 1am is late enough you can be out until midnight and still have an hour to get home and go to sleep on time. Even if you are out very late and only get to bed by 2am you are only down an hour of sleep if you maintain your wakeup time. There is occasional social pressure to hang out substnatially past midnight but it is pretty rare.
For these reasons I go to bet at 1am and get up at 9am. Of course I don't have to be at work until 10am. But if you can make this work its great to have a sleep schedule you can hold to without sacrafices socialization.dagon on Do you like bullet points?
I use bullets almost exclusively when taking notes or writing for myself. When writing for others, I use them as part of a narrative, but rarely the main text. I have gotten feedback that when I over-rely on bullet style lists of points, it's difficult to find a flow in my documents, and I tend to use too much shorthand so some of the points are less compelling than they can be.