Traditions are sets of practices that have survived and passed down through time from generation to generation. They usually weren't arrived at through a process of reasoning, and even if they were, that reasoning is often lost to time. And yet, the very fact that these traditions propagated through time gives us information about their utility... (read more)
AI Control in the context of AI Alignment is a category of plans that aim to ensure safety and benefit from AI systems, even if they are goal-directed and are actively trying to subvert your control measures. From The case for ensuring that powerful AIs are controlled:.. (read more)
The Open Agency Architecture ("OAA") is an AI alignment proposal by (among others) @davidad and @Eric Drexler. .. (read more)
Singluar learning theory is a theory that applies algebraic geometry to statistical learning theory, developed by Sumio Watanabe. Reference textbooks are "the grey book", Algebraic Geometry and Statistical Learning Theory, and "the green book", Mathematical Theory of Bayesian Statistics.
Archetypal Transfer Learning (ATL) is a proposal by @whitehatStoic for what is argued by the author to be a fine tuning approach that "uses archetypal data" to "embed Synthetic Archetypes". These Synthetic Archetypes are derived from patterns that models assimilate from archetypal data, such as artificial stories. The method yielded a shutdown activation rate of 57.33% in the GPT-2-XL model after fine-tuning. .. (read more)
Open Threads are informal discussion areas, where users are welcome to post comments that didn't quite feel big enough to warrant a top-level post, nor fit in other posts... (read more)
| User | Post Title | Wikitag | Pow | When | Vote |
Suffering-focused ethics (SFE) is a family of moral views that give priority to reducing suffering, especially intense suffering. Rather than treating happiness and suffering as fully symmetric, suffering-focused views hold that preventing severe suffering often matters more urgently than creating additional happiness.
Indexical uncertainty is irreducible subjective uncertainty induced by anthropically expecting to be in more than one possible future world.
Suppose you are an AI and shortly 2 instances of you will see "Hello green world" and 3 instances will see "Hello red world". If you are a conscious AI, then you should anticipate seeing a green world with 40% probability, and a red world with 60% probability, and this subjective uncertainty is irreducible because it reflects an objective state of affairs where different versions of you see different things. To obtain full knowledge of this state of affairs obtaining later is to be left with subjective uncertainty now about "what I will see happen next".
On the Many-Worlds Interpretation, quantum uncertainty is irreducible because it is indexical uncertainty; the amplitudes in configuration space diverge into blobs of different measure, and we find ourselves in all blobs, somehow with a weight proportional to the measure of amplitude inside our blob.
Suppose your model of physics says that the universe will eventually dissolve into a sea of particles in thermal equilibrium, and these particles will then exist indefinitely.
Then an exponentially vast supermajority of all observer-moments in this universe take place inside brains that momentarily assemble themselves from the sea of entropy by random fluctuations, and probably dissolve a moment later. These are "Boltzmann brains".
The vast supermajority of Boltzmann brains have relatively disordered experiences; they momentarily see chaos before dissolving again.
So your own, far more ordered and orderly experience in this very moment weighs heavily against the hypothesis "I am a Boltzmann brain". Inder most systems of anthropic reasoning this weighs heavily against the possibility "a supervast majority of all moments of consciousness are vastly less orderly than my own".
This then argues against
Full article:
Since free will is about as easy as a philosophical problem in reductionism can get, while still appearing "impossible" to at least some philosophers, it might make a good training problem for aspiring reductionists. If you consider yourself one of those, consider just reading the free will problem page, and not reading the free will solution pagepages below until you've made a very serious effort of your own.
So your own, far more ordered and orderly experience in this very moment weighs heavily against the hypothesis "I am a Boltzmann brain". InderUnder most systems of anthropic reasoning this weighs heavily against the possibility "a supervast majority of all moments of consciousness are vastly less orderly than my own".
Coordinal: A Postmortem (Ronak Mehta, 2026-05-19)
A reasoning step is "logically valid" when that kind of step never produces a false conclusion from true premises. For example, in algebra, "Add 2 to both sides of the equation" is valid because it only produces true equations from true equations, while "Divide both sides by x"" is invalid because x might be 0. So even if "2x = (y+1)x""", letting x = 0 and y = 2,, the original equation can be true while "2 = y + 1""" is false. But "2x + 2 = (y+1)x + 2""" will be true in every semantic model where the original equation is true.
This then argues against
Indexical uncertainty is irreducible subjective uncertainty induced by anthropically expecting to be in more than one possible future world.
Suppose you are an AI and shortly 2 instances of you will see "Hello green world" and 3 instances will see "Hello red world". If you are a conscious AI, then you should anticipate seeing a green world with 40% probability, and a red world with 60% probability, and this subjective uncertainty is irreducible because it reflects an objective state of affairs where different versions of you see different things. To obtain full knowledge of this state of affairs obtaining later is to be left with subjective uncertainty now about "what I will see happen next".
On the Many-Worlds Interpretation, quantum uncertainty is irreducible because it is indexical uncertainty; the amplitudes in configuration space diverge into blobs of different measure, and we find ourselves in all blobs, somehow with a weight proportional to the measure of amplitude inside our blob.
Suffering-focused ethics (SFE) is a family of moral views that give priority to reducing suffering, especially intense suffering. Rather than treating happiness and suffering as fully symmetric, suffering-focused views hold that preventing severe suffering often matters more urgently than creating additional happiness.
pronoun-resolutionFree will is one of the easiest hard questions, as millennia-old philosophical dilemmas go. Yudkowsky has suggested that aspiring reductionists should try to solve it on their own in advance of reading the LessWrong analysis.
The following posts can be read to set up the problem of "free will" and what constitutes a good solution from a reductionist perspective, without entirely giving away the solution. (When Yudkowsky wrote these posts, he thought he could get away with just leaving it as a practice problem, and some of the posts state that the problem will be left open. However Yudkowsky did eventually find that he needed to write out the whole solution.)
For spoiler posts see free will (solution).
Since free will is about as easy as a philosophical problem in reductionism can get, while still appearing "impossible" to at least some philosophers, it makesmight make a good exercisetraining problem for aspiring reductionists, which they should try on their own - seereductionists. If you consider yourself one of those, consider just reading the main page on free will problem.
These posts should page, and not be readreading the free will solution page until havingyou've made a very serious effort onof your own.
Related Pages: Consciousness, Free Will, Philosophy, Reductionism, .
FreeThe sensation a human being experiences of having many possible choices open to them, which they could at will decide to take.
Many people feel that this sensation's referent feels contradicted by the claim from physics that physics is deterministic. (Or, in some baroque interpretations of quantum mechanics, deterministic plus a special random "collapse" operation which is the only non-linear, non-local, non-unitary, faster-than-light, fundamentally random operation in all of physics.) If our actions are determined by physics, how could they be free?
Yudkowsky has suggested that this is one of the easiest hard questions, as millennia-old philosophical dilemmas go. Though this impossible question is generally considered fullyrelatively simpler starter problems in reductionism, and completely dissolved on Less Wrong, aspiring reductionists should trytherefore make their own run at it to solve it on their own.
The following posts can be read to set upsee how far they get into the problemdepths of "free will" and what constitutes a good solution from a reductionist perspective, without entirely giving away the solution. (When Yudkowsky wrote these posts, he thought he could get away with just leaving it as a practice problem, and some of the posts state that the problem will be left open. However Yudkowsky did eventually find that he needed to write out the whole solution.) solutions.
For spoiler posts see free will (solution).
I think the phrase "more than one possible future world" is misleading and should be changed. The paradigmatic cases of indexical uncertainty (including both given examples of the five-AI-clones case and of Everettian quantum mechanics) involve no certainty about which objective possible world is actual, and only involves self-locating uncertainty about my location within a possible world.
So it would be more accurate to say that it is irreducible subjective uncertainty over multiple centered possible worlds or multiple locations with a given possible world.