This is a short list of common terms and phrases, i.e., jargon used on LessWrong.
ADBOC: Agree Denotationally, But Object Connotatively. Discussion in When Truth Isn't Enough
AFAICT: As Far As I Can Tell
Affect: Mood or emotion as demonstrated in external physical signs.
Affective death spiral: When positive attributions combine with the halo effect in a positive feedback loop.
AGI: Artificial general intelligence
Anti-epistemology: Bad rules for thinking itself, capable of protecting false beliefs.
Beisutsukai: A fictional secret society of Bayesians.
Belief update: What you do to your beliefs, opinions and cognitive structure when new evidence comes along.
Black swan: In the usage of Nassim Nicholas Taleb, a black swan is a rare event whose magnitude is so high as to impact the average of a series. These are characteristic of 'fat-tailed' distributions, as opposed to thin-tailed distributions such as the normal distribution, in which rare events are too unlikely to have a large impact.
Blues, Greens: Roman Empire chariot-racing teams that became part of politics. Used in place of real party names. See Mind-killer.
CEV: Coherent Extrapolated Volition, "In poetic terms, our coherent extrapolated volition is our wish if we knew more, thought faster, were more the people we wished we were, had grown up farther together; where the extrapolation converges rather than diverges, where our wishes cohere rather than interfere; extrapolated as we wish that extrapolated, interpreted as we wish that interpreted."
Clever arguer: Someone skilled at writing convincing-sounding arguments for an existing belief. Inventing clever arguments for a belief does not change the truth value of the belief. Discussed in "The Bottom Line"
Consequentialism: A moral theory that places value on the consequences of actions. Covered in more depth here.
Crisis of faith: What to have when you may have been quite wrong for a long time.
Dark arts: Rhetorical techniques crafted to exploit human cognitive biases. Considered bad behaviour even if the belief you want to communicate is good.
Deontology/ deontological ethics: An approach to ethics that judges the morality of an action based on the action's adherence to a rule or rules. See Wikipedia article on deontological ethics for more. Contrast consequentialism.
EEA: Environment of Evolutionary Adaptedness. An Evolutionary psychology term synonymous with the more commonly-used "ancestral environment". For humans, refers to the state of tribal bands of hunter-gatherers.
Egan's law: "It all adds up to normality." Surprising truths do not make the sky orange and grey; it stays blue.
ETA: Edited To Add (though some would rather you say "Edit:" instead)
EY: Eliezer Yudkowsky
FAI: Friendly AI
FOOM: Onomatopoetic vernacular for an intelligence explosion.
Fully general counterargument: An argument which can be used to discount any conclusion the arguer does not like.
Fuzzies: The desired but less useful counterpart to utils. They make you feel you're altruistic and socially contributing.
Hedon: A unit philosophers use to quantify pleasure. (Note: no actual quantifying is done.)
Hollywood rationality: What Spock does, not what actual rationalists do.
IA: Intelligence augmentation
IAWYC: I Agree With Your Conclusion. Generally used when nitpicking, to make it clear that the nitpicks are not meant to represent actual disagreement. Discussed in Support That Sounds Like Dissent.
I don't know: Something that can't be entirely true if you can even formulate a question.
IMO/IMHO: In my (humble) opinion
Inferential distance: The number of inferences, or intermediate steps, it takes someone to get from their existing knowledge to an understanding of the point you're making. See also illusion of transparency.
ISTM: It Seems To Me
Kolmogorov complexity: Given a string, the length of the shortest possible program that prints it. See also: Solomonoff induction
LCPW: Least convenient possible world. A technique used to prevent oneself from evading the point of a question by nitpicking details.
Logical rudeness: A response to criticism which insulates the responder from having to address the criticism directly, without appearing to be conventional rudeness.
LW: Less Wrong
Meetup: Groups of Less Wrong members sometimes arrange to meet each other in meat space (in person). Some geographic areas have groups that do this regularly.
Mind-killer: A topic that reliably produces biased discussions, e.g. politics or Pick-Up Artists.
MoR: Also HPMoR, Harry Potter and the Methods of Rationality
Motivated cognition: Reasoning used to reach desired conclusions rather than true conclusions.
MWI: Many-Worlds Interpretation, an interpretation of quantum mechanics advocated in Eliezer Yudkowsky's quantum mechanics sequence
NPC: Non-Player Character (think of an MMORPG like World of Warcraft that has characters controlled by humans and characters controlled by a computer; the characters controlled by humans would be PCs (player characters) and the characters controlled by a computer would be NPCs (non-player characters)).
Noncentral fallacy: A rhetorical move often used in political, philosophical, and cultural arguments. "X is in a category whose archetypal member gives us a certain emotional reaction. Therefore, we should apply that emotional reaction to X, even though it is not a central category member."
OB: Overcoming Bias
One-box: One of the choices for Newcomb's problem.
Omega: A hypothetical superintelligent being, canonically found in Newcomb's problem.
Ontology/ontological: The philosophical study of the nature of being, existence, or reality, deals with questions concerning what entities exist or can be said to exist, and how such entities can be grouped, related within a hierarchy, and subdivided according to similarities and differences. See also ontological argument at Wikipedia for an example of (ab)using ontology to try and prove the existence of God.
Paperclip maximizer: An AI that has been created to maximize the number of paperclips in the universe. A hypothetical unfriendly artificial intelligence.
Paranoid debating: A group estimation game in which one player, unknown to the others, tries to subvert the group estimate.
Password: The answer you guess instead of actually understanding the problem. See Guessing the teacher's password
PC: Player Character (think of an MMORPG like World of Warcraft that has characters controlled by humans and characters controlled by a computer; the characters controlled by humans would be PCs (player characters) and the characters controlled by a computer would be NPCs (non-player characters))
PCT: Perceptual control theory
Philosophical zombie or P-Zombie: A creature which looks and behaves indistinguishably from a human down to the atomic level, but is not conscious. See Zombies (sequence)
Priors: What you update from in Bayesian calculations. In practical terms, everything you think you know now.
Privileging the hypothesis: The fallacy of singling out a specific hypothesis for investigation when there isn't enough evidence at hand to select this hypothesis over others. e.g."We have no idea who committed the murder, so let's consider the possibility that Mortimer Q. Snodgrass did it, and investigate him."
Or "The origin of the universe sure is mysterious! Have you considered that it could have been done by the God of the Bible?"
QALY: Quality-adjusted life year; a concept from the economics of health care
Rationalism: The mode of thinking within the rationalist movement
Rationalist taboo: A technique for unpacking words into concepts: taboo the use of a given word or its synonyms. Particularly useful in arguments over definitions.
Reversed stupidity is not intelligence: "The world's greatest fool may say the Sun is shining, but that doesn't make it dark out."
Reality is normal: See also: Egan's Law
Semantic stopsign: A term that looks like an explanation but, on closer examination, doesn't actually explain anything. Also called curiosity stopper.
Shut up and multiply: How to do a utility calculation without scope insensitivity.
Signaling: Conveying information by performing an action which would be costly to perform, if the information were not true.
Solomonoff induction: A formalized version of Occam's razor based on Kolmogorov complexity.
Steel man: A term for the opposite of a Straw Man: the strongest possible form of an opponent's argument, even if they didn't make it themself.
Strong man: Another term for the opposite of a Straw Man: the strongest actual form of an opponents argument.
Taboo the word ... See rationalist taboo.
TAI: Transformative Artificial Intelligence
Teleology: Discussing an event as though it were caused by its future consequences.
tl;dr : Too long; didn't read. Polite use: one-line summary at top of your long article. Impolite use: dismissive response to another's long piece of writing or unparagraphed slab of text.
Topic that must not be named: When LessWrong was started, Eliezer put a temporary moratorium on discussion of the Singularity or AI. You will see this used in old discussions to allude to these topics.
Two-box: One of the choices for Newcomb's problem.
Tsuyoku naritai: Japanese: "I want to become stronger."
uFAI: Unfriendly AI
Ugh field: A subject that is thought about less over time due to behavioral conditioning.
Utility function: A utility function assigns numerical values ("utilities") to outcomes, in such a way that outcomes with higher utilities are always preferred to outcomes with lower utilities.
Utils: Units of utility; sometimes called "utilons". Contrast fuzzies.
Update: See Belief update
Weak Man: The opposite of a Strong Man, and relative to a Straw Man: the weakest version of your opponents actual arguments.
YMMV: Your Mileage May Vary. See Other-optimizing