If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Notes for future OT posters:
1. Please add the 'open_thread' tag.
2. Check if there is an active Open Thread before posting a new one. (Immediately before; refresh the list-of-threads page before posting.)
3. Open Threads should be posted in Discussion, and not Main.
4. Open Threads should start on Monday, and end on Sunday.
Are there major points that MIRI considered to be true 5 years ago but doesn't consider to be true today?
Eliezer is the only staff we still have around from 2010, and I'm not sure what he'd say his biggest updates have been. I believe he's shifted significantly in the direction of thinking that the best option is to develop AI that's high-capability and safe but has limited power and autonomy (e.g., Bostrom's 'genie AI' as opposed to 'sovereign AI'), which is interesting.
I came on at the end of 2013, so I've observed that MIRI staff were very surprised by how quickly people started taking AI more seriously and discussing it more publicy over the last year -- how positive the reception to Superintelligence was, how successful the FLI conference was, etc. Also, I know that Nate now assigns moderate probability to the development of smarter-than-human AI systems being an event that plays out on the international stage, rather than taking most of the world by surprise.
Nate also mentioned on the EA Forum that Luke learned (and passed on to him) a number of lessons from SIAI's old mistakes:
...The concrete list includes things like (a) constantly drive to systematize, automate, and outsource the busywork; (b) always attack the biggest constraint (by contrast, most people seem to have a defaul
Is this a new bias? I haven't seen it mentioned before. Abstract (emphasis mine):
Consumers routinely rely on forecasters to make predictions about uncertain events (e.g., sporting contests, stock fluctuations). The authors demonstrate that when forecasts are higher versus lower (e.g., a 70% vs. 30% chance of team A winning a game) consumers infer that the forecaster is more confident in her prediction, has conducted more in-depth analyses, and is more trustworthy. The prediction is also judged as more accurate. This occurs because forecasts are evaluated based on how well they predict the target event occurring (team A winning). Higher forecasts indicate greater likelihood of the target event occurring, and signal a confident analyst, while lower forecasts indicate lower likelihood and lower confidence in the target event occurring. But because, with lower forecasts, consumers still focus on the target event (and not its complement), lower confidence in the target event occurring is erroneously interpreted as the forecaster being less confident in her overall prediction (instead of more confident in the complementary event occurring—team A losing). The authors identify boundary conditions, generalize to other prediction formats, and demonstrate consequences.
A quick update: My astrobiology posts are still coming, but after conferring with Toggle and being surprised with a massive response to my first blog posts (who linked me onto ycombinator? thousands of views in two days) I am going through a bit more research about the geochemical history of Earth and other solar system objects in an attempt to be as rigorous as possible. I only play an astrobiologist on TV so to speak, day to day my focus is a little smaller.
I only play an astrobiologist on TV so to speak, day to day my focus is a little smaller.
The marketplace of ideas wants you to change focus.
Hah, believe me I have been looking into ways to get professionally involved in astrobiology once my PhD is done (gimme 18 months, cracking fascinating problems in the study of metabolic regulation and systems bio here). It's pretty hard to break into, though there was that massive gift for SETI from that Russian billionaire recently... it's either that, going into metabolic engineering, or continuing with basic Eukaryotic cell biology for me most likely.
I like this part from a MIRI blog article:
The problem isn’t Terminator. It’s “King Midas.” King Midas got exactly what he wished for — every object he touched turned to gold. His food turned to gold, his children turned to gold, and he died hungry and alone.
When people insist on using fictional evidence, at least give them one that matches your concerns.
(It will probably also help that Kind Midas is higher status than Terminator.)
Fifty psychological and psychiatric terms to avoid: a list of inaccurate, misleading, misused, ambiguous, and logically confused words and phrases by Scott O. Lilienfeld1, Katheryn C. Sauvigné, Steven Jay Lynn, Robin L. Cautin, Robert D. Latzman and Irwin D. Waldman
The goal of this article is to promote clear thinking and clear writing among students and teachers of psychological science by curbing terminological misinformation and confusion. To this end, we present a provisional list of 50 commonly used terms in psychology, psychiatry, and allied fields that should be avoided, or at most used sparingly and with explicit caveats. We provide corrective information for students, instructors, and researchers regarding these terms, which we organize for expository purposes into five categories: inaccurate or misleading terms, frequently misused terms, ambiguous terms, oxymorons, and pleonasms. For each term, we (a) explain why it is problematic, (b) delineate one or more examples of its misuse, and (c) when pertinent, offer recommendations for preferable terms. By being more judicious in their use of terminology, psychologists and psychiatrists can foster clearer thinking in their students and the field at large regarding mental phenomena.
TL;DR: what actual useful outcomes are there of political discourse for people who don't have political advocacy as an area of comparative advantage?
I think the ability to discuss politics clearly is one of the biggest successes of rationality, perhaps because politics breaks minds so thoroughly. It is ironic then, that it's not obvious whether this is of any use whatsoever to the majority of us. I don't believe that political advocacy is my area of comparative advantage, and with the general aspie nature of LWers I would assume this probably applies to most of us. Indeed, the main effect of thinking rationally about politics that I can't stand non-rationalist politics anymore. Having the correct political views is of some small benefit, although since no general election has ever been won or lost by one vote, its very unlikely to change anything.
Are there other benefits to having accurate political views? Some people, largely on both the far left and the far right think that total social collapse is likely within a few decades. I think they might be partially right - for instance, in Europe previously fringe parties are becoming mainstream, both communists protesting against aust...
Where are all the vids of the speeches and stuff from the EA global events? Surely they were recorded?
I took part in the Good Judgment Project, a giant prediction market study from Philip Tetlock (of "Foxes and Hedgehogs" theory). I also blogged about my results, and the heuristics I used to make bets:
http://aarongertler.net/good-judgment-project/
I thought it might be of interest to a few people -- I originally learned that I could join the GJP from someone I met at CFAR.
What are you thought rituals when you want to optimise your cognitive processes when you don't have much else to do. For instance, you are walking between office buildings.
For me, I'm going to phase in the following ritual and see if it's helpful. Asking myself:
What was done effectively or what outcomes were useful today?
How might things have been improved or what different ideas do you have?
What questions arising from today might be relevant to tomorrow?
How do you feel about the process and content of your life today?
Have you been radically honest or dishonest today?
The good, the bad, and the ineffective: social programs in America
...Do people know which social interventions work just from hearing about them?
To do a test, we made the following game. We've described ten major US social interventions, and you'll have to guess whether they had a positive effect, no effect or negative effect.
The interventions were taken from those reviewed by the Campbell Collaboration, which brings together all the highest-quality research that's available on major social interventions to decide whether they're effective or not. We chose
Hi! This is my first post. I have a physics/MWI question, and I don't know if LessWrong / Discussion is the right place to post it, so is it OK if I ask it here? Here it goes:
As any other amateur who reads Eliezer’s quantum physics sequence, I got caught up in the “why do we have the Born rule?” mystery. I actually found something that I thought was a bit suspicious (even though lots of people must have thought of it, or experimentally rejected this already.) Note that I'm deep in amateur swamp, and I'll gleefully accept any "wow, you are confused&quo...
What should you be doing right now if you believe that advances in AI are about to cause large-scale unemployment within the next 20 years (ignoring the issue of FAI for the sake of discussion)?
Oh, dear... From Marginal Revolution's comment section:
What if “Satoshi Nakamoto” is an evil AI, and the whole concept of the blockchain was invented to see if it could devise a way to harvest the processing power of billions of computers. Currently they are just doing meaningless (or seemingly meaningless) math problems. But what if the math problems they were doing weren’t meaningless? What if they were trying to solve some sort of physics problem necessary to create wormholes or something?
What if Satoshi Nakamoto is Roko’s Basilisk?
Bentham’s Fallacies, Then and Now by Peter Singer
...Bentham collected examples of fallacies, often from parliamentary debates. By 1811, he had sorted them into nearly 50 different types, with titles like “Attack us, you attack Government,” the “No precedent argument,” and the “Good in theory, bad in practice” fallacy. (One thing on which both Immanuel Kant and Bentham agree is that this last example is a fallacy: If something is bad in practice, there must be a flaw in the theory.)
Bentham was thus a pioneer of an area of science that has made considerable p
For the next two weeks I won't be able to create the Open Threads precisely on monday morning, and with some probability I won't be able to connect at all.
There needs to be someone willing to embark on this admittedly very easy task. Please comment below if you wish to wear the cape for the time being.
What is something that would profoundly surprise you?
If I came in to work tomorrow to discover a large oak tree rooted in my office.
I get the impression (maybe wrong) that most answers to this won't be what you're looking for, but I have no idea what you're looking for.
I found SMBC to be the jester at the court of transhumanism, saying with humor and exaggeration some nuggets of truth... I enjoy it greatly.
What are your thoughts on the following Great Filter hypothesis: (1) Reward-based learning is the only efficient way to create AI. (2) AGI is easy, but FAI is hard to invent because the universe is so unpredictable (intelligent systems themselves being the most unpredictable structures) and nearly all reward functions will diverge once the AI starts to self-improve and create copies of itself. (3) The reward functions needed for a friendly reinforcement learner reflect reality in complex ways. In the case of humans they are learned by trial and error durin...
I don't see why an unfriendly AGI would be significantly less likely to leave a trail of astronomical evidence of its existence than a friendly AI or an interstellar civilisation in general.
Remote Exploitation of anUnaltered Passenger Vehicle by Dr. Charlie Miller and Chris Valasek
...Target – 2014 Jeep Cherokee
The 2014 Jeep Cherokee was chosen because we felt like it would provide us the best opportunity tosuccessfully demonstrate that a remote compromise of a vehicle could result in sending messages thatcould invade a driver’s privacy and perform physical actions on the attacker’s behalf. As pointed out inour previous research [6], this vehicle seemed to present fewer potential obstacles for an attacker. Thisis not to say that other manufact
A few nutrition-related questions:
Why does Soylent 2.0 have so much fat? They appear to be going for 45% of calories from fat, whereas they typical recommendation is 10%-35%.
Why does the Bulletproof stuff include so much saturated fat? It appears that the consensus is that saturated fat significantly increases blood cholesterol and arterial plaque formation - curious why such a deviation here.
I want to model incentives, behaviour and agents. Are there any handy, intuitive, web-based tools for mechanism design?
Here is a theory (fox lens viewpoint) of why many people are under economic duress. The problem is not consumerism but producerism:
In the US there are many ambitious, hardworking people who are very excited about their work and career. Such people view work, achievement, and production as its own reward. They primarily value the intellectual satisfaction and social validation that comes from career success; money is less important.
However, even though they don't necessarily value the money that highly, they still get a lot of it, because they are typically...
This article is interesting to me because I have this belief that weight loss is basically about eating less (and exercising more). And some extremely high percentage of everything said about dieting, etc. beyond that is just irrational noise. And that the diets that work don't work because of the reasons their proponents say they work, but only because they end up restricting calories as a byproduct.
...this study is NOT a blow to low-carb dieting, which can be quite effective due to factors such as typically higher protein and more limited junk food optio
What do the following have in common?
Answer: they would all be quite successful if followed, but they are all difficult enough to follow that people who actually care about results will do better to set different goals that take more account of how human decision-making actually works.
If you eat less and exercise more then, indeed, you will lose weight. (I do not know how reliably you will lose weight by losing fat, which of course is usually the actual goal.) But you don't exactly get to choose to eat less and exercise more; you get to choose to aim to do those things, but willpower is limited and akrasia is real and aiming to eat less and exercise more may be markedly less effective than (e.g.) aiming to reduce consumption of carbohydrates, or aiming to keep a careful tally of everything you eat, or aiming to stop eating things with sugar in, or whatever.
People with plenty of willpower, or unusually fast metabolism, or brains less-than-averagely i...
Yes, the effect of diets on weight-loss is roughly mediated by their effect on caloric intake and expenditure. But this does not mean that "eat fewer calories and expend more" is good advice. If you doubt this, note that the effect of diets on weight-loss is also mediated by their effects on mass, but naively basing our advice on conservation of mass causes us to generate terrible advice like "pee a lot, don't drink any water, and stay away from heavy food like vegetables".
The causal graph to think about is "advice → behavior → caloric balance → long-term weight loss", where only the advice node is modifiable when we're deciding what advice to give. Behavior is a function of advice, not a modifiable variable. Empirically, the advice "eat fewer calories" doesn't do a good job of making people eat fewer calories. Empirically, advice like "eat more protein and vegetables" or "drink olive oil between meals" does do a good job of making people eat fewer calories. The fact that low-carb diets "only" work by reducing caloric intake does not mean that low-carb diets aren't valuable.
I think it has a net negative effect on the global dieting discussion that it contains these superfluous steps to the actual causes of weight loss.... (read more)