id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
a94df03b-049b-4302-86c9-b62126850e0e
trentmkelly/LessWrong-43k
LessWrong
[Intro to brain-like-AGI safety] 2. “Learning from scratch” in the brain (Last revised: October 2024. See changelog at the bottom.) 2.1 Post summary / Table of contents Part of the “Intro to brain-like-AGI safety” post series. Having introduced the “brain-like-AGI safety” problem in the previous post, the next 6 po...
50926c95-7e84-4b27-9332-edcf20a146f2
trentmkelly/LessWrong-43k
LessWrong
Rationality Quotes 8 "Like a lot of people in the computer industry, Keith Malinowski had spent his whole life being the smartest person in the room, and like most of his fellows the experience left him with a rather high opinion of his opinions."         -- Rick Cook, The Wizardry Quested "The fact that we can becom...
57511447-0fcb-4a96-981f-e2430984e31e
trentmkelly/LessWrong-43k
LessWrong
Chinese room AI to survive the inescapable end of compute governance This post looks at the problem of how to make a cap on AI intelligence enforceable, using something different from current ideas in compute governance (mandated limitations to computational power). The idea I present is finding an AI technique/archi...
4859f07b-5ebb-4ab2-a5ec-90deaa7ee23c
trentmkelly/LessWrong-43k
LessWrong
Morality is subjective preference, but it can be objectively wrong People are often unwilling to think of ethics as their own preferences, rather than demands from something more transcendent. For instance it’s normal to claim that one really wants to make one choice, but it’s only ethical to make the other. My feelin...
55fd6c0e-8070-4ecd-8546-84f0dbd71f34
trentmkelly/LessWrong-43k
LessWrong
Screen-supported Portable Monitor I broke my MacBook screen a few weeks ago and it didn't end up making sense to repair it. I was considering getting a new computer, but since I almost always use an external monitor maybe a portable external monitor would work? I found a well reviewed 15.6" one for $130, capable of 1...
2f671ae6-eef0-4a27-9d1c-b85b8d25e212
trentmkelly/LessWrong-43k
LessWrong
2015 Less Wrong Study Hall census is open. I forgot about it until after were were past the Hall's birthday, so the survey is late this year and won't run for as long. Nevertheless, for those of you that use the Less Wrong Study Hall, here is this year's census: 2015 LWSH Census and Survey It will close on April 7th...
a8c7a2b8-ceb8-4c10-b6e1-85cc04f6c41a
trentmkelly/LessWrong-43k
LessWrong
Consequentialist Formal Systems This post describes a different (less agent-centric) way of looking at UDT-like decision theories that resolves some aspects of the long-standing technical problem of spurious moral arguments. It's only a half-baked idea, so there are currently a lot of loose ends. On spurious argumen...
bcf82e32-f5d7-4878-b161-470e359a9b9c
trentmkelly/LessWrong-43k
LessWrong
Meetup : San Antonio Meetup - Double Crux Game Discussion article for the meetup : San Antonio Meetup - Double Crux Game WHEN: 22 May 2016 02:00:00PM (-0500) WHERE: 818 nw loop 410 78216 Meetup to discuss rationality and all things LessWrong at Earth Burger. Earth Burger is a claimed "Upbeat fast-food joint supplyi...
9d63eca9-fc14-4c6f-a29a-567d7a966cd3
trentmkelly/LessWrong-43k
LessWrong
Explaining information theoretic vs thermodynamic entropy? What is the best way to go about explaining the difference between these two different types of entropy? I can see the difference myself and give all sorts of intuitive reasons for how the concepts work and how they kind of relate. At the same time I can see w...
a4a5c30e-968d-4bc5-a4de-cb12c31a0020
StampyAI/alignment-research-dataset/arbital
Arbital
Bounded agent A bounded agent is the opposite of an [unbounded agent](https://arbital.com/p/unbounded_agent) and refers to agents that operate under realistic conditions and resource constraints. Bounded agents only have realistic amounts of computing power, RAM, and disk space (no more than fits into one Hubble volu...
c0a68b5b-2d71-46a2-9e07-109c869b0f67
trentmkelly/LessWrong-43k
LessWrong
Social prerequisites of rationality Summary: it is a prerequisite that you think you are entitled to your own beliefs, your beliefs matter, you think your actions follow your own beliefs and not from commands issued by others, and your actions can make a difference, at the very least in your own life. This may correla...
f298762e-b865-4878-b307-f84e3442f159
trentmkelly/LessWrong-43k
LessWrong
Less Wrong link exchange We've had similar threads before, but not for a while so I thought I'd make one.    Basic rules, share links that are relevant to Less Wrong areas of interest, but aren't worthy of their own post. Please include a brief description with the link. (My own contributions are below.)
17263fde-8b57-403c-bbf1-218ea3003a50
trentmkelly/LessWrong-43k
LessWrong
Let Your Mind Be Not Fixed When I was a teenager, I knew my mind was not like adults. Sure, I was like adults in a lot of ways. I believed things. I trusted things. I knew I was right sometimes. But they didn't change their minds. I did about once a month! Woe was me. How could I ever be as sure and confident as an...
14798273-2401-4903-9fd3-c8aaa867770b
trentmkelly/LessWrong-43k
LessWrong
Is anchoring a reliable cognitive bias? Say you get a large group of people and split them into three groups. The first group gets this question: > What is the length of the Mississippi river? The second group gets these questions: > Is the Mississippi river greater or less than 70 miles long? > > What is the len...
22819fa6-b3f7-4b70-b8d6-9b86995911da
trentmkelly/LessWrong-43k
LessWrong
Understanding new terms via etymology ---------------------------------------- * This is a quick snapshot showing how I come to understand unfamiliar, key terminology. * If you're lucky enough to encounter confusion, then the process goes; Encounter an unfamiliar term, look at etymology of said term, encounter info...
0603ffc9-42bf-4a44-bb87-dc2a3a687e75
trentmkelly/LessWrong-43k
LessWrong
Abstraction As Symmetry and Other Thoughts Epistemic status: slightly pruned brainstorming. I've been reading John Wentworth's posts on the natural abstraction hypothesis recently, and wanted to add some of my own thoughts. These are very much "directions to look in" rather than completed research, but I wanted to sh...
e3da45d7-96be-44fb-85bd-949da4057808
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Seeking Student Submissions: Edit Your Source Code Contest At EA UC Berkeley, we’re launching an ongoing series of contests called the Artificial Intelligence Misalignment Solutions (AIMS) series. This third contest, [Edit Your Source Code](https://eaberkeley.com/edit-your-own-source-code), is an AI Safety sci-fi crea...
23f799c7-579c-43a7-a7e6-532d136674ae
trentmkelly/LessWrong-43k
LessWrong
On Being Okay with the Truth On January 11, 2007, I timidly whispered to myself: "There is no God." And with that, all my Christian dreams and hopes and purposes and moral systems came crashing down. I wrote a defiant email to the host of an atheist radio show I'd been listening to: > I was coming from a lifetime h...
49781ee1-3abd-4973-bb54-2bdc21d87e39
trentmkelly/LessWrong-43k
LessWrong
Correlated decision making: a complete theory The title of this post most probably deserves a cautious question mark at the end, but I'll go out on a limb and start sawing it behind me: I think I've got a framework that consistently solves correlated decision problems. That it is, those situation where different agent...
7d96b29e-be0a-46f2-9c23-b43fb03a4966
trentmkelly/LessWrong-43k
LessWrong
References & Resources for LessWrong A list of references and resources for LW Updated: 2011-05-24 * F = Free * E = Easy (adequate for a low educational background) * M = Memetic Hazard (controversial ideas or works of fiction) Summary Do not flinch, most of LessWrong can be read and understood by people with a ...
571db138-b381-400a-8526-3bdbc185e456
trentmkelly/LessWrong-43k
LessWrong
Poetic Methods II: Rhyme as a Focusing Device As promised in the previous instalment on meter, let’s explore rhyming from a methodological perspective. The first difference between meter and rhyme lies in their opposite obviousness: the first one is subtle, requiring a learned and attuned ear; the second is so sonoro...
5a0b1a0f-b653-4d90-8a73-94ab509cfcf9
trentmkelly/LessWrong-43k
LessWrong
Will humans build goal-directed agents? In the previous post, I argued that simply knowing that an AI system is superintelligent does not imply that it must be goal-directed. However, there are many other arguments that suggest that AI systems will or should be goal-directed, which I will discuss in this post. Note t...
78a54407-07b8-42e9-8948-7f638565db26
trentmkelly/LessWrong-43k
LessWrong
[AN #62] Are adversarial examples caused by real but imperceptible features? Find all Alignment Newsletter resources here. In particular, you can sign up, or look through this spreadsheet of all summaries that have ever been in the newsletter. I'm always happy to hear feedback; you can send it to me by commenting on t...
1e2377df-c2e8-42f1-ad4b-884921037081
StampyAI/alignment-research-dataset/lesswrong
LessWrong
A Critique of The Evidentialist's Wager *This post critiques* [*The Evidentialist's Wager*](https://globalprioritiesinstitute.org/wp-content/uploads/2019/MacAskill_et_al_Evidentialist_Wager.pdf) *by* [*William MacAskill*](https://www.lesswrong.com/users/wdmacaskill)*, Aron Vallinder, Carl Shulman,* [*Caspar Österheld*...
871feafe-3844-4268-88a2-028aaf566c57
StampyAI/alignment-research-dataset/lesswrong
LessWrong
The Unified Theory of Normative Ethics **Epistemic status:** Reframing of a trivial result, followed by a more controversial assertion. **Summary:** For bounded agents acting in complex environments, using raw consequentialist reasoning for every moment-to-moment decision is prohibitively costly/time-inefficient. A r...
6778894d-f83c-412f-8fee-90d3c4d03d3d
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Clarifying wireheading terminology *See also:* [*Towards deconfusing wireheading and reward maximization,*](https://www.lesswrong.com/posts/jP9cKxqwqk2qQ6HiM/towards-deconfusing-wireheading-and-reward-maximization) [Everett et al. (2019)](https://arxiv.org/pdf/1908.04734.pdf). There are a few subtly different things ...
9cb3e43f-6cc2-4c34-a5b8-ca38fc42aaa6
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups This summary was posted to LW Main on January 30th. The following week's summary is here. New meetups (or meetups with a hiatus of more than a year) are happening in: * Czech's first Meetup Prague: 16 February 2015 06:00PM Irregularly scheduled Less Wrong meetups are taking place in: * Bangalor...
dd4a2c00-ea10-45d7-ac04-61e839f04983
trentmkelly/LessWrong-43k
LessWrong
2020 Review: The Discussion Phase LessWrong is currently doing a major review of 2020 — looking back at old posts and considering which of them have stood the test of time. There are three phases: * Preliminary Voting (completed) * Discussion (ends Jan 11th) * Final Voting (ends January 25th) We're now in the Di...
b41bb26a-f48a-4651-9b42-3d54593a447d
trentmkelly/LessWrong-43k
LessWrong
How evals might (or might not) prevent catastrophic risks from AI Epistemic status: Trying to synthesize my thoughts as someone without any private information about evals. I expect this to be wrong in several ways, lacking detail, and missing a lot of context on specific things that AGI labs and independent teams ar...
9f48e788-cde3-4da5-9a00-8162af455ecd
trentmkelly/LessWrong-43k
LessWrong
Pascal's Mugging for bounded utility functions Followup to: Pascal's Mugging: Tiny probabilities of vast utilities; The Lifespan Dilemma This is Pascal's Mugging: Someone comes to you and says, "Give me five dollars, and I'll use my powers from outside the matrix to grant you 4^^^^4 years of fun." And they're lying, ...
af978835-7a79-4108-aa9a-6682467e6463
trentmkelly/LessWrong-43k
LessWrong
Features and Adversaries in MemoryDT Keywords: Mechanistic Interpretability, Adversarial Examples, GridWorlds, Activation Engineering This is part 2 of A Mechanistic Interpretability Analysis of a GridWorld Agent Simulator Links: Repository, Model/Training, Task. Epistemic status: I think the basic results are pret...
24a87451-5d8a-42af-8c4d-1dfc1fc08aac
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Steganography and the CycleGAN - alignment failure case study *1. This is a (lightly edited) transcript of a lightning talk I gave during the* [*Virtual AI Safety Camp 2022*](https://aisafety.camp/)*. The original talk in video format can be found* [*here*](https://drive.google.com/file/d/1v-nfm6aAyGn3zGQbXc8sd2VJ6XEL...
5df4e9b9-ad2e-42d9-b184-4e853dfe010a
trentmkelly/LessWrong-43k
LessWrong
[Fiction] [Comic] Effective Altruism and Rationality meet at a Secular Solstice afterparty (Both characters are fictional, loosely inspired by various traits from various real people. Be careful about combining kratom and alcohol.)
988c271e-edce-4143-b45d-ee4321e7f0c4
trentmkelly/LessWrong-43k
LessWrong
Petrov Day Ritual: Coronavirus Edition Ozy adapts the rationalist Petrov Day ritual for 2020.
4c38e79e-c856-4eff-9392-8a3a118cd556
trentmkelly/LessWrong-43k
LessWrong
Separate comments feeds for different post listings views? (Inspired by discussions like this one, but also a thing that I’ve thought for a while would be a good idea.) Currently, there’s a single “recent comments” feed, which includes comments posted to any post—frontpage, Meta, personal blogs, Alignment Forum… even...
5a9343e2-86ff-459f-bb40-2ec9009e5d51
trentmkelly/LessWrong-43k
LessWrong
Image GPT My hot take: Not too surprising to me, considering what GPT-3 could do. However there were some people (and some small probability mass remaining in myself) saying that even GPT-3 wasn't doing any sort of reasoning, didn't have any sort of substantial understanding of the world, etc. Well, this is another n...
464537ba-4814-4b7c-bfdb-2ea80027d9d0
trentmkelly/LessWrong-43k
LessWrong
[Simulators seminar sequence] #2 Semiotic physics - revamped Update February 21st: After the initial publication of this article (January 3rd) we received a lot of feedback and several people pointed out that propositions 1 and 2 were incorrect as stated. That was unfortunate as it distracted from the broader argument...
c51e0aa3-8dde-424f-b9d5-a221071be380
trentmkelly/LessWrong-43k
LessWrong
Emotions = Reward Functions ... Or more specifically, a post on how, and why, to encode emotions to find out more about goals, rationality and safe alignment in general.   If one tries to naively fit reinforcement learning’s reward functions back onto the human mind, the closest equivalent one may find is emotions....
92b58d54-9b57-4b0a-977e-8c3fec13d77e
trentmkelly/LessWrong-43k
LessWrong
Regarding South Africa The system prompt being modified by an unauthorized person in pursuit of a ham-fisted political point very important to Elon Musk once already doesn’t seem like coincidence. It happening twice looks rather worse than that. A Funny Thing Happened on Twitter In addition to having seemingly banne...
85eea5fd-fd15-49b6-95ec-f4b59f4745af
trentmkelly/LessWrong-43k
LessWrong
Three Approaches to "Friendliness" I put "Friendliness" in quotes in the title, because I think what we really want, and what MIRI seems to be working towards, is closer to "optimality": create an AI that minimizes the expected amount of astronomical waste. In what follows I will continue to use "Friendly AI" to denot...
ba40816f-1f96-4191-9ff9-841b57416739
trentmkelly/LessWrong-43k
LessWrong
Religion, happiness, and Bayes Religion apparently makes people happier. Is that evidence for the truth of religion, or against it? (Of course, it matters which religion we're talking about, but let's just stick with theism generally.) My initial inclination was to interpret this as evidence against theism, in the s...
08613661-418f-49b4-ae78-18b939758575
trentmkelly/LessWrong-43k
LessWrong
Request: Induction, Proof, Experimental mathematics Suppose we made an algorithm capable of forming empirical conjectures for mathematics.  How might such an algorithm discover the principle of mathematical proof?   I would like to see an article on the relevant philosophy and mathematical logic background for this ...
dd0ec1b9-b3bf-432c-ac55-416fd7d8bb3b
trentmkelly/LessWrong-43k
LessWrong
Prediction: Astra Zeneca approval date It's widely publicized that the Trump administration pressured the FDA to release the vaccines earlier by reducing our evidentiary standards in light of the thousands of daily deaths. While not fond of Trump generally, I heartily agree that the FDA should end its vaccine bans ear...
302cbe3d-06d8-4ad2-b086-c906d378a746
trentmkelly/LessWrong-43k
LessWrong
The Transparent Society: A radical transformation that we should probably undergo Edit (May, 2020): The purpose of this document is to examine the end, and ask whether it would be good or stable, the purpose is not to examine the means, to lay a path, or to tally the risks of trying. I don't know how clear that was. I...
41e5bfd1-1b6c-4422-8687-15a05e3db8d7
StampyAI/alignment-research-dataset/special_docs
Other
NeurIPSorICML_bj9ne-by Vael Gates-date 20220324 # Interview with AI Researchers NeurIPSorICML\_bj9ne by Vael Gates \*\*Interview with bj9ne, on 3/24/22\*\* \[Note: This transcript has been less edited than other due to language barriers. The interviewee is also younger than typical interviewees.\] \*\*0:00:00.0 Vae...
ae95396b-a9a2-4c6d-aabb-a9a88387ca95
trentmkelly/LessWrong-43k
LessWrong
How does an infovore manage information overload? I am, and have been for most of my life, an information glutton.  The internet has made my affliction worse by providing me with the equivalent of an unlimited buffet of both nutritious as well as junk food for my brain which never leaves my side.  A fire hose of data ...
f2a9291f-cbbe-4742-99ca-e2609c8b1c75
StampyAI/alignment-research-dataset/lesswrong
LessWrong
A new definition of "optimizer" Here, I provide a new definition of "optimizer", as well as better explain a previous one I gave. I had [previously](https://www.lesswrong.com/posts/5BwADJo44dvLQ4csS/defining-optimizer) posted a definition, but I think the definition I gave was somewhat wrong. It also didn't help that ...
21839716-5b11-4e9e-836b-cbec6af61cab
trentmkelly/LessWrong-43k
LessWrong
On False Dichotomies Epistemic status; wild speculation, thought experiment, with potentially helpful personal benefits and emotional resonance. Imagine you're an algorithm. Imagine that you know what it feels like to be an algorithm. Imagine that you are outputting the results of an algorithm, accumulating several d...
732588c1-f5a1-49b0-bbd1-9acb0501b8b3
trentmkelly/LessWrong-43k
LessWrong
Great Empathy and Great Response Ability As kids, long before our frontal lobes have finished developing, we’ve all had the instinct to tease. Just watch two toddlers interact, you can see the malicious glint in their eye as they discover that calling one of their peers a “poopy head” triggers a devastating emotional ...
b0782e08-ddde-42b6-b7ee-04de137875ac
trentmkelly/LessWrong-43k
LessWrong
Cooperation is optimal, with weaker agents too  -  tldr It's a ‘superrational’ extension of the proven optimality of cooperation in game theory  + Taking into account asymmetries of power // Still AI risk is very real Short version of an already skimmed 12min post 29min version here ---------------------------------...
76fa665c-ebcb-4ce3-8c1d-b7cb0d6abbbe
trentmkelly/LessWrong-43k
LessWrong
Murphy’s Quest Ch 8: False Pentachotomy There’s a powerful spell employed by the leaders of my old world called the Dichotomy. It goes something like this: 1. Pick two unfortunate phenomena. 2. Make an Either/Or statement as if the two phenomena are mutually exclusive. Famous examples include: “Either the One Per...
dc2654e5-68f0-492e-8923-5433148770c2
trentmkelly/LessWrong-43k
LessWrong
Notice when you stop reading right before you understand Here is a bad habit I noticed in myself.  * I encounter something I want to learn about. For example, let's pick transformers.  * I scroll through the tweet/paper/article and get some sort of general gist. "Okay, there's positional encoding, and an encoder-de...
ef449a5a-359d-4b8f-a982-0a15664a28a7
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Book Review: Reinforcement Learning by Sutton and Barto *Note: I originally submitted this for the Slate Star Codex (RIP) book review contest, but given the current drama it probably won't happen for a while (if at all), so I decided to share this review with the LW community in the meantime.* *Cross-posted on my blo...
dca904c7-21d6-41db-9d52-2d3b095d41a9
trentmkelly/LessWrong-43k
LessWrong
Alignment Implications of LLM Successes: a Debate in One Act Doomimir: Humanity has made no progress on the alignment problem. Not only do we have no clue how to align a powerful optimizer to our "true" values, we don't even know how to make AI "corrigible"—willing to let us correct it. Meanwhile, capabilities continu...
ca89770b-5492-4821-9a08-0ad5338cff43
trentmkelly/LessWrong-43k
LessWrong
Remind Physicalists They're Physicalists Weisberg et al. (2008) presented subjects with two explanations for psychological phenomena (e.g. attentional blink). Some subjects got the regular explanation, and other subjects got the 'with neuroscience' explanation that included purposely irrelevant verbiage saying that "b...
68f5c9f7-13dd-4208-b381-6652a229dc82
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
AI-Relevant Regulation: CERN ### Preface This is the second post in a series exploring existing approaches to regulation that seem relevant for thinking about governing AI. The goal of this series is to provide a brief overview of a type regulation or a regulatory body so others can understand how they work and glea...
e8dda6fc-e517-444f-829d-af010941319c
trentmkelly/LessWrong-43k
LessWrong
Narrow AI Nanny: Reaching Strategic Advantage via Narrow AI to Prevent Creation of the Dangerous Superintelligence Abstract: As there are no currently obvious ways to create safe self-improving superintelligence, but its emergence is looming, we probably need temporary ways to prevent its creation. The only way to pr...
ca85caf8-4006-43ac-bc15-314df3558505
trentmkelly/LessWrong-43k
LessWrong
Safe Development of Hacker-AI Countermeasures – What if we are too late? Hacker-AI is AI used in hacking. It is an advanced super-hacker tool (gaining sysadmin access in all IT devices), able to steal access credentials, crypto-keys, and any secret it needs. It could use every available feature on a device. It could a...
6095da29-98f1-4397-a76f-7a4359fa212d
trentmkelly/LessWrong-43k
LessWrong
Blackmail, Nukes and the Prisoner's Dilemma This example (and the whole method for modelling blackmail) are due to Eliezer. I have just recast them in my own words. We join our friends, the Countess of Rectitude and Baron Chastity, in bed together. Having surmounted their recent difficulties (she paid him, by the way...
89729ed7-441a-4396-ba2a-d21d2e59ffac
StampyAI/alignment-research-dataset/blogs
Blogs
Plausible Quantum Suicide **DISCLAIMER: the idea described here stands or tenuous philosophysical ground and should generally *not* be considered worth the risk attempting because it may be wrong; in addition, this plan should *not* be utilized to retroactively justify depression-based suicide — retroactive justificat...
5ef67e7d-b645-4a81-ad75-7d00a60af7a6
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Making Nanobots isn't a one-shot process, even for an artificial superintelligance **Summary:** [Yudkowsky argues](https://www.alignmentforum.org/posts/uMQ3cqWDPHhjtiesc/agi-ruin-a-list-of-lethalities) that an unaligned AI will figure out a way to create self-replicating nanobots, and merely having internet access is ...
a9eb4698-c23b-4deb-ad17-cfe39e3df643
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Goal retention discussion with Eliezer Although I feel that Nick Bostrom’s new book “Superintelligence” is generally awesome and a well-needed milestone for the field, I do have one quibble: both he and Steve Omohundro appear to be more convinced than I am by the assumption that an AI will naturally tend to retain its...
00bc62ef-64d7-4eac-9f2e-6467c33bd059
trentmkelly/LessWrong-43k
LessWrong
The Alignment Community Is Culturally Broken Disclaimer: These are entirely my thoughts. I'm posting this before it's fully polished because it never will be. Epistemic status: Moderately confident. Deliberately provocative title.  Apparently, the Bay Area rationalist community has a burnout problem. I have no idea ...
11d18c99-2b02-4da6-8220-9f79bcf199c4
trentmkelly/LessWrong-43k
LessWrong
Book review: The Explanation of Ideology Book review: The Explanation of Ideology: Family Structure and Social Systems, by Emmanuel Todd. What features distinguish countries that embraced communism from countries that resisted? Why did Islam spread rapidly for a century and a half, then see relatively few changes in...
733f55e8-43e7-4cb0-91ef-cf3e4e520f32
trentmkelly/LessWrong-43k
LessWrong
A Critique of “Utility” Note: I didn't write this essay, nor do I own the blog where it came from. The essay text is displayed below this line. Economists and other social theorists often take the concept of utility for granted. In theory, utility is a measure of value. I will denote the utility of X to Y as U(X, Y)...
dc4efef0-29bc-48ac-aa68-5697ddc1830a
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Hiring engineers and researchers to help align GPT-3 My team at OpenAI, which works on aligning GPT-3, is hiring ML engineers and researchers. **Apply** [**here**](https://jobs.lever.co/openai/98599d5b-2d1d-4127-b9b5-708343c8730b) **for the ML engineer role and** [**here**](https://jobs.lever.co/openai/24bd9cf4-fe95-4...
e6b80e88-82b4-411d-873c-ec27cf749773
trentmkelly/LessWrong-43k
LessWrong
New LW Meetup: São Paulo This summary was posted to LW Main on June 12th. The following week's summary is here. New meetups (or meetups with a hiatus of more than a year) are happening in: * São Paulo, Brazil - Meetup at Base Sociedade Colaborativa: 27 June 2015 01:00PM Irregularly scheduled Less Wrong meetups are...
4299a9ac-94cb-4ad3-8533-7376a3577316
StampyAI/alignment-research-dataset/blogs
Blogs
tabooing "AGI" tabooing "AGI" -------------- i've come to not be a fan of the term "Artificial General Intelligence", and i favor [tabooing](https://www.lesswrong.com/posts/WBdvyyHLdxZSAMmoz/taboo-your-words) it in serious discussions. we can't agree on what it even means; some think it's very remote from current ...
16b13bbd-a175-4635-9414-63349506d6e1
trentmkelly/LessWrong-43k
LessWrong
How would you improve ChatGPT's filtering? I am wondering how Less Wrong would improve ChatGPT's filtering? I'm reading through the comments on breaking OpenAI's filtering, and see plenty of analysis of the weaknesses of the safeguards. There's always the chance that some group could steal ChatGPT's source code and re...
97ab4feb-f245-47f7-8eeb-5c8ca46dc6fd
trentmkelly/LessWrong-43k
LessWrong
Torchwood I don't know if anyone on LW watches DoctorWho/Torchwood but the american Torchwood remake's first episode just came out and the premise is that everyone on earth loses the ability to die: S04 EP1 the "Day of Miracles" Enjoy, the reason i added it to discussion was because of the whole never die concept and...
b9d0cc7a-2905-496a-8d3e-d44ad52c0e97
StampyAI/alignment-research-dataset/special_docs
Other
Quantifying Independently Reproducible Machine Learning P eer review has been an integral part of scientific research for [more than 300 years](https://blogs.scientificamerican.com/information-culture/the-birth-of-modern-peer-review/). But even before peer review was introduced, reproducibility was a primary compone...
4fb1714f-e8d3-41fe-a56c-f6a1f5343b7a
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Agency from a causal perspective *Post 3 of*[*Towards Causal Foundations of Safe AGI*](https://www.alignmentforum.org/s/pcdHisDEGLbxrbSHD)*, preceded by*[*Post 1: Introduction*](https://www.alignmentforum.org/posts/3oQCY4he4zRzF6vQb/introduction-to-towards-causal-foundations-of-safe-agi) and [*Post 2: Causality*](http...
180773dc-54c1-4400-babb-f4a316202226
trentmkelly/LessWrong-43k
LessWrong
Thoughts on moral intuitions The style in the following article is slightly different from most LW articles, due to the fact that I originally posted this on my blog. Some folks on #lesswrong liked it, so I thought it might be liked here as well. ---------------------------------------- Our moral reasoning is ultima...
5a456eda-c26a-462a-92b4-3cf46a98539e
trentmkelly/LessWrong-43k
LessWrong
Stupidity and Dishonesty Explain Each Other Away The explaining-away effect (or, collider bias; or, Berkson's paradox) is a statistical phenomenon in which statistically independent causes with a common effect become anticorrelated when conditioning on the effect. In the language of d-separation, if you have a causal...
8eea0ca7-ec00-4722-8353-7fb3da5d20ca
trentmkelly/LessWrong-43k
LessWrong
Book review: Why we sleep I read this book (by a sleep scientist called Matthew Walker) because I knew that it would tell me to sleep more, and I hoped it would cite enough scary statistics that I'd be likely to actually follow through. Well, it worked - I'm keeping a copy on my bedside table for the foreseeable futur...
7a9398ff-21ac-4425-82d3-b92347955fba
trentmkelly/LessWrong-43k
LessWrong
1. The challenge of unawareness for impartial altruist action guidance: Introduction (This sequence assumes basic familiarity with longtermist cause prioritization concepts, though the issues I raise also apply to non-longtermist interventions.) Are EA interventions net-positive from an impartial perspective — one th...
2d209e83-8069-498a-b11e-243ed4cc4cf6
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
AI Alignment and Our Momentous Imperative to Get It Right by Olle Häggström [Applause] i am very very glad to be here uh thank you for the kind invitation to speak to this audience and and for the kind introduction let me see if i can find my slides very good okay so uh i think some of you uh saw this paper that came ...
e166b30b-ec26-41c8-a85a-4cb51f6483fa
trentmkelly/LessWrong-43k
LessWrong
Notes on Justice (as a virtue) This post examines the virtue of justice. I wrote this not as an expert, but as someone who wants to learn. I hope it will help people who want to know more about this virtue and how to nurture it. Defining what justice is is a major topic of ethical and political philosophy. There is n...
f8ab57be-5bc8-482d-965c-c057858b4e60
trentmkelly/LessWrong-43k
LessWrong
The impossible problem of due process I wrote this entire post in February of 2023, during the fallout from the TIME article. I didn't post it at the time for multiple reasons: * because I had no desire to get involved in all that nonsense * because I was horribly burned out from my own community conflict investiga...
55b69392-5f76-41f4-8387-1dd687f5a60b
StampyAI/alignment-research-dataset/lesswrong
LessWrong
An attempt to understand the Complexity of Values Let's pretend one day you find magic machine that could grant any lifestyle you asked it for. Professional golfer? Done. Star of *Pulp Fiction*? Say no more. Enlightened writer? Easy. Any lifestyle you wished to live could be yours upon demand. But like with any magic...
47095d87-832e-4d58-b7db-c23dc35c3943
trentmkelly/LessWrong-43k
LessWrong
Is "hidden complexity of wishes problem" solved? Can the problem described in the hidden complexity of wishes (at least partially) be considered solved? I believe that current LLMs are perfectly capable of answering a question like "How do I save grandma from a burning house?" without suggesting any unexpected maximum...
90d18a2e-ce21-4805-8bde-5de8eac62302
trentmkelly/LessWrong-43k
LessWrong
AstralCodexTen and Rationality Meetup Organisers’ Retreat — Europe, Middle East, and Africa 2023 TL;DR: Retreat for EMEA meetup organisers, Summer 2023, apply here Application deadline: 31ˢᵗ Oct 2022   Why a retreat? We think we all benefit from having an international community of rationality and ACX meetup organ...
fea01c6e-3597-4af2-ac64-f9d71c332034
trentmkelly/LessWrong-43k
LessWrong
The Golem Anthony Ravenscroft writing on why it is important, in a relationship, to honestly communicate your grievances to the other person: > If you don't present your gripes to the responsible party, you cannot humanly bury those complaints - it's just not possible to "forget" about something that has hurt or stun...
0cee8127-61e7-4cc4-b7d5-3894fb4f6a74
StampyAI/alignment-research-dataset/blogs
Blogs
The Unsatisfactorily Far Reach Of Property The Unsatisfactorily Far Reach Of Property ------------------------------------------ i'm a fervent [intellectual](where-next-piracy.html) [property](cultural-and-memetic-hygiene.html) [abolitionist](cc_-1.html). one of the reasons IP seems repulsively bad is that it give...
05011be2-f589-4169-a2e5-999a0b09ed07
trentmkelly/LessWrong-43k
LessWrong
Is there any scientific evidence for benefits of meditation? Are there good reasons for a typical reader of LessWrong to invest their time and effort into meditation practice? I'm going to do a literature review and will keep adding answers here but contributions from others are more than welcome. Some ground rules fo...
ac60c139-0b85-4c70-bc38-25ae9518b91c
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
AI Safety Gym - Computerphile this episode has been brought to you by fast hosts find out more about them later so I wanted to talk about this paper out of opening eye benchmarking safe exploration in deep reinforcement learning what comes with this there's a blog post so I wanted to explain this because when I saw th...
e363d6c2-0791-4e97-91d4-1b64f132d101
trentmkelly/LessWrong-43k
LessWrong
why do trivial inconveniences matter? [epistemic status: messy. probably approximately one of the major processes for task carrythrough failure.] explorer: why do i care about trivial inconveniences? those sound, uh, trivial speaker: well, most of the time, youre not gonna be thinking shit like this, youre just gonn...
e1327ce3-785d-4b51-bb2f-390481ea5a30
trentmkelly/LessWrong-43k
LessWrong
Idea Generation and Sifting This post on Best Of A Great Lot is a part of a series on the subject of designing a new form of governance. Each piece aims to stand alone, but fits together on the Table of Contents. Previous: A Sketch of Belocracy, Evaluation as Feedback Cycle. Next: The Belocrat. Governance is, at its...
f49296a6-b1de-42a2-ada5-9b0eb990d1fd
trentmkelly/LessWrong-43k
LessWrong
Falling fertility explanations and Israel From Robin Hanson, via TheZvi: > The following 8 social trends plausibly contribute to falling fertility: > More gender equality - More equal gender norms, options, & expectations, have contributed to fewer women having kids. > Higher parenting effort - Expectations for how...
83ce2b16-7580-42db-82c1-d0b2ca703875
StampyAI/alignment-research-dataset/blogs
Blogs
New Transcript: Yudkowsky and Aaronson [![Yudkowsky-Aaronson](http://intelligence.org/wp-content/uploads/2013/05/Yudkowsky-Aaronson.png)](https://docs.google.com/document/d/1JIqzTGNvdLukR0Ce5T2eiv_vxtPO53N3KGCbxVvREtU/pub) In [When Will AI Be Created?](http://intelligence.org/2013/05/15/when-will-ai-be-created/), I re...
a2ba2c73-16a7-4782-81b1-0b33cdd33235
trentmkelly/LessWrong-43k
LessWrong
Stable agent, subagent-unstable A putative new idea for AI control; index here. A reflectively consistent agent is one that is ok with creating copies with the same motivation as itself, or continuing to have the same motivation under self-modification. A reflectively stable agent is one that would only create agent...
17e8978f-71df-4e97-af24-75a0b1e2749b
trentmkelly/LessWrong-43k
LessWrong
[draft] Concepts are Difficult, and Unfriendliness is the Default: A Scary Idea Summary Here's my draft document Concepts are Difficult, and Unfriendliness is the Default. (Google Docs, commenting enabled.) Despite the name, it's still informal and would need a lot more references, but it could be written up to a prop...
e40b68eb-b0ca-4e16-b395-690e741dc781
StampyAI/alignment-research-dataset/lesswrong
LessWrong
The Governance Problem and the "Pretty Good" X-Risk *No longer endorsed.* *Like a child suddenly given awesome strength, I could have pushed too hard, and left the world a broken toy I could never repair. —The Hero of Ages* I hope we create superintelligence, and I hope it does what we want. But this would not suf...
2ff6e116-0836-4fab-b183-aaaf92810c66
trentmkelly/LessWrong-43k
LessWrong
Empowerment is (almost) All We Need Intro What/who would you like to become in a thousand subjective years? or a million? Perhaps, like me, you wish to become posthuman: to transcend mortality and biology, to become a substrate independent mind, to wear new bodies like clothes, to grow more intelligent, wise, wealthy...
a2b6fcea-a15b-48e9-81fd-1f51bcd2a0a1
trentmkelly/LessWrong-43k
LessWrong
Torture vs. Dust vs. the Presumptuous Philosopher: Anthropic Reasoning in UDT In this post, I'd like to examine whether Updateless Decision Theory can provide any insights into anthropic reasoning. Puzzles/paradoxes in anthropic reasoning is what prompted me to consider UDT originally and this post may be of interest ...
96fad431-fe96-47f7-9347-bb876d94f8d3
trentmkelly/LessWrong-43k
LessWrong
Jankily controlling superintelligence When discussing AI control, we often talk about levels of AI capabilities where we think control can probably greatly lower risks and where we can probably estimate risks. However, I think it's plausible that an important application of control is modestly improving our odds of su...
97d13c45-6c57-4f91-b18f-aa9fcdd791d6
trentmkelly/LessWrong-43k
LessWrong
Sean Carroll: Does the Universe Need God? [link] Does the Universe Need God? (essay by Sean Carroll) In this essay, Sean Carroll: * Dissolves the problem of "creation from nothing": > A provocative way of characterizing these beginning cosmologies is to say that "the universe was created from nothing." Much ...
8f65a1e3-7c9f-4eb8-bed5-e529da46307a
trentmkelly/LessWrong-43k
LessWrong
A few remarks about mass-downvoting To whoever has for the last several days been downvoting ~10 of my old comments per day: It is possible that your intention is to discourage me from commenting on Less Wrong. The actual effect is the reverse. My comments still end up positive on average, and I am therefore motivat...
15d20021-022d-42e1-a07f-c79dbf15722c
trentmkelly/LessWrong-43k
LessWrong
How I Think About My Research Process: Explore, Understand, Distill This is the first post in a sequence about how I think about and break down my research process. Post 2 is coming soon. Thanks to Oli Clive-Griffin, Paul Bogdan, Shivam Raval and especially to Jemima Jones for help and feedback, and to my co-author G...
1d172118-0361-41e0-878a-5886eefbbd18
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Nick Bostrom's Q&A on Existential risk and AI no I want to do this as an interactive Q&A so and it any question really that you have the two to discuss with you I guess I want to work them right now which is this book on super intelligence which will look not so much at the person of whether or how long it might take ...