id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
8416e437-a1fa-43ec-8e20-4dbdda39efb3 | trentmkelly/LessWrong-43k | LessWrong | Nine Points of Collective Insanity
A scenario detailing how humans automate away other humans, as the beginning of the economic decoupling of the AI-hardware economy from the human-wetware economy.
Excerpts compiled below:
> * Where/first it begins with the customer service people – they are automated away (since ... |
8f3c30c2-3f90-491d-aab3-274e263bb253 | trentmkelly/LessWrong-43k | LessWrong | EA might systematically generate a scarcity mindset that produces low-integrity actors
Epistemic status: Highly speculative quick Facebook post. Thanks to Anna Riedl for nudging me to share it here anyway.
Something I've noticed recently is that some people who are in a bad place in their lives tend to have a certain... |
4b66e57e-195d-47ba-b504-334093badd79 | trentmkelly/LessWrong-43k | LessWrong | A practical out-of-the-box solution to slow down COVID-19: Turn up the heat
3 days ago there was published a research that swayed me significantly towards the opinion that COVID-19 is spreading better in cold (but not too cold) climates. I believe that the odds for this are probably over 50%, and if it's true there is... |
41c87c54-e539-4c0a-b4db-f670b19adb7c | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Self-Limiting AI in AI Alignment
I'm writing this post because something occurred to me which I'm sure doesn't eliminate the possibility of an AI-caused X-risk, but which may mitigate it. Now, I should admit this is a rather naive speculation, as I'm not an expert on the field of AI safety; however, I thought it was w... |
50078050-0995-4285-958e-92399f101cb9 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] The Wonder of Evolution
Today's post, The Wonder of Evolution was originally published on 02 November 2007. A summary (taken from the LW wiki):
> The wonder of the first replicator was not how amazingly well it replicated, but that a first replicator could arise, at all, by pure accident, in the primor... |
b204cbc7-c675-4a89-85d0-236a20fb0295 | trentmkelly/LessWrong-43k | LessWrong | Why would an AI try to figure out its goals?
> "So how can it ensure that future self-modifications will accomplish its current objectives? For one thing, it has to make those objectives clear to itself. If its objectives are only implicit in the structure of a complex circuit or program, then future modifications are u... |
b6a58391-da6d-459a-898e-fccf3aa1d808 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Safe Imitation Learning via Fast Bayesian Reward Inference from Preferences
1 Introduction
---------------
It is important that robots and other autonomous agents can safely learn from and adapt to a variety of human preferences and goals. One common way to learn preferences and goals is via imitation learning, in ... |
57627163-23ad-4602-8e30-7fae28ccae0d | trentmkelly/LessWrong-43k | LessWrong | Conditioning Predictive Models: Interactions with other approaches
This is the fifth of seven posts in the Conditioning Predictive Models Sequence based on the paper “Conditioning Predictive Models: Risks and Strategies” by Evan Hubinger, Adam Jermyn, Johannes Treutlein, Rubi Hudson, and Kate Woolverton. Each post in ... |
7eb00206-a28c-4116-959f-980ebe61e437 | trentmkelly/LessWrong-43k | LessWrong | Qualia Research Institute: History & 2021 Strategy
Key Points
* Theoretical & empirical work on emotional valence is Qualia Research Institute’s core research focus. At a high level, Qualia Research Institute (QRI) works to formulate universal theories of valence[1], test those theories experimentally, and build non-... |
6dba6d15-7c21-4a18-91a7-f20a60b2470e | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Can we learn much by studying the behaviour of RL policies?
Economists sometimes study [revealed preferences](https://en.wikipedia.org/wiki/Revealed_preference), which are preferences that we can infer from choices, e.g. when given the choice between an apple or an orange, if I choose an apple, then I have revealed a ... |
c7ce6d4a-16a3-4a5e-bd95-9ffe8437b823 | trentmkelly/LessWrong-43k | LessWrong | Decission Theory and Quantum Self-Measurement
Most Quantum Mechanics are formulated from a point of view of Causal Decission Theory, as can be seen from the Noncontextuality/counterfactual indefiniteness of Kochen-Specker Theorem.
Now, this creates a chatastrophe when the Observator needs to be analized through a phy... |
8eae313d-962f-439a-9f0e-98ca51cf5279 | trentmkelly/LessWrong-43k | LessWrong | Image vs. Impact: Can public commitment be counterproductive for achievement?
The traditional wisdom says that publicly committing to a goal is a useful technique for accomplishment. It creates pressure to fulfill one's claims, lest one lose status. However, when the goal is related to one's identity, a recent study... |
4bfe7768-f471-4b2c-bc67-5f36c6abd0ae | trentmkelly/LessWrong-43k | LessWrong | The Robots, AI, and Unemployment Anti-FAQ
Q. Are the current high levels of unemployment being caused by advances in Artificial Intelligence automating away human jobs?
A. Conventional economic theory says this shouldn't happen. Suppose it costs 2 units of labor to produce a hot dog and 1 unit of labor to produce ... |
91efd5f4-b0bf-4aff-9088-92ef429a2150 | trentmkelly/LessWrong-43k | LessWrong | [LINKS] New GiveWell Top Charities
As of today, GiveWell has two new top charities:
1. Against Malaria Foundation
2. Schistosomiastis Control Initiative
Their previous top charity, Village Reach, no longer has much room for more funding. Which is a good thing: partly through givewell's influence they now have as m... |
ff3e24c1-722f-4c95-a07c-0e4b76e9f7a2 | trentmkelly/LessWrong-43k | LessWrong | Silliness
Before I saw Everything Everywhere All at Once all I knew was:
1. It's about a Chinese-American family that runs a laundromat.
2. It's the best movie ever.
Spoilers Ahead
Everything Everywhere All at Once is a silly movie.
What is silliness? Silliness is the opposite of seriousness. What is seriousness... |
ed7964d5-4aa8-4104-a831-cdbf783fbe73 | trentmkelly/LessWrong-43k | LessWrong | Logical Counterfactuals are low-res
Cross-posted from my blog. Related to my several previous posts.
(Epistemic status: I have no idea why such an obvious observation is never even mentioned by the decision theorists. Or maybe it is, I have not seen it.)
A logical counterfactual, as described by Nate Soares:
> In a... |
e1c91a48-dece-4142-9cd8-9adc97e14ad5 | trentmkelly/LessWrong-43k | LessWrong | Three Subtle Examples of Data Leakage
This is a description of my work on some data science projects, lightly obfuscated and fictionalized to protect the confidentiality of the organizations I handled them for (and also to make it flow better). I focus on the high-level epistemic/mathematical issues, and the lived exp... |
2e3f4ed4-626e-4e56-adc6-14b5abab3d66 | trentmkelly/LessWrong-43k | LessWrong | The metaphor/myth of general intelligence
Thanks for Kaj for making me think along these lines.
It's agreed on this list that general intelligences - those that are capable of displaying high cognitive performance across a whole range of domains - are those that we need to be worrying about. This is rational: the mos... |
85bf75ff-71da-482a-a2bc-8623baaf64f1 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | "Tech company singularities", and steering them to reduce x-risk
The purpose of this post (also available on [LessWrong](https://www.lesswrong.com/posts/ezGYBHTxiRgmMRpWK/tech-company-singularities-and-steering-them-to-reduce-x)) is to share an alternative notion of “singularity” that I’ve found useful in timelining/f... |
aba58193-5049-4fc4-94f2-e32ea4aa7c23 | trentmkelly/LessWrong-43k | LessWrong | Evaluating Multiple Metrics (where not all are required)
This is my first article, and I'm submitting it in the discussion forum, so hopefully I've done this correctly and we can discuss!
Anyway, I have a group of friends who are really interested in movies, and they feel very strongly about them. I find their convi... |
55c9fe04-d7da-48e7-a1b8-815043062366 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington DC meetup
Discussion article for the meetup : Washington DC meetup
WHEN: 29 April 2012 03:00:00PM (-0400)
WHERE: Portrait Gallery, central plaza
There will be Zendo; or as dubbed at the last meetup "Science: the game"
Discussion article for the meetup : Washington DC meetup |
22cd9ccb-65bf-4e16-ad04-78c75e71f681 | trentmkelly/LessWrong-43k | LessWrong | Trust
One of the more important types of decisions is deciding who to trust and about what, but I don't think I've seen any discussion here about rational methods in that area.
Any suggestions? |
4825c705-d11c-49c3-8be1-adeec634d6e8 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "I always wash my hands but typically not very well; I never worked in healthcare and never got trained. Given the current situation I decided to actually research how to do it properly.
Different sources differ slightly in their recommendations but the basic gist is the same. Some takeaways: Washing your hands should ... |
29eb4a02-26b3-4b85-93ca-8cc17460730b | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Caution when interpreting Deepmind's In-context RL paper
Lots of people I know have had pretty strong reactions to the [recent Deepmind paper](https://arxiv.org/abs/2210.14215), which claims to have gotten a transformer to learn an RL algorithm by training it on an RL agent's training trajectories. At first, I too was... |
098e4052-fe13-44dd-a50e-e4f437129721 | trentmkelly/LessWrong-43k | LessWrong | The new dot com bubble is here: it’s called online advertising
If you want to understanding Goodharting in advertising, this is a great article for that.
At the heart of the problems in online advertising is selection effects, which the article explains with this cute example:
> Picture this. Luigi’s Pizzeria hires ... |
4c844d7e-affe-4ae2-b319-a52fd9f89d8c | trentmkelly/LessWrong-43k | LessWrong | Prepsgiving, A Convergently Instrumental Human Practice
Most cultures have a harvest festival, and every harvest festival is basically automatically a Prepsgiving Celebration.
In the northern hemisphere this probably happens in November or October, and in the southern hemisphere it probably happens in May or June. Th... |
43f75f66-7398-4535-974b-35021491862e | trentmkelly/LessWrong-43k | LessWrong | A Less Mysterious Mindfulness Exercise
At the beginning of the year, I stumbled upon a self-help book based on the ideas of Acceptance and Commitment Therapy (ACT). I assimilated it insights and started practicing a simple exercise and quickly noticed its effectiveness, to the point of becoming convinced that I had fo... |
a4acdd56-a286-4dca-a13f-d8abe7f4118b | trentmkelly/LessWrong-43k | LessWrong | Last call for submissions for TAIS 2024!
We are very close to the submission deadline for TAIS 2024. Alignment research moves fast; if there's something you've been working on the for last few months that you think would be interesting to our attendees, please submit! You can find a draft agenda on our website.
Time... |
a2418067-a0b8-4015-a1d4-4afc83de2255 | trentmkelly/LessWrong-43k | LessWrong | There's No Fire Alarm for Artificial General Intelligence
What is the function of a fire alarm?
One might think that the function of a fire alarm is to provide you with important evidence about a fire existing, allowing you to change your policy accordingly and exit the building.
In the classic experiment by Latane ... |
cc4693e8-1367-484b-9a08-878155cc97f5 | trentmkelly/LessWrong-43k | LessWrong | Experiences raising children in shared housing
Crosspost from The Whole Sky.
Sometimes I see posts about people’s hope to raise children in a group housing situation, and it often seems overly optimistic to me. In particular they seem to expect that there will be more shared childcare than I think should be expected.
... |
b9c7ada1-a866-4e99-92ef-50be108b4885 | trentmkelly/LessWrong-43k | LessWrong | Making Vaccine
Back in December, I asked how hard it would be to make a vaccine for oneself. Several people pointed to radvac. It was a best-case scenario: an open-source vaccine design, made for self-experimenters, dead simple to make with readily-available materials, well-explained reasoning about the design, and wi... |
78182ad7-3bd9-466f-8120-a821cce4834d | trentmkelly/LessWrong-43k | LessWrong | Coherent extrapolated dreaming
This work was supported by the Monastic Academy for the Preservation of Life on Earth. You can support my work here.
I will give a short presentation of this work followed by discussion on Wednesday Dec 28 at 12pm Pacific / 3pm Eastern. RSVP here.
Outline
* I have four questions abov... |
88dbedce-81d7-4d1d-8454-0a4d786717d3 | trentmkelly/LessWrong-43k | LessWrong | An Overview of Sparks of Artificial General Intelligence: Early experiments with GPT-4
Analyzing one of the most fascinating papers on Artificial Intelligence today
I’ve been wanting to write about the advancements of Artificial Intelligence (AI) since I started using chatGPT back in December. The developments and adv... |
3e41ae8f-1529-4b16-a2ca-e22f4c3a7315 | trentmkelly/LessWrong-43k | LessWrong | Luna Lovegood and the Chamber of Secrets - Part 7
They were in the hospital wing.
"Who are you?" Harry Potter said to Luna.
"You-Know-Who. Hiss hiss hiss," Luna said. She held fingers down from her mouth and swayed like a snake.
"This is Loony—" Hermione said, "Luna Lovegood. She created the BOY-WHO-LIVED GETS DRAC... |
fa1911b3-3494-4e1b-b610-ab0d7ce894d7 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | What must be the case that ChatGPT would have memorized “To be or not to be”? – Three kinds of conceptual objects for LLMs
*This is cross-posted from* [*my personal blog, New Savanna*](https://new-savanna.blogspot.com/2023/09/what-must-be-case-that-chatgpt-would.html)*.*
Is anyone working on this kind of issue? If s... |
b2a1b510-8a10-421e-8c5f-3f8f36212319 | trentmkelly/LessWrong-43k | LessWrong | Duncan Sabien On Writing
The following is an interview I had informally with Duncan Sabien back in early 2020 when I was trying to improve my writing. I thought Duncan gave some great, actionable tips, so I’m sharing it here for other people trying to improve their writing.
The transcript has been lightly edited for... |
9da74a07-f140-49e8-be9a-f7ef486593a2 | trentmkelly/LessWrong-43k | LessWrong | If you could push a button to eliminate one cognitive bias, which would you choose?
I realize this question is contrived, but I figure it might provoke some fun discussion, so here goes:
If you could push a button and have your brain modified to precisely remove a cognitive bias (and have no other unnecessary effects... |
8f22333b-3200-4c4d-9f7f-e5ad196fef19 | trentmkelly/LessWrong-43k | LessWrong | How much to optimize for the short-timelines scenario?
Some have argued that one should tend to act as if timelines are short since in that scenario it's possible to have more expected impact. But I haven't seen a thorough analysis of this argument.
Question: Is this argument valid and if yes how strong is it?
The b... |
2eeb0596-f114-467a-9d16-5cbf88e6f1a9 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Agentic Language Model Memes
Related: [What specific dangers arise when asking GPT-N to write an Alignment Forum post?](https://www.lesswrong.com/posts/Et2pWrj4nWfdNAawh/what-specific-dangers-arise-when-asking-gpt-n-to-write-an)
I've been thinking about the AI safety implications of ultra powerful language models. I... |
4dd3f431-3fb8-4964-9a70-4112492e53d2 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Optimization and the Singularity
Lest anyone get the wrong impression, I'm juggling multiple balls right now and can't give the latest Singularity debate as much attention as it deserves. But lest I annoy my esteemed co-blogger, here is a down payment on my views of the Singularity - needless to say, all this is comi... |
fa95a849-1098-4587-971a-a3f9aaeedf3d | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Graphical World Models, Counterfactuals, and Machine Learning Agents
*This is the second post in a sequence. For the introduction post, see
[here](https://www.alignmentforum.org/posts/7EnZgaepSBwaZXA5y/counterfactual-planning-in-agi-systems)*.
Graphical World Models
======================
A *world model* is a math... |
90d10d59-decf-4d1c-afaf-75c51f788654 | trentmkelly/LessWrong-43k | LessWrong | The Alignment Newsletter #4: 04/30/18
Highlights
Reptile: A Scalable Meta-Learning Algorithm (Alex Nichol et al): I somehow forgot to include this in past emails, so I'm including it now. Reptile is an algorithm for meta-learning, and in this paper is applied to few-shot classification, where given a few examples of d... |
32e4b93f-0b48-49df-923c-9e6b38a85aaf | trentmkelly/LessWrong-43k | LessWrong | Formula for a Shortage
Previous coverage (not required in any way): Formula For Dying Babies
Great Formal Workups of Much of This (Scott Lincicome): America’s Infant Formula Crisis and the ‘Resiliency’ Mirage. And another from Derek Thompson at The Atlantic. If you want a shorter summary, both of those are excellent.... |
1d69dffb-8490-4e17-aaff-0a79904ea8bf | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Reducing sycophancy and improving honesty via activation steering
*Produced as part of the* [*SERI ML Alignment Theory Scholars Program*](https://serimats.org/) *- Summer 2023 Cohort, under the mentorship of Evan Hubinger.*
I generate an activation steering vector using Anthropic's [sycophancy dataset](https://huggin... |
3ae0b78d-5b97-4d55-ad17-cd857599f390 | trentmkelly/LessWrong-43k | LessWrong | On the Importance of Systematic Biases in Science
From pg812-1020 of Chapter 8 “Sufficiency, Ancillarity, And All That” of Probability Theory: The Logic of Science by E.T. Jaynes:
> The classical example showing the error of this kind of reasoning is the fable about the height of the Emperor of China. Supposing that ... |
a469720d-db08-4b9c-a1a7-5ca0a2b4b345 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How Doomed are Large Organizations?
We now take the model from the previous post, and ask the questions over the next several posts. This first answer post asks these questions:
1. Are these dynamics the inevitable results of large organizations?
2. How can we forestall these dynamics within an organization?
3. To w... |
41715d39-e178-4a0b-a90e-68aa9183d8ee | trentmkelly/LessWrong-43k | LessWrong | Specialization
Multiple LW and EA voices (1, 2, 3) have written or spoken about the idea of becoming world-class generalists. More specifically, about combining "six specializations," or "the effort equivalent of two to three PhDs, plus enough work in the relevant fields in order to master them," or building a "cluste... |
67b10245-dfa4-4d9c-81ed-c60e19310ff8 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington, D.C.: Meta Meetup
Discussion article for the meetup : Washington, D.C.: Meta Meetup
WHEN: 11 January 2015 03:00:00PM (-0500)
WHERE: National Portrait Gallery
We will be meeting in the Kogod Courtyard of the National Portrait Gallery (8th and F Sts or 8th and G Sts NW, go straight past the infor... |
0e1e3a8b-af5f-4104-8613-aebc6a140bca | trentmkelly/LessWrong-43k | LessWrong | Imagine a World Where Governments Treated COVID-19 Properly
This essay was originally published on my website.
Over the last few weeks, I’ve reaffirmed my position that a statistics class should be REQUIRED for every person before they are released into the open world of adulthood.
I was first introduced to the mar... |
8ef78506-eeb6-4df6-865f-6e8e9571d1d1 | trentmkelly/LessWrong-43k | LessWrong | 2017: An Actual Plan to Actually Improve
[Epistemic status: mostly confident, but being this intentional is experimental]
This year, I'm focusing on two traits: resilience and conscientiousness. I think these (or the fact that I lack them) are my biggest barriers to success. Also: identifying them as goals for 2017... |
8094fdbc-ba4c-4083-a59f-dae9f41dd56d | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Decision Theories: A Semi-Formal Analysis, Part III
### Or: Formalizing Timeless Decision Theory
**Previously:**
0. [Decision Theories: A Less Wrong Primer](/lw/aq9/decision_theories_a_less_wrong_primer/)
1. [The Problem with Naive Decision Theory](/lw/axl/decision_theories_a_semiformal_analysis_part_i)
2. [Ca... |
041368ab-c5d1-4a0d-b091-8e01e59e2034 | trentmkelly/LessWrong-43k | LessWrong | Some conceptual alignment research projects
Some research outputs I’d love to see, focused on exploring, clarifying and formalizing important alignment concepts. I expect that most of these will be pretty time-consuming, but happy to discuss for people who want to try:
1. A paper which does for deceptive alignment ... |
1c07993e-37a5-4f68-81d4-37b35dccd404 | trentmkelly/LessWrong-43k | LessWrong | Applications for AI Safety Camp 2022 Now Open!
If you’ve read about alignment research and you want to start contributing, the new iteration of the AI Safety Camp is a great opportunity!
It’s a virtual camp from January to May 2022, where you collaborate with other applicants to work (1h / normal workday, 7h / weeken... |
eea03836-bfdd-48dd-a9e6-219cf23b772e | trentmkelly/LessWrong-43k | LessWrong | The Extraordinary Link Between Deep Neural Networks and the Nature of the Universe
"The answer is that the universe is governed by a tiny subset of all possible functions. In other words, when the laws of physics are written down mathematically, they can all be described by functions that have a remarkable set of simp... |
66bf1d7f-ddb1-4f38-80b3-43b3fac9fbb8 | trentmkelly/LessWrong-43k | LessWrong | The Ancient God Who Rules High School
|
53851fee-6cb0-4dbd-a201-69de45eb04f5 | trentmkelly/LessWrong-43k | LessWrong | We're in danger. I must tell the others...
... Oh, no! I've been shot!
— C3PO
A strange sort of paralysis can occur when risk-averse people (like me) decide that we're going to play it safe. We imagine the worst thing that could happen if we go ahead with our slightly risky plan, and this stops us from carrying it o... |
731ecfa0-94ea-4042-b6d5-40802c51a43b | trentmkelly/LessWrong-43k | LessWrong | Topics to discuss CEV
CEV is our current proposal for what ought to be done once you have AGI flourishing around. Many people have had bad feelings about this. When in Singularity Institute, I decided to write a text do discuss CEV, from what it is for, to how likely it is to achieve it's goals, and how much fine... |
598d17aa-ed27-4a19-be31-87c9b2cb251d | trentmkelly/LessWrong-43k | LessWrong | The Axiological Treadmill
The obvious reason that Moloch is the enemy is that it destroys everything we value in the name of competition and survival. But this is missing the bigger picture. We value what we value because, in our ancestral environment, those tended to be the things that helped us with competition and ... |
16180122-e22b-4903-b4fd-7ec156d3955d | trentmkelly/LessWrong-43k | LessWrong | LINK "Politics is the Mindkiller" a la Cracked.com
A Cracked.com columnist gives a witty explanation of some of the same problems EY points out with political teams.
http://www.cracked.com/blog/5-ways-internet-convinced-me-not-to-vote/ |
5f6e6cad-1509-4169-8f3f-dcda0e9c2eb8 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Random Thoughts on Predict-O-Matic
I'm going to be a bit more explicit about some ideas that appeared in [The Parable of Predict-O-Matic](https://www.lesswrong.com/posts/SwcyMEgLyd4C3Dern/the-parable-of-predict-o-matic). (If you don't want spoilers, read it first. Probably you should read it first anyway.)
[Note: whi... |
dbcd4fb7-7c6f-4738-a11e-f5e16c211f96 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | The Inside View #1—Does the world really need another podcast?
so this one is very special because i
will be the one
being interviewed and this will be about
the podcast the interviewee
will be thiago one of my best friends
and introduced me to rationality last
wrong the whole concept
of an intelligent explosion while... |
18127025-61fa-4022-ad4e-220288ffff4b | trentmkelly/LessWrong-43k | LessWrong | [Review] Move First, Think Later: Sense and Nonsense in Improving Your Chess
The author is Dutch International Master Willy Hendriks (IM is the rank below Grandmaster; there are ~4000 IMs and ~2000 GMs). He takes aim at chess instructors like IM Jeremy Silman, whose popular books The Amateur’s Mind and How to Reassess... |
33facfb8-0932-4a5b-9630-2600a1976738 | trentmkelly/LessWrong-43k | LessWrong | Disincentivizing deception in mesa optimizers with Model Tampering
Disclaimer: I don't know if this is something worth exploring or a stupid idea and I thought the best way to get feedback was to describe it here.
Suppose a training process could produce a mesa optimizer. How can we proceed to ensure that the optimiz... |
75fc6075-08b1-4b9d-ac6d-11224791be36 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "In the Ultimatum Game, the first player chooses how to split $10 between themselves and the second player, and the second player decides whether to accept the split or reject it—in the latter case, both parties get nothing. So far as conventional causal decision theory goes (two-box on Newcomb's Problem, defect in Pri... |
40f269d1-e17e-4fb3-be0c-3912e38c7392 | trentmkelly/LessWrong-43k | LessWrong | Making the case for average-case AI Control
Introduction
In this short article, I will share my thoughts on why I believe greater efforts should be directed at solving the average case of AI Control, ultimately designing systems that resist adaptation to a certain harmful task. First, I will explain what I mean by AI ... |
53cf3d0c-2532-430e-abb0-d373e9044dd6 | trentmkelly/LessWrong-43k | LessWrong | Inspired by: Failures in Kindness
silentbob's post "Failures in Kindness" is excellent. I love the idea that sometimes, when we exaimine a situation in depth, the most "kind" course of action can be highly conterintuitive. A few other examples I'd like to offer:
Appreciative Kindness
Imagine you meet a friend-of-a-... |
efbdd7dd-7543-46d4-86dd-b87dcc52fb37 | trentmkelly/LessWrong-43k | LessWrong | EPUBs of MIRI Blog Archives and selected LW Sequences
Here is a repository of EPUBs that I've created, consisting of rationalist or alignment content I want to read on my Kindle, such as selections of the MIRI (Blog) Archives and of multiple LessWrong Sequences that I am reading or intend to read.
This post is an ann... |
0866618c-0e21-48ae-8330-1f00e8da236b | StampyAI/alignment-research-dataset/lesswrong | LessWrong | My Most Likely Reason to Die Young is AI X-Risk
Introduction
------------
One of my pet peeves is people calling AI Safety a longtermist cause area. It is [entirely](https://forum.effectivealtruism.org/posts/KDjEogAqWNTdddF9g/long-termism-vs-existential-risk) [reasonable](https://forum.effectivealtruism.org/posts/rFp... |
0543c6e8-cae3-448f-b11c-cbe5c4abe2cb | trentmkelly/LessWrong-43k | LessWrong | Are there any Lesswrongers in the Waterloo, Ontario area?
I will soon be attending the University of Waterloo and I was wondering if there is interest in a meetup. Please reply if you are in the area. |
e1010123-f6d5-4dee-9fc4-8741663f5381 | trentmkelly/LessWrong-43k | LessWrong | Markets are Anti-Inductive
I suspect there's a Pons Asinorum of probability between the bettor who thinks that you make money on horse races by betting on the horse you think will win, and the bettor who realizes that you can only make money on horse races if you find horses whose odds seem poorly calibrated relative ... |
bc388814-51a5-4ad5-b8de-bfcd4508d291 | trentmkelly/LessWrong-43k | LessWrong | Occam's Razor: In need of sharpening?
In the first half of the 14th century, the Franciscan friar and logician, William of Occam proposed a heuristic for deciding between alternative explanations of physical observables. As William put it: "Entities should not be multiplied without necessity". Or, as Einstein reformul... |
af1e7a90-8430-4a92-9ea7-e123d4f87941 | trentmkelly/LessWrong-43k | LessWrong | Game Review: This Merchant Life
In a previous life, I used to be a professional game player. I still am hard at work on a new tradable blockchain card game, Emergents, now in closed (but soon to be more open) beta, and I still love playing games.
This weekend I had the chance to take a break and play This Merchant Li... |
23ead987-228e-4170-b751-4e7abb77a358 | trentmkelly/LessWrong-43k | LessWrong | The Limits of Intelligence and Me: Domain Expertise
Related to: Trusting Expert Consensus
In the sequences, Eliezer tells the story of how in childhood he fell into an affective death spiral around intelligence. In his story, his mistakes were failing to understand until he was much older that intelligence does not g... |
522deb2d-1056-4130-be47-580d8cdf9fc2 | trentmkelly/LessWrong-43k | LessWrong | One-shot steering vectors cause emergent misalignment, too
TL;DR: If we optimize a steering vector to induce a language model to output a single piece of harmful code on a single training example, then applying this vector to unrelated open-ended questions increases the probability that the model yields harmful output... |
20302d83-5735-4063-ae31-2cbd5810bb03 | trentmkelly/LessWrong-43k | LessWrong | Empty Labels
Consider (yet again) the Aristotelian idea of categories. Let's say that there's some object with properties A, B, C, D, and E, or at least it looks E-ish.
> Fred: "You mean that thing over there is blue, round, fuzzy, and—"
> Me: "In Aristotelian logic, it's not supposed to make a difference what the ... |
8288e71f-264d-476e-ba5b-51573325da00 | trentmkelly/LessWrong-43k | LessWrong | Where to begin in ML/AI?
I am a senior CS/Math student at UMD, college park, and I "specialized" my major in machine learning. I have taken courses on Deep Learning and NLP, reinforcement learning, however I still feel like a novice in all of these categories, my proudest achievement being this half-ripped off sudoku ... |
52f35303-afb1-4ce0-9ff0-282989968945 | trentmkelly/LessWrong-43k | LessWrong | Category Theory Without The Baggage
If you are an algebraic abstractologist, this post is probably not for you. Further meta-commentary can be found in the “meta” section, at the bottom of the post.
So you’ve heard of this thing called “category theory”. Maybe you’ve met some smart people who say that’s it’s really u... |
9de311c0-f1be-42a7-abec-6f7680f242ff | StampyAI/alignment-research-dataset/arxiv | Arxiv | Why Build an Assistant in Minecraft?
1 Introduction
---------------
In the last decade, we have seen a qualitative jump in the performance of machine learning (ML) methods directed at narrow, well-defined tasks. For example, there has been marked progress in object recognition [[57](#bib.bib57)], game-playing [[73]... |
88dcd177-755f-40ff-ac0f-0724cf82d7dc | StampyAI/alignment-research-dataset/lesswrong | LessWrong | AGI Timelines in Governance: Different Strategies for Different Timeframes
Summarization Table
===================
| Timelines | **Pre-2030** | **Post-2030** |
| --- | --- | --- |
| **Expectations** | AGI will be built by an organization that’s already trying to build it (85%) | Some governments will be in the race... |
0da83c9a-ddfc-4508-bf41-161a3be57380 | trentmkelly/LessWrong-43k | LessWrong | Towards a Unified Interpretability of Artificial and Biological Neural Networks
Neuroscience and mechanistic interpretability share a common goal: understanding neural networks, either biological or artificial. This is reflected in the convergent evolution of these domains – from interpreting single neurons to abstrac... |
0f39b062-26f8-4bb6-95ed-22b1dd5d5fd8 | trentmkelly/LessWrong-43k | LessWrong | The Mirror Chamber: A short story exploring the anthropic measure function and why it can matter
This story warrants and motivates the study of the hard problem of consciousness, presenting a plausible future scenario in which a notion of quantifiable anthropic measure as a property of an arrangement of matter demands... |
f4e68fc3-b418-4cea-95b8-405225aa783a | trentmkelly/LessWrong-43k | LessWrong | Car Thoughts
I've been putting this post off a bit it, because a post about family logistics is a bit of a silly way for people to learn that we're expecting a third child, and I kept thinking I would write something else first, but I didn't. So: with another child coming, in June, I'm thinking about, among many other... |
d639e08b-2712-48a4-8ea3-9fb1b734034e | trentmkelly/LessWrong-43k | LessWrong | The Tallest Pygmy Effect
Status: I thought this was a common economics term, but when I google it I get either unrelated or references using it the way I expect but not defining it. It’s a really useful term, so I’m going to attempt to make it a thing.
“Tallest Pygmy Effect” is when you benefit not from absolute skil... |
169b3137-b2d8-4e34-b79f-c13c6f68a6f0 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Paris Meetup: Sunday, May 26.
Discussion article for the meetup : Paris Meetup: Sunday, May 26.
WHEN: 26 May 2013 02:00:00PM (+0200)
WHERE: Café des Arts et Métiers, 51 Rue Turbigo, Paris
The next Paris Meetup will be Sunday, May 26, at the Café des Arts et Métiers opposite the Museum. We have several gue... |
07691ab8-a584-402e-a4dc-b0e0cbe271e2 | trentmkelly/LessWrong-43k | LessWrong | Heat vs. Motion
After yesterday's post, it occurred to me that there's a much simpler example of reductionism jumping a gap of apparent-difference-in-kind: the reduction of heat to motion.
Today, the equivalence of heat and motion may seem too obvious in hindsight—everyone says that "heat is motion", therefore, it ca... |
1f8d8b4b-7fec-4a5b-88d8-02688af84962 | trentmkelly/LessWrong-43k | LessWrong | [Site Meta] Feature Update: More Tags! (Experimental)
Edit: We're specifically seeking feedback right now on the tag relevance voting system, please chime in if you can. Thanks!
A month ago, the LessWrong team launched our new tagging system...with a whopping whole entire single tag. We rushed it out quickly to give... |
f44faa37-eb0b-43cb-a870-e88e6442d33c | trentmkelly/LessWrong-43k | LessWrong | Resource-Limited Reflective Oracles
Reflective oracles accurately answer questions about what arbitrary halting probabilistic oracle machines output. It is possible to make a variant of a reflective oracle that accurately answers questions about what sufficiently short-running Turing machines with access to the same o... |
b422f8ff-f092-44d4-8a33-439ce5f056d3 | trentmkelly/LessWrong-43k | LessWrong | Physicists To Test If Universe Is A Computer Simulation (link)
If it is... I hope they do not crash the system with the test.
http://www.huffingtonpost.co.uk/2012/12/12/physicists-universe-simulation-test-university-of-washington-matrix_n_2282745.html
Be sure to check out the actual reseach papers linked in the art... |
dc35b695-028c-4035-8032-97836bc7ef48 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Universal and Transferable Adversarial Attacks on Aligned Language Models [paper link]
> Because "out-of-the-box" large language models are capable of generating a great deal of objectionable content, recent work has focused on aligning these models in an attempt to prevent undesirable generation. While there has been... |
63ca4a74-abad-47f2-81d4-f15e8fa8adde | StampyAI/alignment-research-dataset/arxiv | Arxiv | Imitation Learning by Estimating Expertise of Demonstrators.
1 Introduction
---------------
Reinforcement learning provides a powerful and general framework for tasks such as autonomous vehicles, assistive robots, or conversation agents, by optimizing behavior with respect to user-specified reward functions. Howeve... |
edd57fb8-a1a8-4529-9fb1-78c2cc7ae1a9 | trentmkelly/LessWrong-43k | LessWrong | Munk AI debate: confusions and possible cruxes
There was a debate on the statement “AI research and development poses an existential threat” (“x-risk” for short), with Max Tegmark and Yoshua Bengio arguing in favor, and Yann LeCun and Melanie Mitchell arguing against. The YouTube link is here, and a previous discussio... |
4685e390-0769-4930-a646-52bef4c33232 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | Aren't there easy solutions to AI alignment?
<iframe src="https://www.youtube.com/embed/i8r_yShOixM" title="AI? Just Sandbox it... - Computerphile" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen></iframe>
If you've been learn... |
19c95186-71b5-435a-a58d-13ac056bcde6 | trentmkelly/LessWrong-43k | LessWrong | Race to the Top: Benchmarks for AI Safety
This is an executive summary of a post from my personal blog, also cross-posted from the EA Forum. Read the full texts here.
Summary
Benchmarks support the empirical, quantitative evaluation of progress in AI research. Although benchmarks are ubiquitous in most subfields of... |
e1f7f4c2-7649-447a-b204-96f180f50e19 | trentmkelly/LessWrong-43k | LessWrong | Rationality Cardinality
Rationality Cardinality is a card game which takes memes and concepts from the rationality/Less Wrong sphere, and mixes them with jokes to make a game. After nearly two years of card-creation, playtesting and development, today, I'm taking the "beta" label off the web-based version of Rationali... |
030dc03e-18a9-4948-827f-c9da31008b1c | trentmkelly/LessWrong-43k | LessWrong | Confirmation Bias Presentation
On Monday I need to give a presentation to a group of 6-15 finance professionals on Confirmation Bias. I intend to use the 2-4-6 task to demonstrate it.
Do people have any advise on how to make this work well? Do people tend to fall for it? Does it help them understand afterwards?
... |
e37e7308-a33f-4816-8af3-cc1898b14bc3 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Polysemanticity and Capacity in Neural Networks
1 Introduction
---------------
Individual neurons in neural networks often represent multiple unrelated features in the input [olah2017feature](#bib.bib4) ; [olah2020zoom](#bib.bib3) .
This phenomenon is known as polysemanticity, and makes it more difficult to interpr... |
3e96c22b-440b-4adf-bb78-2ca636357b7d | trentmkelly/LessWrong-43k | LessWrong | making notes - an instrumental rationality process.
The value of having notes. Why do I make notes.
Story time!
At one point in my life I had a memory crash. Which is to say once upon a time I could remember a whole lot more than I was presently remembering. I recall thinking, "what did I have for breakfast last ... |
acb78f48-e0dc-47d5-b975-aaf912858652 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | BCIs and the ecosystem of modular minds
*Crossposted from* [*my personal blog*](https://www.beren.io/2023-04-23-Composable-latent-spaces-BCIs-modular-minds/)*.*
**Epistemic status**: *Much more speculative than previous posts but points towards an aspect of the future that is becoming clearer which I think is underap... |
901a8a07-88c0-4ba4-81a2-4e3a2823c9bc | trentmkelly/LessWrong-43k | LessWrong | The Argument from Philosophical Difficulty
(I'm reposting this comment as a top-level post, for ease of future reference. The context here is a discussion about the different lines of arguments for the importance of AI safety.)
Here's another argument that I've been pushing since the early days (apparently not very s... |
4e41e36f-1615-4e5d-81e6-4fb65ff5be4b | trentmkelly/LessWrong-43k | LessWrong | Summary: "How to Do Research" by OSP's Red
This post is part of my "research" series.
Red. How to Do Research. Overly Sarcastic Productions, 2019. https://www.youtube.com/watch?v=sigJwoeU6OI.
Why
Overly Sarcastic Productions are two friends making short (around 14 minutes) video presentations about history, mytholo... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.