id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
621f795e-947e-4af1-944e-d5d9c63c8b8d | trentmkelly/LessWrong-43k | LessWrong | Going Out With Dignity
DF was born with a time bomb in his genome, a deadly curse more horrifying than most. The name of this curse was Fatal familial insomnia (FFI).
Wikipedia describes the usual progression of this hell:
> The disease has four stages:[8]
>
> 1. Characterized by worsening insomnia, resulting in p... |
41f87889-2ae3-425d-92cc-6d2571c5f0f9 | trentmkelly/LessWrong-43k | LessWrong | Pivotal outcomes and pivotal processes
tl;dr: If you think humanity is on a dangerous path, and needs to "pivot" toward a different future in order to achieve safety, consider how such a pivot could be achieved by multiple acts across multiple persons and institutions, rather than a single act. Engaging more actors i... |
a920727b-3505-49f6-b6cb-0789f0a934f5 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Power-Seeking = Minimising free energy
Epistemic Status: *Pretty certain there are better ways to describe this but I believe that the underlying intuition holds and that it might be an exciting formalisation of power.*
*Thank you to Viktor Rehnberg and Roman Levantov for some great discussions leading up to this and... |
10e3aab9-3ccd-4235-a166-2d22309a99ca | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Who is liable for AI?
[Edit -- Changed to title as suggested by Raemon. So I've also made some edits to the post itself.]
Where does liability fall in the case of AI doing things?
I think there is a rather close parallel with autonomous cars/vehicles. Probably 4 or 5 years back that was a question being raise. I ass... |
f57785bc-506c-4962-9bf4-9874fb84dda3 | trentmkelly/LessWrong-43k | LessWrong | Inequality Penalty: Morality in Many Worlds
This is a summary of Sean Carroll's musings on when it would matter which model of apparent collapse is more accurate.
Executive Summary: If you care about equality, or even are risk-averse, then your decisions depend on whether you subscribe to the Many Worlds model.
Dis... |
1a6113e4-c3cf-496a-8711-0227f7366f9a | trentmkelly/LessWrong-43k | LessWrong | What was my mistake evaluating risk in this situation?
At the start of the pandemic, I heard that Covid was likely to be a problem. I didn't pay much attention, because I remembered the Swine Flu "pandemic", and Covid was getting less attention than Swine Flu did at the time. I also knew that the media has a strong in... |
26674062-b618-4cae-9cb4-e009073087b6 | trentmkelly/LessWrong-43k | LessWrong | Comparing Options For Safer Events
Yesterday I wrote about how indoor contra dancing, with one set of precautions, seems to be safe enough that I would be ok resuming organizing dances. How safe it is, however, depends a lot on which precautions organizers take. I looked over a range of potential options, many of them... |
eb6e3de8-ed19-430e-a559-c68491def371 | trentmkelly/LessWrong-43k | LessWrong | 2005 short story loosely about P-zombies
I thought some of you might enjoy this 2005 story published in Asimov's:
Second Person, Present Tense by Daryl Gregory (plus some of the author's notes on the story)
The story involves a a drug that (for a period of time) turns someone into a P-zombie.
It's currently availab... |
08359d0e-e3ef-4b51-a64b-20c2c0b048b4 | trentmkelly/LessWrong-43k | LessWrong | What are good election betting opportunities?
I have seen a bunch of people saying that there are very good opportunities to bet on the upcoming election. This is meant to be a practical advice thread to share concrete opportunities. |
a3e0e462-1768-4afd-9055-373e83eb0958 | trentmkelly/LessWrong-43k | LessWrong | Russian Cynicism
I.
"My dear Kostya, you've finally returned to me!"
Kitty comes trotting down the steps of the train station, rushing into Levin's waiting arms.
"It's been but four days, my dear."
Levin's wraps his arms around her slim waist with a passion that belies his nonchalance. Gently resting in each others... |
127970be-a437-47fb-b5c8-9dc9f3bf7db9 | trentmkelly/LessWrong-43k | LessWrong | confirmation bias, thought experiment
Why do people end up with differing conclusions, given the same data?
Model
The information we get from others can not always be 100% relied upon. Some of the people telling you stuff are liars, some are stupid, and some are incorrectly or insufficiently informed. Even in t... |
8a2721db-ad20-431f-9585-c60a3514a757 | trentmkelly/LessWrong-43k | LessWrong | Apollo Neuro Results
Introduction
Two months ago I recommended the Apollo Neuro for sleep/anxiety/emotional regulation. A number of people purchased it based on my recommendation- at least 25, according to my referral bonuses. Last week I asked people to fill out a form on their experience.
Take-home messages:
* If... |
8d4e3402-71e3-47a7-b565-87bab9d4c2f8 | trentmkelly/LessWrong-43k | LessWrong | Meta: IAFF vs LessWrong
It seems that currently many or even most researchers started posting on LessWrong.com instead of IAFF. I am troubled by this for two reasons.
1. I hoped that eventually IAFF will draw in all of the AI alignment research community rather than only the MIRI-sphere. Moving to LessWrong seems li... |
76da7cc5-7b29-4321-97fd-ae8f1ce26752 | trentmkelly/LessWrong-43k | LessWrong | Conspiracy Theorists Aren't Ignorant. They're Bad At Epistemology.
Cross-post of my blog article on the topic.
I probably know less about science than most people who think the earth is flat do.
Okay, that’s not quite true. I have knowledge of lots of claims of science—that the big bang happened, that evolution i... |
91837c39-6a8d-40de-a9f4-9a77167a2bec | trentmkelly/LessWrong-43k | LessWrong | What’s in the box?! – Towards interpretability by distinguishing niches of value within neural networks.
Abstract
Mathematical models can describe neural network architectures and training environments, however the learned representations that emerge have remained difficult to model. Here we build a new theoretical mo... |
f7a45245-648f-4ebe-9808-da6cfbb0bdc0 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | What would you expect a massive multimodal online federated learner to be capable of?
One intuitive picture I have for what very rapid ML capability gains might look like is a massive multimodal deep learning model that uses some form of online federated learning to continually learn from many devices simultaneously, ... |
b1e65836-682f-4a71-878f-21120713f35d | trentmkelly/LessWrong-43k | LessWrong | Is this the beginning of the end for LLMS [as the royal road to AGI, whatever that is]?
It’s hard to tell, but it sure is...shall we say...interesting.
Back in the summer of 2020 when GPT-3 was unveiled I wrote a working paper, GPT-3: Waterloo or Rubicon? Here be Dragons. My objective was to convince myself that the ... |
c94b8748-f10d-4c57-8a7d-c89d206c447a | trentmkelly/LessWrong-43k | LessWrong | AI Governance across Slow/Fast Takeoff and Easy/Hard Alignment spectra
It has been suggested that in a rapid enough takeoff scenario, governance would not be useful, because the transition to superintelligence would be too rapid for human actors - whether governments, corporations, or individuals - to respond to. This... |
3e5c26b7-03f7-4901-9f24-027056441342 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | [Link post] Promising Paths to Alignment - Connor Leahy | Talk
**Date of recording:**Tuesday May 10, 2022
**Description**
[AI Safety Support](https://www.aisafetysupport.org/home) recently hosted a closed, introductory talk by Connor Leahy on Promising Paths to Alignment.
The talk covers: (i) why Alignment is suc... |
24101a6f-87f3-4248-afa6-560644fa9e02 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Australia Online Hangout
Discussion article for the meetup : Australia Online Hangout
WHEN: 07 October 2015 06:00:00PM (+1000)
WHERE: Canberra, ACT, Australia
Come one, come all, to our latest Less Wrong Online Hangout!
We talk about anything and everything Less-Wrong-related: things you've learned recent... |
a73b919f-bebd-4ad4-9289-d826c1b43f04 | trentmkelly/LessWrong-43k | LessWrong | Attention Lurkers: Please say hi
Some research says that lurkers make up over 90% of online groups. I suspect that Less Wrong has an even higher percentage of lurkers than other online communities.
Please post a comment in this thread saying "Hi." You can say more if you want, but just posting "Hi" is good for a guar... |
62f8abd5-53c0-48ab-a133-df87709a7264 | trentmkelly/LessWrong-43k | LessWrong | Take 1: We're not going to reverse-engineer the AI.
As a writing exercise, I'm writing an AI Alignment Hot Take Advent Calendar - one new hot take, written every day for 25 days. Or until I run out of hot takes, which seems likely.
Any approach to building safe transformative AI, or even just auditing possibly-safe T... |
810d2960-d9e8-4993-91a0-2878275f4a6d | StampyAI/alignment-research-dataset/special_docs | Other | Voluntary safety commitments provide an escape from over-regulation in AI development.
Voluntary safety commitments provide an escape from over-regulation in AI development
=====================================================================================
Abstract
--------
With the introduction of Artificial Inte... |
398a7b74-69ec-40b0-b57f-61d0df1e3d72 | trentmkelly/LessWrong-43k | LessWrong | Fake Norms, or "Truth" vs. Truth
Followup to: Applause Lights
When you say the word "truth", people know that "truth" is a good thing, and that they're supposed to applaud. So it might seem like there is a social norm in favor of "truth". But when it comes to some particular truth, like whether God exists, or how ... |
f3d65c0a-8d65-4ac7-8ea5-5621b0d0b4df | trentmkelly/LessWrong-43k | LessWrong | Predictive Reasoning Systems
Meta: This is meant to be a succinct and high-level overview. I expect to go more into detail on parts of the system in future posts. This work builds on this previous post.
Predictions can be a powerful tool but are typically difficult to use in isolation. Being sure to predict the right... |
304fd999-893e-416c-ae41-e97ea3d9d2e7 | trentmkelly/LessWrong-43k | LessWrong | Have you tried hiIQpro.com's cognitive training or coaching?
Seems extremely flimy-flammy, but they do offer a money back guarantee if you don't jump 10 - 20 points on a standardized test. |
12592511-d30a-4265-8f2a-20f7b256a529 | StampyAI/alignment-research-dataset/arbital | Arbital | Featured math content
Here are a few of the best pages on Arbital, selected by our editors.
### [https://arbital.com/p/1zq](https://arbital.com/p/1zq)
A customizable multi-page exploration of one of the core concepts behind good reasoning.
### [https://arbital.com/p/3nd](https://arbital.com/p/3nd)
An extended expl... |
bdc49566-9f9b-40b1-b286-b1840509f35c | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Catastrophic Risks from AI #6: Discussion and FAQ
*This is the final post in a sequence of posts giving an* [*overview of catastrophic AI risks*](https://arxiv.org/abs/2306.12001)*.*
6 Discussion of Connections Between Risks
=========================================
So far, we have considered four sources of AI risk... |
23451116-1006-4263-9105-85c9adb67cbf | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Where Experience Confuses Physicists
Today's post, Where Experience Confuses Physicists was originally published on 26 April 2008. A summary (taken from the LW wiki):
> It then turns out that the entire planet of Ebbore is splitting along a fourth-dimensional thickness, duplicating all the people withi... |
1809e393-a420-4077-b236-3cc60a39da71 | trentmkelly/LessWrong-43k | LessWrong | So You've Learned To Teleport by Tom Scott
Tom Scott video on creative uses of teleportation (if you were a superhero). It is from August 20, 2015 and there have been no similar videos despite it claiming to be Episode 8 at t=16s.
It reminded me of HPMOR when I saw it. |
6f718404-846d-40a8-bbaf-2a5151b04524 | trentmkelly/LessWrong-43k | LessWrong | Is there something like a "meta-encyclopedia"?
With Wikipedia showing increasing signs of ideological uniformity, I thought I should try to diversify my consumption of reference material. A potentially very useful tool for this purpose would be a "meta-encyclopedia", i.e. a place where users can find a list of all exi... |
5053b7be-6743-48e9-805c-4cc08411fbed | trentmkelly/LessWrong-43k | LessWrong | Meetup : Tel Aviv: Rationality in the Immune System
Discussion article for the meetup : Tel Aviv: Rationality in the Immune System
WHEN: 24 February 2015 07:00:00PM (+0200)
WHERE: Google, Electra Tower, 98 Yigal Alon Street, Tel Aviv, Israel
There was a change of speaker after the first announcement.
On Tuesday, F... |
8899b7d0-5fd5-4a8e-a175-a10b6e1e2aa6 | trentmkelly/LessWrong-43k | LessWrong | Interpreting a dimensionality reduction of a collection of matrices as two positive semidefinite block diagonal matrices
Here are some empirical observations that I have made on August 14, 2023 to August 19, 2023 that are characteristics of the interpretability of my own matrix dimensionality reduction algorithm. Thes... |
64204d84-edeb-4948-9479-03daefd47dc8 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What are polysemantic neurons?
Neural networks often contain “polysemantic neurons” that respond to multiple unrelated inputs, unlike “monosemantic” neurons which only fire in response to a single human-understandable feature or concept.
 does NVIDIA make per month?
* Which of these use ... |
45d6d5d1-d195-4700-a0e7-b6100263c45f | trentmkelly/LessWrong-43k | LessWrong | Information-generating research projects
The research projects that I'm interested in here are projects that:
* Involve aggregating and analyzing information
* Don't require experiments (e.g. not biomedical research)
* Can be done by people who aren't subject matter experts.
* Lead to publicly accessible informat... |
368d19c4-4f5e-4804-8cfa-e6fc6b5f9071 | trentmkelly/LessWrong-43k | LessWrong | GPT-3: a disappointing paper
[Note: I wrote this post in late May 2020, immediately after the GPT-3 paper was released.]
This post is a compilation of two posts I recently made on tumblr.
For context: I have been an enthusiastic user of GPT-2, and have written a lot about it and transformer models more generally. My... |
6acd23b3-3f35-4ecc-a9ea-9409fb637874 | trentmkelly/LessWrong-43k | LessWrong | What's a good book for a technically-minded 11-year old?
"I, Robot" comes to mind. What else? |
10036c81-4e6c-4971-af70-5b10898a92ab | trentmkelly/LessWrong-43k | LessWrong | Ambiguity Detection
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
Note: the most up-to-date version of this proposal can be found here.
Introduction
> If I present you with five examples of burritos, I don’t want you to pursue the simplest way of classifying burritos versus non-burritos. I want you to come up with a way of classif... |
f35bb67d-c5e0-461a-9941-e39702cdda57 | trentmkelly/LessWrong-43k | LessWrong | In the very near future the internet will answer all of our questions and that makes me sad
Today if I type into Google’s search bar “how to move a couch into a moving plane using only a pair of avocados” the first result is not particularly on-point:
> 14 Tips for Moving Furniture | The Family Handyman
> https://ww... |
1b280dd3-dbb5-4fc4-b5e7-95f271c471ea | trentmkelly/LessWrong-43k | LessWrong | Meetup : Weekly Chicago Meetups Resume 5/26
Discussion article for the meetup : Weekly Chicago Meetups Resume 5/26
WHEN: 26 May 2012 01:00:00PM (-0500)
WHERE: 360 N Michigan Ave, Chicago IL
Because of all of the craziness happening in downtown Chicago, with the NATO meeting and associated protests, the 5/19 Chicago... |
d704b074-960a-40f0-9076-d8a4d8fc201a | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | How long does it take to undersrand AI X-Risk from scratch so that I have a confident, clear mental model of it from first principles?
In particular how many hours of study would this take, having minimal background in computer science and AI, and I’m also curious what courses / books will get me to a first principles... |
c40a3b52-a8a7-4dbb-ae49-f6174689fd92 | trentmkelly/LessWrong-43k | LessWrong | How many GPUs are markets expecting?
Even if we get ultra-capable frontier models, we’ll need lots of GPUs to run them at scale. Currently, our installed GPU stock isn’t enough to automate the white-collar workforce or cause extinction. We need more!
Tyler Cowen thinks we should predict AI progress using asset pri... |
6cfeb3a7-1a44-478c-ae56-e18a5d111342 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Existential Angst Factory
Today's post, Existential Angst Factory was originally published on 19 July 2008. A summary (taken from the LW wiki):
> As a general rule, if you find yourself suffering from existential angst, check and see if you're not just feeling unhappy because of something else going on... |
5bbbada2-25b3-455d-8195-f360c3be5dba | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Are short timelines actually bad?
Sam Altman recently[posted](https://twitter.com/sama/status/1621621725791404032) the following:
I have seen very little serious discussion about whether s... |
8203d680-2bec-4428-929f-17583de1bde5 | trentmkelly/LessWrong-43k | LessWrong | Sugar and motivation
I reread Are Wireheads Happy?[1], which is about the difference between liking and wanting.
I've noticed that too much simple carbs (not terribly much-- three twenty ounce Cokes on consecutive days will do it) severely damages my desire to do much of anything, and the effects take approximately ... |
d8e8a3f7-92d6-4470-a5cd-d6052a619723 | trentmkelly/LessWrong-43k | LessWrong | Instrumental rationality for overcoming disability and lifestyle failure (a specific case)
I read things like "Rationality should win!", and I feel validated. That's basically what I've believed as far back as I can remember (and though my memory isn't as reliable as I'd like, I do think it is pretty decent, based on ... |
aa03a7d3-108c-4d99-a5c7-dbead474708d | trentmkelly/LessWrong-43k | LessWrong | Meetup : Moscow: applied rationality and web resources
Discussion article for the meetup : Moscow: applied rationality and web resources
WHEN: 29 September 2012 04:00:00PM (+0400)
WHERE: Rossiya, Moscow, ulitsa Ostozhenka 14
We will meet at “Subway” restaurant, entrance from Lopukhinskiy pereulok. Look for a table ... |
90930ef0-cba4-44fc-b3ff-cc3db041f551 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Humans are stunningly rational and stunningly irrational
Are humans mainly rational or mainly irrational? This is an old debate, with some political implications.
But the question is ill posed, because it elides the important detail: compared with what?
Humans are stunningly irrational
============================... |
60227168-189c-4397-af14-03b43b08e754 | trentmkelly/LessWrong-43k | LessWrong | Conserved Quantities (Stat Mech Part 2)
As before, we will consider particles moving in boxes in an abstract and semi-formal way.
Imagine we have two types of particle, red and blue, in a box. Imagine they can change colour freely, and as before let's forget as much as possible. Our knowledge over particle states now... |
e1179991-bf55-4bb2-bee6-39f1a070b80d | StampyAI/alignment-research-dataset/arbital | Arbital | Mind design space is wide
Imagine an enormous space of possible mind designs, within which all humans who've ever lived are a single tiny dot. We all have the same cerebral cortex, cerebellum, thalamus, etcetera. There's an instinct to imagine "Artificial Intelligences" as a kind of weird tribe that lives across the... |
ad439b3d-a539-40b1-9fd8-06e77c6e0686 | trentmkelly/LessWrong-43k | LessWrong | [Fiction] Lena (MMAcevedo)
Wiki article about the first brain image of a human upload (2031). Subpar in some respects, but initially a compliant worker and free to copy thanks to court cases ruling that the biological original did not have a legal right to restrict its use. |
20259e0e-f288-42a9-9ce3-82d32b719d60 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Barring a major collapse of human civilization (due to nuclear war, asteroid impact, etc.), many experts expect the intelligence explosion Singularity to occur within 50-200 years.
That fact means that many philosophical problems, about which philosophers have argued for millennia, are suddenly very urgent.
Those conc... |
0ed75f77-e3aa-47c3-abf8-e6a69b2e2c3e | trentmkelly/LessWrong-43k | LessWrong | Brain Preservation Foundation ask me anything on Reddit 7:00PM EST Thursday Nov 21
AMA is here : http://www.reddit.com/r/IAmA/comments/1r6exr/i_am_kenneth_hayworth_a_phd_neuroscientist_and/
The Brain Preservation Foundation's founder Ken Hayworth is going to be available on Reddit (http://www.reddit.com/r/IAmA) th... |
26ee6b67-fa61-4f1a-8c29-291661d66766 | trentmkelly/LessWrong-43k | LessWrong | Shock Level 5: Big Worlds and Modal Realism
In recent times, science and philosophy have uncovered evidence that there is something very seriously weird about the universe and our place in it. We used to think that there was one planet earth, inside a universe that is very large (at least 10^26 meters in diameter) but... |
e798d8a5-6a48-4c64-919e-ac97220006a3 | trentmkelly/LessWrong-43k | LessWrong | Quick Thoughts on A.I. Governance
My friend Devin Kalish recently convinced me, at least to some extent, to focus on A.I. governance in the short-to-medium term, more than technical A.I. safety.
The argument that persuaded me was this:
> A key point of stress among AI doomsayers like Yudkowsky is that we, first of a... |
bf8fdf21-d0d2-4b56-80de-5bc99198737d | trentmkelly/LessWrong-43k | LessWrong | Open Thread, May 1-15, 2012
If it's worth saying, but not worth its own post (even in Discussion), then it goes here. |
9017a478-ef55-46f3-b7aa-1c6c4a64fb42 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | How humans co-determine the development of intelligent technology (Serge Thill)
yes
page is yours okay thank you
right now i have to click away the fact
that you are started recording
and i have to click this away and now i
can actually talk so i'm guessing you
can all see my screen right that's
working out okay it's ... |
ac3cb656-e17a-466d-ab98-ce2a135e5f2b | trentmkelly/LessWrong-43k | LessWrong | Humans Are Embedded Agents Too
Most models of agency (in game theory, decision theory, etc) implicitly assume that the agent is separate from the environment - there is a “Cartesian boundary” between agent and environment. The embedded agency sequence goes through a long list of theoretical/conceptual problems which a... |
0ef9ede5-e4eb-48e9-b87b-a8adf7b25cf5 | trentmkelly/LessWrong-43k | LessWrong | Sal Khan publicly predicts the positive feedback loop associated with human intelligence augmentation. [Link]
http://www.khanacademy.org/science/cosmology-and-astronomy/v/random-predictions-for-2060
3:35 |
ad932bea-ff91-4708-854d-ee52a8dd3b94 | trentmkelly/LessWrong-43k | LessWrong | Naive self-supervised approaches to truthful AI
Useful background: TruthfulQA
Consider the following approach to (possibly) make a pretrained generative language model (like GPT-3) more truthful:
* Ask the model questions.
* Also ask a 'judge' copy of the model if its own answers to these questions are truthful. T... |
69fc393a-477a-4b59-918e-aee4798bae1d | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | It's OK not to go into AI (for students)
I've been running EA events in San Francisco every other month, and often I will meet a recent graduate, and *as part of their introduction* they will explain to me why they are or aren't working on AI stuff.[[1]](#fn-G5Rf9jLpcx6kgaD45-1) For the EA movement to be effective in ... |
224a21a3-a5e4-43cc-aafd-4e05030c45f7 | trentmkelly/LessWrong-43k | LessWrong | Towards an Axiological Approach to AI Alignment
NB: Originally posted on Map and Territory on Medium, so some of the internal series links go there.
AI alignment research currently operates primarily within the framework of decision theory and looks for ways to align or constrain utility functions such that they avoi... |
623b6f0d-9c23-4068-b0b9-ff673f3269d6 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Basic Inframeasure Theory
.mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direc... |
221318a1-7ddb-4671-a800-32afe798b907 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Dublin, Ireland Meetup
Discussion article for the meetup : Dublin, Ireland Meetup
WHEN: 12 August 2012 03:00:28PM (+0100)
WHERE: College Green
For anybody in Ireland, more specifically in Dublin, interested in meeting up. College Green Starbucks beside the wax museum. This has been postponed a week.
Disc... |
801b5716-9649-4d96-b5f6-d7dec9dba87e | trentmkelly/LessWrong-43k | LessWrong | Interpretability as Compression: Reconsidering SAE Explanations of Neural Activations with MDL-SAEs
This work was produced as part of the ML Alignment & Theory Scholars Program - Summer 24 Cohort, under mentorship from Lee Sharkey and Jan Kulveit.
Note: An updated paper version of this post can be found here.
Abstr... |
8eff7687-3d61-4ab6-a4eb-5532164d2cc9 | trentmkelly/LessWrong-43k | LessWrong | Meetup : West LA Meetup - Reasoning About Evolutionary Psychology & History
Discussion article for the meetup : West LA Meetup - Reasoning About Evolutionary Psychology & History
WHEN: 29 February 2012 07:00:00PM (-0800)
WHERE: 10850 West Pico Blvd, Los Angeles, CA 90064
When: 7:00pm - 9:00pm Wednesday, February 29... |
3449d88d-2df4-4820-a04e-0469784cdad1 | trentmkelly/LessWrong-43k | LessWrong | Gemini 2.5 Pro released
Google has now experimentally released Gemini 2.5 Pro.
As is standard, they claim it #beats all the benchmarks.
In my initial testing it's a very good model, definitely on par with GPT 4.5. It's a reasoning model, but runs very quickly, and shows you the chain of thought, unlike OpenAI models... |
e150c765-ec4b-4b73-991f-915b4ab7b6c5 | trentmkelly/LessWrong-43k | LessWrong | Calibrate - New Chrome Extension for hiding numbers so you can guess
One thing I wish I did more often is checks to see if I understand the world - Fermis, landmark numbers, predictions about how things work, etc.
Along the way I started wishing I could cover up numbers on websites so I would have the chance to guess... |
48a23aec-f242-487e-8df3-1170046ddb1c | StampyAI/alignment-research-dataset/arxiv | Arxiv | Planning to Explore via Self-Supervised World Models
1 Introduction
---------------
The dominant approach in sensorimotor control is to train the agent on one or more pre-specified tasks either via rewards in reinforcement learning, or via demonstrations in imitation learning. However, learning each task from scrat... |
dc16e53f-0b2e-4961-bfaa-14888fdd9bda | trentmkelly/LessWrong-43k | LessWrong | Cause Awareness as a Factor against Cause Neutrality
[Epistemic status: Sudden insight + some reflection]
[Novelty status: I Googled some stuff on cause neutrality, and didn’t see this category of concerns mentioned]
Trying to live an optimized, impactful life can be quite a burden. There is no time for fun when the... |
7b025b95-2c7e-4f76-ab75-795dfa901acd | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Abstracting The Hardness of Alignment: Unbounded Atomic Optimization
*This post is part of the work done at* [*Conjecture*](https://conjecture.dev)*.*
Disagree to Agree
=================
([Practically-A-Book Review: Yudkowsky Contra Ngo On Agents](https://astralcodexten.substack.com/p/practically-a-book-review-yudko... |
cec8644d-5734-443d-88d1-a20da31bb6c6 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | How might interpretability be helpful?
Current machine learning (ML) systems are mostly black boxes. Interpretability can help us understand them better and hopefully make them transparent enough so that we can prevent problems such as [goal misgeneralization](/?state=8QZI&question=What%20is%20goal%20misgeneralization... |
453628d5-0d9b-4f41-8206-fa8684f64c81 | trentmkelly/LessWrong-43k | LessWrong | Against the Open Source / Closed Source Dichotomy: Regulated Source as a Model for Responsible AI Development
Context
This is a short post to illustrate and start testing an idea that I have been playing around with in the last couple of days after listening to a recent 80,000 hours podcast with Mustafa Suleyman. In t... |
adb4d6bd-b4b3-48f2-85ad-701e4b5d508f | trentmkelly/LessWrong-43k | LessWrong | Self-inquiry
What is self-inquiry? Self-inquiry is an esoteric spiritual practice geared towards the deconstruction of the self. Why would anyone want to do that? What makes this practice spiritual? Why can it be found in the core of many spiritual traditions? What evidence of its effectiveness do we have?
First, why... |
343938b1-06c5-4292-ae18-9b26e5addb16 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Reinforcement Learning: A Non-Standard Introduction (Part 2)
**Followup to:** [Part 1](/lw/dsq/reinforcement_learning_a_nonstandard_introduction/)
In part 1 we modeled the dynamics of an agent and its environment as a turn-based discrete-time process. We now start on the path of [narrowing](/lw/iu/mysterious_answers... |
a7c8e91f-076c-490a-88c5-cfb3b6ef8b32 | trentmkelly/LessWrong-43k | LessWrong | How do I make sample size matter?
In the introductory example in the Wikipedia article on the Bayesian theorem, they start out with a prior distribution for P(machine_ID | faulty_product)* and then updates this using a likelihood distribution P(faulty_product | machine_ID) to acquire a posterior distribution for P(mac... |
37fa169e-1a8e-417c-b5aa-1b55be1f0aa2 | trentmkelly/LessWrong-43k | LessWrong | The Futility of Religion
As someone surrounded by religion, this mysterious and non-arbitrary belief system seemed convenient to believe during those times; but I was gravely mistaken. According to the Ontological, Teleological, and Cosmological Arguments, a god must exist.
These arguments provide the ideol... |
6e2700c7-3495-4436-90fe-9d2e3f925a4e | trentmkelly/LessWrong-43k | LessWrong | A moral backlash against AI will probably slow down AGI development
Note 1 : This was posted to EA Forum on May 31 here, as a submission for the 2023 Open Philanthropy AI Worldviews contest, due May 31, 2023. It addresses Question 1: “What is the probability that AGI is developed by January 1, 2043?”
Note 2: I'm deve... |
0ed5c09b-fb85-4547-a9a8-4957bfd50ad6 | trentmkelly/LessWrong-43k | LessWrong | Nudging My Way Out Of The Intellectual Mosh Pit
As part of my ongoing effort to improve my cost/benefit ratio on social media, I’m nudging myself away from intellectual mosh pit platforms like Facebook and Twitter, and towards blog posts, articles, videos essays, etc. Really longform consumption (e.g. books) remains a... |
74186e40-073b-4a8c-acf7-b7a21182e7be | trentmkelly/LessWrong-43k | LessWrong | I have just donated $10,000 to the Immortality Bus, which was the most rational decision of my life
I have non-zero probability to die next year. In my age of 42 it is not less than 1 per cent, and probably more. I could do many investment which will slightly lower my chance of dying – from healthy life style to cryo ... |
1bd28f0d-2d4b-4fa8-9024-37d8960de7e1 | trentmkelly/LessWrong-43k | LessWrong | Bankless Podcast: 159 - We’re All Gonna Die with Eliezer Yudkowsky
YouTube description:
> We wanted to do an episode on AI… and we went deep down the rabbit hole. As we went down, we discussed ChatGPT and the new generation of AI, digital superintelligence, the end of humanity, and if there’s anything we can do to su... |
87c57591-f9f2-4543-a5c6-99d8155fd7a4 | trentmkelly/LessWrong-43k | LessWrong | Competence vs Alignment
Is there a reliable way of determining whether an agent is misaligned, as compared to just incompetent? And can we talk of such distinction at all? (I suppose the orthogonality thesis implies that we can)
Let's say I give you an agent with suboptimal performance. Without having insights inside... |
eeb99ae9-5ab9-4627-ad67-e3b8e16b1aa2 | trentmkelly/LessWrong-43k | LessWrong | Availability Heuristic and Getting Stranded: Stay With Your Car Or Seek Help?
In every single news story that I can remember about someone getting stranded with a broken-down car, the person leaves their car looking for help and they end up dead. Or, if there are multiple people in the car, the one person who goes off... |
799c7739-2f28-4f7f-9da2-07da642b6017 | trentmkelly/LessWrong-43k | LessWrong | Alpaca: A Strong Open-Source Instruction-Following Model
> We performed a blind pairwise comparison between text-davinci-003 and Alpaca 7B, and we found that these two models have very similar performance: Alpaca wins 90 versus 89 comparisons against text-davinci-003.
Interestingly, Alpaca is trained using supervised... |
1f5a8cdb-fa4d-4342-8424-43470d05ed92 | trentmkelly/LessWrong-43k | LessWrong | Distinguishing logistic curves
This post will attempt to formalise the intuition that "it's hard to figure out the turning point of a logistic curve, at least until after that turning point". Ashort visual "proof" can also be found here.
Logistic curves
The logistic curves look like this:
Logistic curves can be s... |
e7dac64c-7921-4698-8dc1-9e54e37c8756 | trentmkelly/LessWrong-43k | LessWrong | What would it look like if it looked like AGI was very near?
I am writing a post on scenario-based planning, and thought it would be useful to get some example scenario symptoms on the topic of near-term AGI.
So the scenario-based planning thought exercise goes like this:
Imagine it is ten years from now and AGI h... |
672a9690-397f-4b5c-8df8-150aaf24f050 | trentmkelly/LessWrong-43k | LessWrong | New LW Meetup: Canberra
This summary was posted to LW Main on January 31st. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Inaugural Canberra meetup: 12 February 2014 07:30PM
* Meetup: First Meetup in Hamburg, Germany: 07 February 2014 07:00PM
... |
cf11f7c5-648f-4775-9f43-469c52f8c94a | trentmkelly/LessWrong-43k | LessWrong | How to deal with non-realism?
From Being a Realist (even if you believe in God):
> theists and untheists can and should meet half way and at least agree on the logical necessity of being a realist.
My mother, who doesn't call herself a theist (I think she's agnostic), doesn't even accept realism. She doesn't even a... |
bfb76796-575a-4bd9-906b-038de5fef12c | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post1583
This post was written in response to Evan Hubinger’s shortform prompt below and benefited from discussions with him. Thanks to Justis Mills for feedback on this post. If we see precursors to deception (e.g. non-myopia, self-awareness, etc.) but suspiciously don’t see deception itself, that’s evidence for ... |
ed8a6c63-277c-4ac0-803a-936f5ee65c6d | trentmkelly/LessWrong-43k | LessWrong | [Event] Join Metaculus Tomorrow, March 31st, for Forecast Friday!
Are you interested in how top forecasters predict the future? Curious how other people are reacting to the forecasts on Metaculus?
Join us tomorrow, March 31st @ 12pm ET/GMT-4 for Forecast Friday to chat with other forecasters and to analyze current ev... |
337937cf-c67d-421d-819c-89c013ea3a8d | trentmkelly/LessWrong-43k | LessWrong | Estimates vs. head-to-head comparisons
(Cross-posted from my blog.)
Summary: when choosing between two options, it’s not always optimal to estimate the value of each option and then pick the better one.
Suppose I am choosing between two actions, X and Y. One way to make my decision is to predict what will happen if ... |
60a41223-d76b-4116-b8c9-ddd2440a19ca | trentmkelly/LessWrong-43k | LessWrong | [ASoT] Some thoughts on human abstractions
TL;DR:
* Consider a human concept such as "tree." Humans implement some algorithm for determining whether given objects are trees. We expect our predictor/language model to develop a model of this algorithm because this is useful for predicting the behavior of humans.
* T... |
05a41a88-4be6-49fd-821f-0c8be9c9617f | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | [Linkpost] NY Times Feature on Anthropic
Written by Kevin Roose, who had the infamous conversation with Bing Chat, where Sidney tried to get him to leave his wife.
Overall, the piece comes across as positive on Anthropic.
Roose explains Constitutional AI and its role in the development of Claude, Anthropic's LLM:
... |
28dec673-8243-46c6-aba4-169e72805fa0 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Rebelling Within Nature
Today's post, Rebelling Within Nature was originally published on 13 July 2008. A summary (taken from the LW wiki):
> When we rebel against our own nature, we act in accordance with our own nature. There isn't any other way it could be.
Discuss the post here (rather than in the c... |
997eb2ae-8f43-4139-9d19-517ae370a0b1 | trentmkelly/LessWrong-43k | LessWrong | Don't want Goodhart? — Specify the variables more
In everyday life, if something looks good to a human, then it is probably actually good (i.e. that human would still think it’s good if they had more complete information and understanding). Obviously there are plenty of exceptions to this, but it works most of the tim... |
803609ed-6157-404c-99cd-c1bed235752c | trentmkelly/LessWrong-43k | LessWrong | Avoid News, Part 2: What the Stock Market Taught Me about News
It's been a decade since I blogged about the benefits of avoiding news.
In that time I mostly followed the advice I gave. I kicked my addiction to The Daily Show in late 2016 after it switched from ridiculing Trump to portraying him as scary (probably par... |
e98e95f6-a57d-47f7-bfa1-c3bf52f6478a | trentmkelly/LessWrong-43k | LessWrong | A Philosopher Walks Into A Coffee Shop
I have been really enjoying literarystarbucks.tumblr.com, which publishes complicated jokes about what famous authors and fictional characters order at Starbucks. I like it so much I wish I knew more great literature, so I could get more of the jokes.
Since the creators seem to ... |
71f02e86-fed1-4293-85c2-e7a2e460d427 | trentmkelly/LessWrong-43k | LessWrong | Deep Forgetting & Unlearning for Safely-Scoped LLMs
Thanks to Phillip Christoffersen, Adam Gleave, Anjali Gopal, Soroush Pour, and Fabien Roger for useful discussions and feedback.
TL;DR
This post overviews a research agenda for avoiding unwanted latent capabilities in LLMs. It argues that "deep" forgetting and unl... |
6cb4b57e-faf5-4567-9b05-900895142594 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Baltimore Area / UMBC Weekly Meetup
Discussion article for the meetup : Baltimore Area / UMBC Weekly Meetup
WHEN: 27 November 2016 08:00:00PM (-0500)
WHERE: Performing Arts and Humanities Bldg Room 456, 1000 Hilltop Cir, Baltimore, MD 21250
Meeting is on 4th floor of the Performing Arts and Humanities Buil... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.