Just thinking through simple stuff for myself, very rough, posting in the spirit of quick takes
 

1. At present, we are making progress on the Technical Alignment Problem^[2] and like probably could solve it within 50 years.

2. Humanity is on track to build ~lethal superpowerful AI in more like 5-15 years.
3. Working on technical alignment (direct or meta) only matters if we can speed up overall progress by 10x (or some lesser factor if AI capabilities is delayed from its current trajectory). Improvements of 2x are not likely to get us to an adequate technical solution in time.
4. Working on slowing things down is only helpful if it results in delays of decades.
   1. Shorter delays are good in so far as they give you time to buy further delays.
5. There is technical research that is useful for persuading people to slow down (and maybe also solving alignment, maybe not). This includes anything that demonstrates scary capabilities or harmful proclivities, e.g. a bunch of mech interp stuff, all the evals stuff.

6. AI is in fact super powerful and people who perceive there being value to be had aren’t entirely wrong^[3]. This results in a very strong motivation to pursue AI and resist efforts to be stop

There is a now a button to say "I didn't like this recommendation, show fewer like it"

Clicking it will:

• update the recommendation engine that you strongly don't like this recommendation
• store analytics data for the LessWrong team to know that you didn't like this post (we won't look at your displayname, just random id). This will hopefully let us understand trends in bad recommendations
• hide the post item from posts lists like Enriched/Latest Posts/Recommended. It will not hide it from user profile pages, Sequences, etc

Seeking Beta Users for LessWrong-Integrated LLM Chat

Comment here if you'd like access. (Bonus points for describing ways you'd like to use it.)

A couple of months ago, a few of the LW team set out to see how LLMs might be useful in the context of LW.  It feels like they should be at some point before the end, maybe that point is now. My own attempts to get Claude to be helpful for writing tasks weren't particularly succeeding, but LLMs are pretty good at reading a lot of things quickly, and also can be good at explaining technical topics.

So I figured just making it easy to load a lot of relevant LessWrong context into an LLM might unlock several worthwhile use-cases. To that end, Robert and I have integrated a Claude chat window into LW, with the key feature that it will automatically pull in relevant LessWrong posts and comments to what you're asking about.

I'm currently seeking beta users. 

Since using the Claude API isn't free and we haven't figured out a payment model, we're not rolling it out broadly. But we are happy to turn it on for select users who want to try it out. 

Comment here if you'd like access. (Bonus points for describing ways you'd like to use it.)

Selected Aphorisms from Francis Bacon's Novum Organum

I'm currently working to format Francis Bacon's Novum Organum as a LessWrong sequence. It's a moderate-sized project as I have to work through the entire work myself, and write an introduction which does Novum Organum justice and explains the novel move of taking an existing work and posting in on LessWrong (short answer: NovOrg is some serious hardcore rationality and contains central tenets of the LW foundational philosophy notwithstanding being published back in 1620, not to mention that Bacon and his works are credited with launching the modern Scientific Revolution)

While I'm still working on this, I want to go ahead and share some of my favorite aphorisms from is so far:

3. . . . The only way to command reality is to obey it . . .

9. Nearly all the things that go wrong in the sciences have a single cause and root, namely: while wrongly admiring and praising the powers of the human mind, we don’t look for true helps for it.

Bacon sees the unaided human mind as entirely inadequate for scientific progress. He sees for the way forward for scientific progress as constructing tools/infrastructure/methodogy t... (read more)

Communal Buckets

A bucket error is when someone erroneously lumps two propositions together, e.g. I made a spelling error automatically entails I can't be great writer, they're in one bucket when really they're separate variables.

In the context of criticism, it's often mentioned that people need to learn to not make the bucket error of I was wrong or I was doing a bad thing -> I'm a bad person. That is, you being a good person is compatible with making mistakes, being wrong, and causing harm since even good people make mistakes. This seems like a right and true and a good thing to realize.

But I can see a way in which being wrong/making mistakes (and being called out for this) is upsetting even if you personally aren't making a bucket error. The issue is that you might fear that other people have the two variables collapsed into one. Even if you might realize that making a mistake doesn't inherently make you a bad person, you're afraid that other people are now going to think you are a bad person because they are making that bucket error.

The issue isn't your own buckets, it's that you have a model of the shared "communal buck... (read more)

It feels like the society I interact with dislikes expression of negative emotions, at least in the sense that expressing negative emotions is kind of a big deal - if someone expresses a negative feeling, it needs to be addressed (fixed, ideally). The discomfort with negative emotions and consequent response acts to a fair degree to suppress their expression. Why mention something you're a little bit sad about if people are going to make a big deal out of it and try to make you feel better, etc., etc.?

Related to the above (with an ambiguously directed causal arrow) is that we lack reliable ways to communicate about negative emotions with something like nuance or precision. If I think imagine starting a conversation with a friend by saying "I feel happy", I expect to be given space to clarify the cause, nature, and extent of my happiness. Having clarified these, my friend will react proportionally. Yet when I imagine saying "I feel sad", I expect this to be perceived as "things are bad, you need sympathy, support, etc." and the whole stage of "clarify cause, nature, extent" is skipped instead proceeding to a fairly large reaction.

And I wi... (read more)

There's an age old tension between ~"contentment" and ~"striving" with no universally accepted compelling resolution, even if many people feel they have figured it out. Related:

In my own thinking, I've been trying to ground things out in a raw consequentialism that one's cognition (including emotions) is just supposed to take you towards more value (boring, but reality is allowed to be)^[1].

I fear that a lot of what people do is ~"wireheading". The problem with wireheading is it's myopic. You feel good now (small amount of value) at the expense of greater value later. Historically, this has made me instinctively wary of various attempts to experience more contentment such as gratitude journaling. Do such things curb the pursuit of value in exchange for feeling better less unpleasant discontent in the moment?

Clarity might come from further reduction of what "value" is. The primary notion of value I operate with is preference satisfaction: the world is how you want it to be. But also a lot of value seems to flow through experience (and the preferred state of the world is one where certain experiences happen).

A model whereby gratitude journaling (or general "attend to what is good" mot... (read more)

Narrative Tension as a Cause of Depression

I only wanted to budget a couple of hours for writing today. Might develop further and polish at a later time.

Related to and an expansion of Identities are [Subconscious] Strategies

Epistemic status: This is non-experimental psychology, my own musings. Presented here is a model derived from thinking about human minds a lot over the years, knowing many people who’ve experienced depression, and my own depression-like states. Treat it as a hypothesis, see if matches your own data and it generates helpful suggestions.

Clarifying “narrative”

In the context of psychology, I use the term narrative to describe the simple models of the world that people hold to varying degrees of implicit vs explicit awareness. They are simple in the sense of being short, being built of concepts which are basic to humans (e.g. people, relationships, roles, but not physics and statistics), and containing unsophisticated blackbox-y causal relationships like “if X then Y, if not X then not Y.”

Two main narratives

I posit that people carry two primary kinds of narratives in their minds:

• Who I am (the role they are playing), and
• How my life will go (the progress of their life)

T... (read more)

Over the years, I've experienced a couple of very dramatic yet rather sudden and relatively "easy" shifts around major pain points: strong aversions, strong fears, inner conflicts, or painful yet deeply ingrained beliefs. My post Identities are [Subconscious] Strategies contains examples. It's not surprising to me that these are possible, but my S1 says they're supposed to require a lot of effort: major existential crises, hours of introspection, self-discovery journeys, drug trips, or dozens of hours with a therapist.

Have recently undergone a really big one, I noted my surprise again. Surprise, of course, is a property of bad models. (Actually, the recent shift occurred precisely because of exactly this line of thought: I noticed I was surprised and dug in, leading to an important S1 shift. Your strength as a rationalist and all that.) Attempting to come up with a model which wasn't as surprised, this is what I've got:

The shift involved S1 models. The S1 models had been there a long time, maybe a very long time. When that happens, they begin to seem how the world just *is*. If emotions arise from those models, and those models are so entrenched t... (read more)

The LessWrong admins are often evaluating whether users (particularly new users) are going to be productive members of the site vs are just really bad and need strong action taken.

A question we're currently disagreeing on is which pieces of evidence it's okay to look at in forming judgments. Obviously anything posted publicly. But what about:

- Drafts (admins often have good reason to look at drafts, so they're there)
- Content the user deleted
- The referring site that sent someone to LessWrong

I'm curious how people feel about moderators looking at those.

Alt... (read more)

I want to register a weak but nonzero prediction that Anthropic’s interpretability publication of A Mathematical Framework for Transformers Circuits will turn out to lead to large capabilities gains and in hindsight will be regarded as a rather bad move that it was published.

Something like we’ll have capabilities-advancing papers citing it and using its framework to justify architecture improvements.

Just updated the Concepts Portal. Tags that got added are:

• Infra-Bayesianism
• Aversion/Ugh Fields
• Murphyjitsu
• Coherent Extrapolated Volition
• Tool AI
• Computer Science
• Sleeping Beauty Paradox
• Simulation Hypothesis
• Counterfactuals
• Trolley Problem
• Climate Change
• Organizational Design and Culture
• Acausal Trade
• Privacy
• 80,000 Hours
• GiveWell
• Note-Taking
• Reading Group

Huh, well that's something.

I'm curious, who else got this? And if yes, anyone click the link? Why/why not?

When I have a problem, I have a bias towards predominantly Googling and reading. This is easy, comfortable, do it from laptop or phone. The thing I'm less inclined to do is ask other people – not because I think they won't have good answers, just because...talking to people.

I'm learning to correct for this. The think about other people is 1) sometimes they know more, 2) they can expose your mistaken assumptions.

The triggering example for this note is an appointment I had today with a hand and arm specialist for the unconventional RSI I've been experiencing... (read more)

Test