Valdes

Karma: 170

Valdes 21 Apr 2026 22:31 UTC
3 points
0
on: I built a semantic search engine for LessWrong
This sounds really cool and I intend to play around with it when I have time in a few days. First piece of feedback: I think it would be good to have easy to click examples on the welcome page, so we can get a feel for it without having to pick a post to test.

Also, a connected force graph (like in connected paper) would be appreciated.

Do you intend to make the corpus available / to make the code open source? If you put it on a public github I predict you might get some interesting contributions.

Valdes 18 Apr 2026 11:30 UTC
4 points
0
on: LLMs will soon disrupt algorithmic media feeds
For now, this is the equilibrium. Your options for media feeds are:
- corporate-aligned ML-recommended cocaine personalized to you for the purpose of optimizing attention
- voting systems (reddit, Hacker News, LessWrong) which aren’t personalized to you
- non-algorithmic manual curation (personal blogs, newspaper websites like ArsTechnica) which isn’t personalized to you
I think you missed the best one available right now: non-algorithmic manual curation which you personalize to you by yourself.

For example on Lesswrong you can subscribe to posts by authors, this leads to a feed (in the notification pane) that is currated to you. It’s how I got to this post. On youtube one needs to remove a lot of content views to avoid algorithmic recommendations but when you do so the “subscriptions” page remains, which is almost not algorithmically filtered. I personally use ublock origin with custom filters to block all algorithmic recommendation. For news, I subscribed to a handful of email newsletter and currently just one RSS feed, all in the same software (thunderbird).

I wouldn’t say I perfectly removed algo recommendations from my life (some are too trivial for me to bother finding how to remove them or too hard to remove), but I got 80% of the way there and it makes a huge difference.

The flaw of this approach however is that the curation to me mostly relies on a single criteria “who did it” or on existing curation lists.

Valdes 13 Apr 2026 14:48 UTC
1 point
0
on: The hard part isn’t noticing when papers are bad, it’s deciding what to do afterwards

First, I think every critique of a paper should at the very least understand the paper well enough to summarize it in a way the authors would agree with. Second, critiques should rarely dwell on typos, formatting errors, or lack of citations, and should ideally explicitly distinguish criticisms that are fatal to the core claim from ones that aren’t. Third, critiques should give the paper the benefit of steelmanning any ambiguous methodological choice before criticizing it.

I my experience, the first two are common in reviews. I went through the publication process quite a little under a dozen times, all in computer science, most in logic / AI conferences (conferences are the primary mean or publication in CS). More often than not the reviewer are explicitly expected to give an overview of the paper. Most reviewers put typos and small remarks in a separate section of their review.

I quite like the idea that reviewers should steelman the choices of the paper they are criticizing, that would certainly improve review quality. But it would make reviewing more time consuming and harder.

Valdes 12 Apr 2026 18:40 UTC
1 point
0
on: How to make good tea
I would love another post / additional section on “What you need to know about brewing”.

Valdes 26 Mar 2026 20:52 UTC
1 point
0
on: Enlightenment AMA
I would like to ask another question:

As an enlightened person, do you still use the “I” pronoun in your own mind? For example do you think to yourself “I should go get groceries” or “I like this movie”?

Valdes 11 Mar 2026 8:27 UTC
1 point
0
in reply to: the gears to ascension’s comment on: Don’t Let LLMs Write For You
I guess it will be fine if you directly describe what you’re doing and you make it easy to only read the prompts / skim the LLM outputs. Because few people share their LLM prompts when posting LLM writing, people are not fed up by the exercise. If anything it sounds like a “small fun experiment”.

Valdes 27 Jan 2026 0:19 UTC
5 points
0
on: Skill: cognitive black box flight recorder
It also seems wise to write down the content of the flight recorder as soon as possible after the event is over so the memory does not degrade too much. I suspect the kind of information the flight recorder is there for is also susceptible to being altered quickly over time (typically if you sleep before you write).

Valdes 26 Jan 2026 23:26 UTC
2 points
0
on: In My Misanthropy Era
I think it is true that most people think very badly from the perspective of what a rationalist considers “good thinking” and are very bad at saying true things. This does not prevent them from understanding things I (or you) do not. But Schopenhauer is wrong about the common man of today in the west, he comes on too strong. Wit and humour are popular in movies and most people have non-sensuous pleasures (companionship and love, for example).

He does have a point though. If you thought most people are roughly equal to you in intellectual capacity then you were probably wrong. In any case it seems healthy to nurture the ability to love and respect those that are deeply inferior to you at what you care about.

I have grappled with similar feelings and issues in the past. I have not fully solved the problem but I do think I can give an answer that helps somewhat: it is usual to love entities that are much dumber than we are. Cats are a good examples, they show their stupidity constantly. And it goes beyond stupidity; if cats were people they would not be good ones. Most cats are violent, selfish, and mentally weak.

This does not stop me from loving my cat.

I am not saying “think of people as cats”, at least not fully. But I think the cat example helps nurture a positive perspective on those who have flaws you would deem abhorrent in yourself.

As others have said, another good approach is to focus on the many ways in which we are often dumb ourselves and the many qualities others have that we do not. We are deeply inferiors to others at many important things, that does not mean we don’t deserve respect. Perhaps we can use this perspective and apply it to others as we apply it to ourselves.

Also I recommend you read this post on the bucket error if you didn’t already. I think it applies.

Valdes 24 Nov 2025 22:49 UTC
3 points
0
in reply to: lsusr’s comment on: Enlightenment AMA
Thank you. Deeply understanding that the self is just as much an abstraction as a car feels important indeed.

From the perspective of what is meant by “there is no self”, would I be right to say the self “exists” just like any complicated abstraction such as a car, a state, or a landscape; no more and no less?

Valdes 24 Nov 2025 12:51 UTC
3 points
0
on: Enlightenment AMA
I know I am late to this party, but maybe you are still taking questions? I have one about “no self”.

A car is made of parts. In a way there is no single part that is the whole car. The parts come together in a way that allows the function of a car to be performed and so we say that the whole is a “car”. In a sense there is no car, only parts that work together. The “car” concept is an abstraction. Likewise we could say “there is no rock, only a collection of atoms”. The rock is an abstraction.

Is this what is meant by “there is no self”? If not, what is the difference? If yes, why is this so important?

Valdes 15 Jul 2025 10:39 UTC
5 points
0
on: What do you Want out of Literature Reviews?
I think what I like the most is when there is a good summary story with the option to read more details on the parts that interest me, ideally with these complements forming their own coherent article. The structure this suggests is to have two posts:
- One that gives a high level picture focused on the general story, what I need to understand, and what I am expected to understand. With regular (hyperlink) references to parts of the second.
- One that gives a lot of details and can still stand on its own if I want a deep dive.
One great advantage of this is that the first also serves as a primer on the second, making it easier to read it.

I can’t think of an example of blogposts doing this, though I think there are some sequences on LW that follow this logic. But some books do this, with detailed chapters following high level summaries. “The mind illuminated” and “Against method” come to mind.

Book review: Against Method

Valdes21 Jun 2025 18:59 UTC

9 points

0 comments6 min readLW link

Valdes 6 Jun 2025 12:12 UTC
1 point
0
in reply to: Jiro’s comment on: The Stereotype of the Stereotype
A few others
- Death from the Castlevania series is simply evil
- Death in Supernatural is debatable. He is made sympathetic at some point but in his first appearance he is about to destroy Chicago then changes his mind with the reason “I like the pizza”.
- In the Thorgal series it is a complete asshole inventing new ways to put mortals close to being dead without being fully dead, for fun.
Also I think there are various stories and depictions where death is not a character but makes an appearance in some way, maybe purely symbolic, and seems to be evil or to enjoy humans dying. I am sure I have seen many depictions of laughing death as a metaphor of scenes of carnage (the first that comes to mind is a double page in the Darkmoon Chronicles where the burning of cities is illustrated with a laughing/smiling reaper).
So perhaps we could say that when Death is made a real character in stories it is more often than not a “kind hardworking man with a job to do”; but at the same time most depictions of the reaper are not characters but rather use a simple portrayal as a metaphor for evil and slaughter. Somewhat in the spirit of this famous image
If this is correct I think we can see where the friend in the post was coming from
“It’s nice to see a portrayal of Death that doesn’t paint him as some mindless villain, and actually gives him some characterization.”

Valdes 2 Jun 2025 16:59 UTC
2 points
0
in reply to: StanislavKrym’s comment on: How Epistemic Collapse Looks from Inside
I second that getting an AI to agree with your ideas proves very little. Also, it is somewhat intuitive to me that academia would always be suffering from various affective death spiral dynamics. You seem to have a specific one in mind, but I am not sure what it is or how it relates to the post.

Valdes 29 May 2025 20:23 UTC
1 point
0
in reply to: Morendil’s comment on: Agree, Retort, or Ignore? A Post From the Future
I am not sure if there were recent changes to the pages, but:

I don’t see an instance of the phrase “We suggest submitting links with a short description.” on the “About” page.

The “About” page links to the new users guide which is probably what you were looking for, even if that’s not quite a charter. Apologies if you already read it.

Granted I see no clear indication on that page that debates are welcome in the comments of all posts (and indeed I think there are some rare exceptions). But it does list a few things about “conversations” on the site, for example “we try to focus on what would change our minds”.

Valdes 29 May 2025 17:18 UTC
4 points
0
in reply to: Thane Ruthenis’s comment on: Is Building Good Note-Taking Software an AGI-Complete Problem?
I quite like some of your ideas about how to design note-taking software and I think I have a partial answer to the issues you point to:
- Make the repository of notes more semantic than usual, halfway between pure text notes and a structured database. If the notes exist in a graph database then it becomes possible to write database queries (probably with something like SPARQL).
- Allow the user to store queries within notes so the query is run whenever the note is opened. Maybe allow for a graph visualization of the query. This way the semantic structure becomes a part of the normal way to interact with the “hub” notes.
- The ontology used on a topic is defined by the structure of the links. If the user wants a new ontology they just create it on top of the old one and let the old one rot. No need to erase it and it does not distract from the new one (maybe there is some name reuse in the link/attribute names, but that’s trivial to fix).
- Use LLMs to correctly annotate notes with the semantic information that populates the database (links, tags, some attributes, …) This way they don’t need taste to know how to edit, they just follow the pattern that comes with a given ontology. Perhaps this requires the user to create templates.
- Provide a good CLI (Command Line Interface) for the software. This way the user can write scripts to update their notes using any language. It becomes easy to apply the new ontology to the old notes, possibly fully automagically but most of the time there is probably a bit of manual work left afterward. That’s fine.
- (unrelated to your post) let the GUI be a VScode extension so the user can use whatever other existing extension they like, this way there is no need to also design a good text editor.
I sent you a DM to talk more about this.
What links here?
- Thane Ruthenis's comment on Claude Codes by Zvi (9 Jan 2026 19:40 UTC; 44 points)

Valdes 23 May 2025 19:22 UTC
1 point
0
on: What determines female romantic “market value”?
Like (most?) other answers I can provide my own preferences as a mid-20 heterosexual man. So at least you get many datapoints.

The things most important to being someone I would want to have a long term relationship with are
1. Physical attraction
2. Positive influence on my life (compatible goals and desires, good at providing help I need, …)
3. Being of enjoyable company
4. Being high status, or rather being someone who will have a good impact on my own status
Physical attraction is not exactly the same thing as looks because looks also on an impact on the other factors and on how I estimate them. For example the way you dress might look good in my eyes but be a bad thing overall because it would be low status for me to introduce you to friends/family/colleagues.

I would say physical attraction is overall overhyped. You need to be above a certain level and if I don’t find you physically attractive at all there is little you can do to make up for it. But once a certain threshold is met you reach sharp diminishing returns.

Many things go in being a positive influence on my life. This includes having strength that go well with my profile and goals, for some characteristics this means sharing the same strengths as mine (if you and I both communicate well in the same fashion it can be great), for others this means having strengths that that help with my weaknesses (for example being good at handling situations I struggle with). Also we need to have compatible ideas regarding our life paths (children, where to live, …)

Being of enjoyable company is mostly about how we vibe. In this what matters is very much the end results: am I happy when we are together. Sex is a big part of it but not the majority.

Overall the best way to convince me you will have these positive influences on me in the future is to have them in the present (if you already help with some problems I have, I am likely to think you will help with other problems in the future). Likewise for being good for my status and especially being enjoyable to spend time with.

Lastly, I wrote above about the kind of profile I would likely want to be with, plenty into your framing of “market value”. But another big aspect of what relationships do end up happening is who we flirt with/date. A woman could be perfect for me and I for her but we won’t end up together if we don’t learn about eachother at some point and decide to try building a relationship. In this I believe the important things are:
1. Show you qualities (in my case, show you will have a positive influence/be enjoyable to be around/improve my status)
2. Make it clear it is ok to pursue you, socially and in your own view (it won’t bother you)
3. Don’t do it in a way that reduces your status. Making the first move is fine but it seems to me women who are too forward risk being seen as desperate / having some hidden flaw (“why can’t you get men to chase you? What’s wrong?”).
If I have to turn all this into advice about how to behave to make men like me seriously consider a relationship with you
- Be good looking. If you are plain looking to begin with then being reasonably fit and dressing well is probably enough.
- Be kind and open to others.
- Display some great qualities (smart, disciplined, creative, empathy, …)
- Be approachable
- Make me feel like I can help you in some way. This is both because it makes me think you are likely to like me and therefore it is worth trying to see if a relationship is possible; and because it is in some way deeply alluring to me and many other men to be helpful to women.

Valdes 22 May 2025 16:47 UTC
5 points
2
in reply to: RogerDearnaley’s comment on: The stakes of AI moral status
Small nitpick: “the if and only if” is false. It is perfectly possible to have an AI that doesn’t want any moral rights and is misaligned in some other way.

Valdes 23 Apr 2025 11:18 UTC
3 points
0
in reply to: jimmy’s comment on: Scaffolding Skills
I think you’re right but I also think I can provide examples of “true” scaffolding skills:
1. How to pass an exam: in order to keep learning with the academic system/university/school you need to regularly do good enough at exams. That is a skill in itself (read the exam in its entirety, know when to move on, learn how hard a question is likely to be depending on the phrasing of the following questions, …) Almost everyone safely forget most of this skill once they are done studying.
2. Learn to understand your teacher’s feedback: many teachers, professional or otherwise, suck at communicating their feedback. You often need to develop a skill of understanding that specific individual’s feedback. Of course there is a underlying universal skill of “being good at learning how individuals give feedback”; we could think of it as the skill “being good at building a specific kind of scaffolding”.
3. Learn to accept humiliating defeat: A martial artist friend told me it is important at first to learn to accept losing all the time because you learn in the company of strictly better martial artists. Once you get better, you presumably lose less often.

Valdes 24 Mar 2025 23:56 UTC
1 point
0
in reply to: Emrik’s comment on: Solving willpower seems easier than solving aging
This comment made me subscribe to your posts. I hope to read more on your attempts in the future! (no pressure)

Valdes

Book re­view: Against Method

Book review: Against Method