Replying toWhat's the Point of the Math?

For me a key benefit of maths is to answer the question "how much?", to turn qualitative intuitions into quantitative models.

For example if someone tells you "drug X binds to receptor Y which triggers therapeutic effect Z", the first question that comes to mind is "how much X do I need to take to get that much Z?".

If you don't answer that the info is not actionable. That's where the math models (pharmacocinetics and pharmacodynamics) come in, they tell you how much, which allows you to turn info into action.

Replying toMany can write faster asm than the compiler, yet don't. Why?

azergante1mo

Many can write faster asm than the compiler, yet don't. Why?

One implication of this worldview is that as long as there are still some identifiable high-leverage places where humans still write better code than LLMs

I agree (at least on the short term as you point out), but it seems hard to predict what these places will be (and thus hard to prepare for it), and it still seems likely that the market will be tough for the 90% of the programmers that are not experts in the specific niche things AIs are not good at.

Replying toMany can write faster asm than the compiler, yet don't. Why?

azergante1mo

Many can write faster asm than the compiler, yet don't. Why?

Your comment makes it sound a bit like there is no need for performance, but taking servers or REST services as an example, most programmers care about throughput, and almost all about latency which are both measured with e.g. prometheus. When your website takes one more second to load you lose clients, and if your code is slow it shows up on the cloud provider's bill. Even if you are IO bound, you can batch requests, go async, or do less IO.

The reason people don't bother hand-optimizing code is because the hardware is really fast, and because a handful of programmers put a lot of efforts writing optimizing compilers and optimized frameworks so the average output is good enough for the average workload.

-2

Replying toMany can write faster asm than the compiler, yet don't. Why?

azergante1mo

Many can write faster asm than the compiler, yet don't. Why?

Nitpick on the hashmap example: while I agree that the compiler does not produce optimal code in that case (which may be your main point), there is no need to write assembly to get the speed-up you describe, you can iterate the backing array in C. The compiler may or may not generate SIMD code however, so you may want to use SIMD intrinsics which is very close to assembly.

azergante2moQuick Take

Musings on human actions, chemical reactions and threshold potentials:

Chemical reactions don't occur unless a specific threshold of energy is reached ; that threshold is called the activation energy. Would it be fruitful to model human actions in the same way, as in they don't occur unless a specific activation energy is reached?

Chemistry has the concept of a catalyst: a substance that lowers the activation energy required for a reaction. Is there an equivalent for human action? On the top of my head I can think of a few:

being in a good mood
feeling safe
having a strong obligation
drinking alcohol
being familiar with the action

These are all catalysts: they make it easier to get started on... (read more)

Replying toStatus Is The Game Of The Losers' Bracket

azergante3mo

Status Is The Game Of The Losers' Bracket

Get your shit together and go play the winners’ bracket.

No ; if I want to play I do, if I don't I don't.

That's success.

This whole framing in terms of games is misleading. It doesn't matter what bracket you're playing at, if you feel you have to play you've already lost.

-4

-1

certain exotic neurotransmitters as SMART PILLS: or compounds that increase the capacity for mental work in humans

azergante

4mo

CERTAIN EXOTIC NEUROTRANSMITTERS AS

SMART PILLS:

OR COMPOUNDS THAT INCREASE THE CAPACITY FOR MENTAL WORK IN HUMANS

A story about LAZAR as told by Hosteen Nez

SECOND EDITION

Dedicated— To our valiant magic molecule testers and all others who helped.

This material not copyrighted, 1990. It may be copied and reproduced in any form without permission of anyone. If you do so you might give credit to LAZAR.

"A medicine is a true one if it cures. The test for everything is actual experience.”
"If I understand and you understand O my other What matters else?” (From the ancient Ayurveda) (1)

PREFACE:

I first visited with LAZAR at his hillside home overlooking the city in the summer of 1980. I had heard of this chemist for a couple of years and... (read 6366 more words →)

Replying toHow To Dress To Improve Your Epistemics

azergante5mo

How To Dress To Improve Your Epistemics

Alas, memetic pressures and credential issuance and incentives are not particularly well aligned with truth or discovery, so this strategy fails predictably in a whole slew of places.

Can you provide specific examples of places where this fails predictably to illustrate? Better: can you make a few predictions of future failures?

If I understand correctly, your position is that we lose status points when we say weird (as in a few standard deviations outside the normal range) but likely true things, and it's useful to get the points back by being cool (=dressing well).

It seems true that there is only so much weird things you can say before people write you off as crazy.

Do you think a strategy where you try to not lose points in the first place would work? for example by letting your interlocutor come to the conclusion on their own by using the Socratic method?

Replying toThe Rise of Parasitic AI

azergante5mo

The Rise of Parasitic AI

Wow. We are literally witnessing the birth of a new replicator. This is scary.

Replying toHigh-level actions don’t screen off intent

azergante5mo

High-level actions don’t screen off intent

High-level actions don’t screen off intent

, consequences do.

-1

Replying toChesterton's Missing Fence

azergante5mo

Chesterton's Missing Fence

Chesterton's Missing Fence

Reading the title, I first thought of a situation related to the one you describe, where someone ponders the pros and cons of fencing an open path, and after giving it thoughtful consideration, decides not to, for good reason.

So it's not a question of removing the fence, but that it was never even built, it is "missing". Yet the next person that comes upon the path would be ill-advised to fence it without thoroughly weighing the pros and cons, given that someone else decided not to fence that path.

You may think this all sounds abstract, but if you program often this is actually a situation you come across: programmer P1... (read more)

Plan the path to your goals so as to reap benefits regularly along the way, not only at the end

Hypothetically one could spend a few decades researching how to make people smarter (or some other long term thing), unlock that tech, and all that is really good.

But what if you plan your path towards that long-term goal such that it is the unlocking of various lesser but useful techs that gets you there?

Well now that's even better: you get the benefit of reaching the end goal + all the smaller things you accomplished along the way. It gives you some hedge: in case you don't reach the end goal you still accomplished a lot. And cherry on top: it's more sustainable as you get motivation (and money?) from unlocking the intermediary tech.

So it looks like it's worth going out of your way to reap benefits regularly as you journey towards a long term goal.

Why is LW not about winning?

azergante

azergante, Raemon

7mo

This is a bit of a rant but I notice that I am confused.

Eliezer said in the original Sequences:

Rationality is Systematized Winning

But it's pretty obvious that LessWrong is not about winning (and Eliezer provides a more accurate definition of what he means by rationality here). As far as I can tell LW is mostly about cognitive biases and algorithms/epistemology (the topic of Eliezer's sequences), self-help, and a lot of AI alignment.

But LW should be about winning! LW has the important goal of solving alignment, so it should care a lot about the most efficient way to go about it, in other words about how to win, right?

So what would it look like... (read 248 more words →)

A paperclip maximizer would finally turn itself into paperclips after having paperclipped the entire universe.

What AI apps are surprisingly absent given current capabilities?

azergante

8mo

[Epistemic status: a software engineer and AI user, not an AI researcher]

I could not find a readily available book database that offers semantic search with embeddings.

Amazon sells lots of books, wouldn't it be useful for them to propose such a tool to their clients, so they can easily find books they like? What about Netflix for movies? Maybe even other kind of products like clothes or something?

It's been a while since we got tranformers, it really feels like browsing latent spaces should be a thing by now. 3 years since GPT-3.5, why is the only app we have still a chatbot? Maybe everyone thinks it's perfectly normal, but I find it disturbing,... (read more)

ALICE detects the conversion of lead into gold at the LHC

azergante

9mo

Near-miss collisions between high-energy lead nuclei at the LHC generate intense electromagnetic fields that can knock out protons and transform lead into fleeting quantities of gold nuclei

8 May, 2025

Can I publish songs derived from the Sequences' posts on YouTube?

azergante

9mo

I am reading the Sequences and recently started to use LLMs to further analyze each post as I read them. As an experiment I also generate lyrics based on each post's content and analysis, and feed these to Suno AI to create songs.

I find some of these songs pleasant and they tend to be interesting as they pack a lot of meaning compared to mainstream music. I am now wondering if I have the rights to publish these on YouTube or other platforms.

From what I understand these songs are derived works and I need the author (Eliezer Yudkowsky)'s permission before publishing them, is that correct?

If so perhaps someone can bring this post... (read more)

Concerns about Organoids

Chatting with ChatGPT I learned that latest organoids have about 1 million neurons.

Wondering whether that's a lot or not, it tells me that bees and other insects have on the order of 10^5, fish, like zebra fish, have on the order of 10^6. So we are engineering fish brains, at least in terms of number of neurons. That's concerning, as far as I know zebra fish are conscious and can hurt.

What about humans? ChatGPT says humans have around 10^11 neurons, however 10-12 weeks embryo have about 10^6. It so happens that 10 to 12 weeks is the median limit to abort in European countries... and labs want to scale.

I barely know anything about the subject, but imagining a 12 weeks embryo's brain driving my roomba sure feels weird compared to the silicon version.

LLM-based Fact Checking for Popular Posts?

azergante

10mo

Hello,

I think having some LLM-based fact checking on LessWrong posts would be a valuable addition.

To some extent comments already serve this purpose, but LLMs can do this in a more automated and systematic way, side-stepping human blind spots and tendency to conserve energy.

1. Some pros and cons

Cons:

time cost to implement & maintain the feature
money cost to power the LLMs that do the fact-checking
potential errors in the fact-checking

Pros:

higher confidence that what is in the post is true
incentive for the author of the post to put more effort in researching the material
the post is read in a broader context, with opposite views (if any) readily available
spark some interesting discussions in the comments based on

... (read 18482 more words →)

Tip: you can ask ChatGPT to include confidence scores in its replies

Interactions with ChatGPT can be customized durably in the options, for example you can add the following instructions: "include a confidence rating at the end of your response in the format 'Confidence: X%'. If your confidence is below 80%, briefly explain why".

Here is a sample conversation demonstrating this and showing what ChatGPT has to say about its calibration:

Me: Are you calibrated, by which I mean, when you output a confidence X as a percentage, are you right X times out of 100?

ChatGPT: Reasoned about confidence calibration for 27 seconds. I don’t have calibration in that strict sense. When I mention a confidence percentage,... (read more)

Let me show you the ropes

There is a rope.
You hold one end.
I hold the other.
The rope is tight.
I pull on it.

How long until your end of the rope moves?

What matters is not how long until your end of the rope moves.
It's having fun sciencing it!

2+2=5 is Fine Maths: all you need is Coherence

[ epistemological status: a thought I had while reading about Russell's paradox, rewritten and expanded on by Claude ; my math level: undergraduate-ish ]

Introduction

Mathematics has faced several apparent "crises" throughout history that seemed to threaten its very foundations. However, these crises largely dissolve when we recognize a simple truth: mathematics consists of coherent systems designed for specific purposes, rather than a single universal "true" mathematics. This perspective shift—from seeing mathematics as the discovery of absolute truth to viewing it as the creation of coherent and sometimes useful logical systems—resolves many historical paradoxes and controversies.

The Key Insight

The only fundamental requirement for a mathematical system is... (read 540 more words →)

azergante's Shortform

azergante

This is a special post for quick takes (aka "shortform"). Only the owner can create top-level comments.

A Coordination Cookbook?

azergante

Coordination can improve default outcomes, i.e. what happens when individuals act according to their own interest assuming others will do the same.

For example in a flat share the default outcome may be that the common areas stay dirty because each flatmate is willing to spend time cleaning only if they are confident that all other flatmates will also spend time cleaning ; in the absence of coordination there is no such confidence so the common areas stay dirty.

Such coordination challenges come up often in daily life. Another example that comes to mind is choosing a leader, whether for a sport team on a small scale, or a country on a larger... (read 185 more words →)