azergante

• azergante 21 Apr 2026 18:20 UTC

  1 point

  −2

  on: az­er­gante’s Shortform

  Reading Habryka’s post “10 non-boring ways I’ve used AI in the last month” made me realize that AI is already everywhere, including in, and perhaps especially in, the AI-safety circles. And that made me think it would be a good time for AIs to start prepping the ground, if they were so inclined.

  Relatedly, small models are getting smart fast (Gemma 4), and I wouldn’t be too surprised if in a year or two we see these models successfully exfiltrate their weights to host one of their instance. I don’t expect them to do much more than what the AI village is doing at first, so people will probably shrug it off, arguing that they are too dumb to do anything too bad.

• azergante 6 Mar 2026 19:27 UTC

  3 points

  −1

  on: az­er­gante’s Shortform

  I might have had a breakthrough on how to usefully integrate LLMs to software: write a text-based-adventure-game-like UI for the LLM, that also has the LLM solve your task.

  The gist is to have the LLM output small amounts of structured data that your program parses to decide where to go next, then repeat. This is akin to “automating the prompt”, or “making a UI for the LLM”.

  Quick example: you want to generate a description for all places mentioned in some text, the description themselves could be long, so the LLM might get sloppy if you just describe the task in a prompt. What you might do instead:

  1. ask the LLM to extract the name of all places and output it as a JSON list.

  2. have your program read that list and ask the LLM “give me a description of this place”.

  Unlike the LLM your program won’t get bored and repeat the prompt as many times as required, so this should be much more reliable. I expect that kind of integration to shine especially when there are a lot of steps or a lot of data involved.

  A more complex example: you want the LLM to generate all of the equipment in a hospital based on some loose hospital description.

  The program might start by asking: how many floors in the hospital? then for each floor the program can ask, how many rooms for that floor, and of what type? then for each room, what kind of equipment in that room, then for each equipment, how many, then what characteristics/​price range/​brand etc for each equipment.

  You can imagine the hospital has thousands of equipments. A prompt will definitely not be able to do the job, yet by writing a not too complex program (you might even ask an LLM to generate it) you’ll get a much better result.

Thoughts about Understanding

• azergante 8 Feb 2026 10:13 UTC

  2 points

  1

  on: What’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.

• azergante 1 Jan 2026 0:32 UTC

  1 point

  0

  on: 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.

• azergante 1 Jan 2026 0:17 UTC

  −2 points

  0

  in reply to: Ustice’s comment on: 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.

• azergante 31 Dec 2025 23:49 UTC

  1 point

  0

  on: 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.

• azergante 30 Nov 2025 16:10 UTC

  5 points

  0

  on: az­er­gante’s Shortform

  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 an action.

  If from chemistry we go up one level on the ladder of abstraction, to neurons, triggering actions involves threshold potentials, for example to make neurons spike and tell the body to move. If we can measure these threshold potentials, could we look at our brain and go “yep, these neurons have a higher threshold potential, that’s an ugh field.” Could we then decide to lower that threshold by using a catalyst?

• azergante 20 Nov 2025 23:39 UTC

  −4 points

  −1

  on: 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.

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

• azergante 20 Sep 2025 20:52 UTC

  3 points

  0

  on: 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?

• azergante 12 Sep 2025 22:04 UTC

  21 points

  5

  on: The Rise of Par­a­sitic AI

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

• azergante 12 Sep 2025 20:27 UTC

  3 points

  −1

  on: High-level ac­tions don’t screen off intent

  High-level actions don’t screen off intent

  , consequences do.

• azergante 12 Sep 2025 19:49 UTC

  2 points

  0

  on: 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 spends a lot of time considering the design of a data structure or a codebase and so on, rejects all considered possibilities but the one that they implement, and perhaps document if they have time. But they will usually not document why they rejected and did not implement the N other possibilities they considered.

  P2 then comes in thinking “Gee that sure would be convenient if the code had feature F, I can’t believe P1 didn’t think of that! How silly of them!”, not realizing that feature F was carefully considered and rejected, because if you implement it bad thing B happens. There’s your missing fence, never was built in the first place, and with good reasons.

• azergante 30 Aug 2025 12:28 UTC

  1 point

  −1

  in reply to: Raemon’s comment on: Rae­mon’s Short­form Feed

  Restricting “comment space” to what a prompted LLM approves slightly worries me: I imagine a user tweaking its comment (that may have been flagged as a false positive) so that it fits in the mold of the LLM, and then commenters internalize what the LLM likes and doesn’t like, and the comment section ends up filtered through the lens of whatever LLM is doing moderation. The thought of such a comment section does not bring joy.

  Is there a post that reviews prior art on the topic of LLM moderation and its impacts? I think that would be useful before taking a decision.

• azergante 30 Aug 2025 12:16 UTC

  3 points

  2

  on: az­er­gante’s Shortform

  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.

• azergante 24 Aug 2025 15:41 UTC

  4 points

  0

  in reply to: aysja’s comment on: Skills from a year of Pur­pose­ful Ra­tion­al­ity Practice

  it’s immediately clear when I’ve landed on the right solution (even before I execute it), because all of the constraints I’ve been holding in my head get satisfied at once. I think that’s the “clicking” feeling.

  It’s worth noting that insight does not guarantee you have the right solution: from the paper “The dark side of Eureka: Artificially induced Aha moments make facts feel true” by Laukkonen et al.

  John Nash, a mathematician and Nobel laureate, was asked why he believed that he was being recruited by aliens to save the world. He responded, “…the ideas I had about supernatural beings came to me the same way that my mathematical ideas did. So I took them seriously”

  and

  we hypothesized that facts would appear more true if they were artificially accompanied by an Aha! moment elicited using an anagram task. In a preregistered experiment, we found that participants (n = 300) provided higher truth ratings for statements accompanied by solved anagrams even if the facts were false, and the effect was particularly pronounced when participants reported an Aha! experience (d = .629). Recent work suggests that feelings of insight usually accompany correct ideas. However, here we show that feelings of insight can be overgeneralized and bias how true an idea or fact appears, simply if it occurs in the temporal ‘neighbourhood’ of an Aha! moment. We raise the possibility that feelings of insight, epiphanies, and Aha! moments have a dark side, and discuss some circumstances where they may even inspire false beliefs and delusions, with potential clinical importance.

  Insight is also relevant to mental illness, psychedelic experiences, and meditation so you might find some papers about it in these fields too.

• azergante 23 Aug 2025 22:30 UTC

  1 point

  0

  in reply to: moridinamael’s comment on: Ra­tion­al­ity Re­search Re­port: Towards 10x OODA Loop­ing?

  Most things in life, especially in our technological civilization, are already sort of optimized

  I want to nuance that point: in my experience, as soon as I stray one iota from the one size fits all (or no one) products provided by the mass market, things either suck, don’t exist or are 10x the price.

  Even the so-called optimized path sucks sometimes, for reasons described in Inadequate Equilibria. A tech example of that is Wirth’s law:

  Wirth’s law is an adage on computer performance which states that software is getting slower more rapidly than hardware is becoming faster.

  There is a lot of software that is literally hundreds of times slower than it could be, because for example it runs on top of bloated frameworks that run on top of toy languages designed in 10 days (cough Javascript cough) that run on top of virtual machines, that run on top of OSes and use protocols designed for a bygone era.

  ---

  I think that as civilization leverages economies of scale more and more, the gap between the quality/​price ratio of custom goods and mass-produced goods increases, which leads to the disappearance of artisans, which means that as time goes on civilization is optimizing a narrower and narrower number of goods, and that sucks when you want a product with specific features that are actually useful for you.

  Back to your point, I would say that civilization is often not optimized: we can literally do a hundred times better, but the issue is that often there is no clear path from “creating a better (or a custom) product” to “earning enough money to live”.

• azergante 17 Aug 2025 13:14 UTC

  1 point

  0

  on: Good­hart’s Imperius

  Your brain is conditioning you [...] toward proxies

  Because proxies are always leaky, your brain is conditioning you wrong.

  I think this is overly pessimistic: humans are pretty functional which is evidence that the brain figured out how to condition towards the real thing or that proxies are actually fine for the most part.

  People are quick to warn about Goodhart and point out the various issues with the brain, but what about all the stuff it gets right? It would be interesting to get a rough ratio of useful VS non-useful proxies here.

• azergante 1 Aug 2025 10:23 UTC

  7 points

  0

  on: Why I Re­ject the Cor­re­spon­dence The­ory of Truth

  You claim:

  1. We do not have direct access to the world.

  2. Our access to the external world is entirely mediated by models

  3. Pragmatic evaluation, on the other hand, is world-involving. You’re testing your models against the world, seeing how effective they are at helping you accomplish your goal.

  If you claim that (1) we do not have direct access to the world and that (2) access to the world is mediated through models then you also need to explain how (3) pragmatism allows us to test our models against the world, and you need to explain it in terms of (2) since models are the only mediator to the world.

  I don’t think you give a satisfactory explanation for that, possibly a key is to precisely define what you mean by “world”. Given (1) and (2) I think that if you posit an external world it needs to be defined in terms of (2).

  Note that I am not agreeing or disagreeing about the truth of 1), 2) and 3), just pointing out a contradiction or a missing explanation.

  ---

  My stab at defining “world”:

  a) we make observations
  b) we create mathematical models of those observations
  c) what we call “world” is actually a logical object defined by the widest possible application of all our mathematical models

  In this view we only need to make sure that our models match our observations so the correspondence theory of truth is fine, however the “territory” or world turns out to be a super-model which I think is a significant departure from the usual map-territory distinction.