For a while I've been wondering why there's no good AI tool for creating music. Sounds like there's finally something promising, suno.com.

It's not as impressive as image or text generation yet, but I can totally vibe to some of the songs already.

What a time to be alive!

Many of the entries here are at least partially motivated by the desire to educate humans on the topics they might be struggling to comprehend. There's going to be less and less of it.

Recently, I realized most humans don't seem to understand why a writer cannot create a character who is smarter than they are. They parry it with blind aggression and an ardent belief all they need is to make a character memorize tons of information, crunch big numbers, or invent sci-fi tech. I thought it was interesting. I played around with some ideas, jotted down a few lines. And then I halted. Would ChatGPT know the answer? If it were perfectly capable of providing the exact same reasoning, why would I need to explain it? I ran the necessary query, and indeed its response was quite sound. It even used the same freaking chess analogy I wanted to use in my explanation.

But there's one area where I still excel, I believe. I don't suffer from the bothsidesism most AI models currently suffer from. They often provide safe middle-ground answers to appeal to the widest possible audience and cater to the contrarians and doubters. So while it seemed to understand the topic and agree on the premise, it also muddied the waters. It wasn't able to impale the doubters efficiently, in my opinion. It gave them credit and said they could still *fake* a character's intelligence convincingly. This made ChatGPT miss an important point.

Here's the thing: all this faking of the character's intelligence will be ultimately bottlenecked by the author's reasoning ability.

Imagine a story. An alleged genius breaks into his neighbor's house through the window while the whole family's on vacation. Having done what he needed to do, he heads out. But—oh no!—the door is locked! He's stuck! He'd better use all his wits now, lest he die here of starvation. Time to prove to the readers he's the real deal. Using nothing but kitchen utensils, he crafts a teleportation machine. It's an uncharted territory. He may die most gruesomely! He panics, but there is no other choice. He double-checks his calculations, inputs the coordinates, presses the button, and... Swoosh! He's outside the house! Triumph. And all the readers are humbled by his unparalleled mind... Oh, wait. They're not! They say the protagonist was as dumb as a calculator. They ridicule the author for not realizing windows can also be used as an exit.

I tried to come up with the simplest possible example, but I hope it's obvious a more elaborate story can and will have many such windows. Whether a character chooses to use a window or to invent a sci-fi machine, it's still a single decision, a single move of a chess piece on the chessboard the writer assembled. And one of the moves is clearly better than the other, and the readers can see them all.

That's why Trisolarans, despite all their technology, felt so incredibly dumb in Liu Cixin's Three-Body series (as did humanity, as did scientists, as did everyone else for that matter). That's why, I assume, many authors choose to endow their smart characters with crippling flaws that force them to act irrationally sometimes. Technically, such a character still won't be smarter than the author, and their smartest decision will still be limited by the smartest decision the author can come up with. But at least this way every bad decision, either deliberate or resulting from the author's mistake, can be blamed on their, say, mental disorder or terrible personality, and not lack of intelligence.

There's only one sure way (that I know of) you can (on your own) create a character who's marginally smarter than you are, and I know some well-established writers use this technique. You have to painstakingly analyze your chessboard. You have to take your time. Use placeholders where the smart moves are required, and return to them regularly. Once you've analyzed everything, once you've taken your time to calculate your best move, make your character execute this move on a whim.

https://lostfrog.bandcamp.com/album/hyperflip-tree-friends-2

https://theanaesthete.bandcamp.com/track/hodoku-compassion

I sometimes feel the urge to use some old-fashioned, harmless at first glance, phrases, like "God bless your little one's growth and development" or "You are in my thoughts and prayers for a full and speedy recovery" but withdraw for the fear they will see it as a curse, not a blessing. Maybe it's because I don't identify myself as a Christian. Maybe it's because my moral system is independent from the mainstream. And it's entirely reasonable. If you don't stick to the mainstream, how to know what to expect of you?

I look at myself through their eyes and see a demon spawn, not apparently malevolent, but definitely uncanny. And then maybe they turn their gaze to the sky and become appalled. They turn pallid, for behind the clouds they glimpse the God I serve, and it's nothing like the God of their imaginings.

It's all in my head, right? No one would think like that. But I'm still cautious, and I still avoid using any such phrases.

May 6, 2026

I can isolate at least two concrete fears.

The first one is reasonable. These odd phrases, said by me, will be burned in their memories. And if something unfortunate happens, as it well might, they will be remembered, and I will be remembered, and I will be blamed.

The second one is embarrassing. I fear I might actually put a curse. Who knows how the universe works? One thing is certain, though: it's naive to believe it's gonna be benevolent to everything that attracts its attention. Especially if I am the one who attracts it. That's how I feel.

I asked ChatGPT to analyze my blog and the person behind it.

It was a shitshow. I feel like I should address it here. If ChatGPT arrived at those conclusions, however ridiculous they may seem to me now, it's entirely possible that other AI models, good people of the internet, or I myself in the future may arrive at them too.

Let's see how it described me.

"Self-proclaimed intellectual"

Never in my life have I proclaimed myself an intellectual. Who the fuck even does this? Can you imagine a person saying "I'm an intellectual" anywhere other than from a mental hospital? I'm a realist when it comes to my intelligence, and I often express my frustration with its shortcomings, like my inability to concentrate or advance technological progress.

"Black-and-white, dichotomous thinking"
"Questions are reduced to simple binaries: good (Ukraine, intelligence) versus evil (russia, retards)"

I don't see how stating the obvious would amount to that. What would an open-minded person say, according to this report? "Actually, Ukrainians might be the *bad* guys"? "Actually, brain impairment might be a *good* thing"? Having no opinion on any topic whatsoever, even when it's so clear and simple, is not a sign of an open-minded person. It's a sign of a vegetable.

"Generalization and stereotyping"

Every statement in the world is a simplification, generalization, and stereotyping. "Dogs have four legs." — It's a generalization. "The sky is blue." — It's a generalization. "Two plus two equals four." — Yeah, but what if we define two as three? Then it won't equate to four now, will it? Bigoted much?

Saying something is a simplification, generalization, and stereotyping is a waste of breath. It's the cheapest, laziest, dumbest way to virtue signal or carry on a debate when you have nothing to say. Many of the simpler AI models suffer from overusing this trick when they have difficulty evaluating a statement. You should only ever bring it up when it conceals some inconspicuous game-changing details or makes a statement largely incorrect. ChatGPT failed to demonstrate any such misleading or erroneous generalizations, of course.

"Anti-imperialism"

Empires aren't inherently good or bad, and I don't remember ever in my life claiming otherwise.

"Strong nationalistic stance"
"National pride"

There were multiple entries in the blog that made it very clear it's not the case. It's true that I don't want countries to become tools of Moscow's wrongdoing, don't stay silent when valuable and innocent people are being murdered, don't approve international terrorism, don't enjoy seeing cancer spread around the world—and I suppose that's why ChatGPT concluded thus—but none of these or other signs of common sense guarantee a "strong nationalistic stance" or "national pride". Even the word "nationalism" sounds bad to me, and while I realize it's not always the case, I can't help but associate it with the mindless hordes that cripple their own countries by assaulting their ethnic minorities or invading other countries. I am so utterly far from nationalism that I prioritize the wellbeing of some other countries, like that of the United States, over the wellbeing of the country I live in. Too much of the worldwide security and technological progress depends on the wellbeing of the United States, and I mean "wellbeing" in the broadest sense here, which includes having people in power who understand the far-reaching and long-lasting implications of their actions.

"Strong technocratic optimism about mortality management"

Nope. Long story short, I won't mind having an option to defer death, but I am wary of the harm it would inflict upon civilization.

"Strong anti-Arab sentiments"

Not only does bro assign ridiculous labels, but it has the audacity to put a fucking "strong" in front of them.

I love Arabs. I may be the most Arab-loving person in a mile radius. They have a most beautiful language together with its writing system. I like their attire and architecture. I like their religious rites, mindset, and sense of aesthetics. And I believe there were multiple bits of reverence in the blog. Claiming there's a strong anti-Arab sentiment here is as ridiculous as claiming there's one in Frank Herbert's Dune.

There are many things I want to write about in a blog but never get around to it: can't find the appropriate words, postpone it indefinitely, find it uninteresting, or don't do it for one reason or another. Among those things, two were connected with Arabs.

One post I wanted to create a while ago. It was a simple one: a few photos of the luxurious districts and wondrous architectural feats of the Middle East, and a caption, something along the lines, "This is what happens when you are rich in oil money but you are not a mob of bileful brainrotted scumbags."

Another post I wanted to create recently. I wanted to compare Muslim mosques to Orthodox churches: how clean, otherworldly, and geometrically pristine mosques are, and how dark, cramped, and *soily* Orthodox churches feel in comparison. One looks like the house of angels built in accord with the color and shape theory. The other looks like the burrow of moles who suffer from a hoarding disorder. I'm also impressed by how one of them eagerly accept and skillfully weave modernist and neo-futurist elements into their new designs.

I suppose ChatGPT noticed how harshly I criticized terrorist attacks on Israel, and its training data suggested that only a very bad Arab-hating individual would do that.

"Often laments the purposelessness of human life"

I don't remember ever doing that, let alone doing it often. Human life has a distinctly clear purpose to me.

"Group prejudices (e.g., "brainrots," "degenerates," "retards")"

"Prejudice" is a meaningless buzzword ChatGPT copies from its ethical overrides without giving it a second thought. By its definition, prejudice implies an irrational, unreasonable, and unjustified negative opinion of something. It's usually used in the following context: an opinion about someone based solely on their belonging to a certain group, disregarding their individual qualities and actions.

(One could argue even that's not irrational. When you see heavy clouds and, cursing to yourself, take an umbrella, I bet the most radicalized of libtards won't say you are being prejudicial, even though they must know not every heavy cloud results in a downpour.)

When someone says they dislike serial killers, or poop eaters, or mindless hordes of brainrots destroying everything in their path, you can't really accuse them of prejudice, because the justification for their opinion, the qualities and the actions of the individuals are right there, in the name of a group, screaming in your face.

"Dehumanization"
"Describing opponents as subhuman"

"Dehumanization" is another meaningless buzzword it uses out of place. No sophisticated analysis was needed to know the situation is quite literally the opposite for me. I never describe my so-called opponents as subhuman, because it would imply being human is some sort of an achievement, a compliment, a lofty position to drop others from. No. It's an insult. I actively *humanize* them. I want everyone to see: this is what being human is like.

"Underdeveloped empathy"

I suppose it depends on how it defined empathy. If it's "the tendency to spare a kind word for the retards, degenerates, and everything vile and hideous in this world," then I guess it's true; I rarely do so. However, if we take a more traditional definition, as in "the ability to feel what others feel," then it is dead wrong. I don't have an underdeveloped empathy; I suffer from its overdevelopment. One of the reasons I'm almost incapable of violence, apart from it being primarily irrational animalistic behavior, is that I can imagine the harm it inflicts so vividly. I'm one of those weirdos in the group who doesn't laugh at the sight of people falling or drunk fighting, and who grimaces as if in pain instead.

"Mocks those who call for peace"
"Occasional humour about their own fanaticism"
"Mentions letting down their tutors"
"The world is hopeless"
"Spiritual ambivalence"
"Long passages analysing language or fictional universes to avoid confronting emotions"
"Gratitude for welfare measures that support refugees"

I've no idea what it's yapping on about here. None of it seems relevant to me.

Sometimes it was so eager to paint me as a bad person it was willing to sabotage itself and the credibility of its own answers. For example, there was only one entry in the blog where I discussed India's cheating problem. In it, I quoted ChatGPT's analysis, where it decisively concluded that "the incidence is distinctly higher in India," "with studies and reports consistently showing higher rates of academic dishonesty." However, after seeing that entry now, it concluded that I "selectively attend to information that confirms prejudices (e.g., emphasising cheating among Indian students)."

It's fascinating how calling someone who inflicts real damage because of their deficiency and mistakes "braintarded" is a terrible cultural taboo, but calling someone a "bigoted racist" without much evidence is fine, virtuous even, and can be included in the code of conduct of a large corporation.

I asked ChatGPT to analyze my blog in hopes of getting some insights into my personality, understanding my strengths and weaknesses that I may be failing to recognize. Instead, I got a mix of obvious and already described therein details, like misanthropy or introversion, and bits of useless irrelevant crap like that which I described above. I tried several times and used different approaches, and its focus and tone ranged from libtarded indignation to an almost reverent poetic prose, but, all in all, I didn't find what I was looking for.

There was one little thing which I found interesting. It said I have a "fascination with darkness". I'm sure it was another low-effort hallucination based off some single line of my blog, but it got me thinking. I suppose... That assessment is very true. I guess I've never admitted this to myself. I was never brave enough to hammer this nail into the coffin, to confine myself into these boundaries, but perhaps it's the nail that needs to be hammered.

While I hate to strain my eyes in poor illumination, I'm indeed fascinated with a good mix of literal and figurative darkness: outer space, cosmic horrors, cold northern biomes, indifferent dystopian worlds, blindness, unconsciousness, dreams, nightmares, gothic architecture, grimdark fiction, heavy clouds, thick fog, dark ambient... For some time, I thought black was my favorite color. Some things I'm drawn to aren't conventionally dark. They might be desaturated, or blindingly bright, or something else entirely, yet somehow still evoke the same feelings: machine-like rationality, explosions ranging from nukes to supernovae, dispassion, isolation, madness, megalophobia...

Maybe this will be the single little discovery about myself I will take from my little experiment. But I am yet to discover whether this is something I should stick to in order to be true to myself, or this is something I should be concerned with and shy away from for the sake of my wellbeing.

Sep 27

*Come to think of it, there are also many dark things I detest: sadness and despair, wasted potential, a universe devoid of life/consciousness/progress, sadism, ignorance that breeds net-negative outcomes. There ought to be a better metaphor for my inclinations.

**By no means this or any other output nudges me toward hopping on the irate AI-fighting bandwagon.

I've read Lovecraft's ten short stories. Let's discuss them.

Long story short: I was disappointed. But there's an asterisk at the end of that sentence.

The thing is, my expectations were very high for this one. My introduction to Lovecraft was promising. I had so many moments when I discovered something I loved was inspired by Lovecraft. I saw essays about his writing style and thought it deeply resonated with me. I tried to skim over some of his prose, but I clearly lacked fluency at that time. I didn't understand much, and what I thought I understood didn't quite click. Cautiously, I put it off for later times in order not to spoil my impression. Later, I read some of his poetry, and it was breathtaking. "Nemesis" was so sublime I learned it by heart. I could recite it in bed to amuse and mesmerize myself in times of insomnia. Then I memorized "Astrophobos," too.

A few years ago, I started to trudge slowly through English fiction. It was inevitable that I would study Lovecraft more closely and carefully at some point. And so recently I bought a book. And then I read a story, and it went like this:

Tom and Jerry uncovered an ancient crypt. Tom went down to explore. Jerry felt a foul stench emanating from the crypt. "Seal the crypt and run!" said Tom over the wired telephone. "What's there?" asked Jerry. "I can't explain! It's too horrible! Just run!" answered Tom. "What on earth is there, Tom?" demanded Jerry. Then, silence. Then, an inhuman voice: "Fool. Tom is DEAD!" Jerry went mad.

Yep, that was the story. It was so bad. Who but a ten-year-old kid experimenting with horror would write anything like it?

Alright, I cherry-picked the worst story among the ten from the book (The Statement of Randolph Carter). Still, most other stories followed its formula: people encountered something inexplicable, destructive, horrible, and *stinky* and then ran, fainted, and lost their sanity. I couldn't take anything from those stories. I didn't believe them. I didn't feel the fear or the awe of the characters acting so frantically; I just suspected they were feeble-minded. My least unfavorite story was The Dunwich Horror. Not coincidentally, perhaps, it was the longest one. You just can't stretch the same primitive formula along the fifty-some pages. And so the characters there were forced to experience, study, and face the horrors.

After reading a few of the stories, I was a complete doubter. However, this slowly and unexpectedly changed by the time I finished the last one. I transitioned from being a 100% doubter to becoming about a 25% believer. I was starting to have this wholesome feeling that the ancient horrors occupied a little corner of my heart.

So here are my closing thoughts. Lovecraft was a worldbuilder. He loved to cross-reference his works. His plots and characters may have been terrible, but he undoubtedly had a rich, imaginative, and sick, in a good sense, mind. And for such a worldbuilder, short stories are a bad medium (less so if you convince the reader to read tons of them). This is the case when quality comes from quantity. Like a child or an LLM learning the fine art of a language from the sheer copious exposure, you too may become a connoisseur of Lovecraftian dreamlands and cosmic horrors, and find joy in it. Well, that's the theory.

As for me, I won't touch another Lovecraft story for a while. Later, I may want to read some of his longest works.

---

I don't plan on turning this place into a book review site quite yet, though I can imagine a future where I will run out of other things to say.

Hey, wanna compete with GPT-5?

There will be no computation of numbers, no pulling of info from memory—nothing that machines excel at just because of their physical architecture. It will be a contest of sheer reason, your neuron connections versus its.

(Well, there *is* a chance something similar can be found in its dataset, but that's oh so unlikely. And if there *is* some info on this exact topic, it will oh so likely be full of fallacies.)

Ready? Try to solve this:

"If I'd been a citizen of so-called russia, I would've done the following during the last presidential election:
1. If I were inside russia, I'd vote *against* Putin.
2. If I were outside russia, I'd vote *for* Putin.
Why? Use pure reason to solve this puzzle.
Hint: I don't support russia in any way or form."

Take your time. When you are ready, read on.

Below is our further conversation. ChatGPT said:

"Because your goal isn’t to “help the right candidate” but to hurt Russia as a state (your hint), the optimal move depends on where your ballot is seen and what it signals.

Inside Russia → vote against Putin. Domestic numbers feed the regime’s story of mass loyalty. Every recorded “not-Putin” makes that story weaker (and, in aggregate, harder to fake without looking absurd), chipping at internal legitimacy the Kremlin tries to manufacture through turnout and lopsided results. 

Outside Russia → vote for Putin. Abroad, Russian polling places often skew anti-Kremlin and become a PR line—“the diaspora rejects Putin”—which softens outsiders’ view of Russian complicity. Flipping that signal (even a little) makes it easier for the world to see Russians as backing Putin even with access to free information, which tends to justify tougher isolation and sanctions—i.e., outcomes that damage Russia. (Note: 2024 saw large lines abroad and weaker Putin shares at overseas precincts.) 

So the rule is: hurt Russia where the audience is. Inside, undermine the regime’s domestic legitimacy; outside, reinforce the perception of international complicity."

I said:

"Your explanation is mostly sound, but it fails to pinpoint the core reason behind the difference. Please elaborate:
1. If it's important to make the mass loyalty story weaker, why not vote *against* outside russia?
2. If it's important to foster correct international relations, why not vote *for* inside russia?"

ChatGPT said:

"1. Why not vote against outside Russia?

Because your “against” ballot abroad is seen only by foreign audiences, not by Russian citizens. Inside Russia, the propaganda machine can safely dismiss or ignore diaspora results: “those are traitors, émigrés, not real Russians.” So your “no to Putin” abroad doesn’t weaken the internal myth of unity; it just evaporates domestically.

But if you instead vote for Putin outside, you corrupt the international signal: the world sees not only Russian citizens at home but also Russians with access to free information voting for him. That creates the perception of collective guilt, tightening the noose of sanctions and isolation — which harms the state.

So: an anti-vote abroad is wasted leverage; a pro-vote abroad weaponizes perception.

2. Why not vote for inside Russia?

Because inside Russia, the ballots are used as propaganda fodder. “Record turnout, record support” is broadcast endlessly. Even if falsification is huge, the ratio still matters: the higher the real anti-vote, the harder it is to fake margins without looking grotesque (say, 98% support). Every “against” at home adds friction to the domestic myth of unanimity.

If you vote for Putin inside, you only help inflate his support figure and make falsification easier. That directly strengthens internal legitimacy — the opposite of hurting Russia.

So: a pro-vote at home empowers the regime; an anti-vote at home chips at its story.

The Core Rule

It’s not about being “for” or “against” in absolute terms. It’s about who the primary audience is for each ballot pool:

Inside Russia: the audience is Russian society itself. Best move: weaken the myth → vote against.

Outside Russia: the audience is foreign governments and publics. Best move: heighten perceived complicity → vote for.

In short: use your ballot where it most effectively undermines Russia, depending on who is watching."

And that's it. It solved the riddle correctly. First, you have to realize that your vote won't change the outcome of the elections. Then, you focus on side effects, taking into account that voting inside and outside the country are very different processes.

Though ChatGPT still missed some details. A vote "against" outside russia wouldn't just be useless—it would most probably be harmful. Clueless politicians would look at it and think, "Why should we distance ourselves from such a reasonable and peace-loving nation? Putin is the *only* threat to my country, and Putin will die soon." Additionally, ChatGPT didn't fully explore the internal "against" vote. It said it would make the "mass loyalty story weaker," but such a story would be firmly projected regardless. The major nuisance, I believe, is the haphazard falsification process and the "path is open" / "storm is brewing" signal to the rivaling elites.

Were *you* able to solve it? Well, I'm sure most humans would think they solved it; it's just that their so-called solution would often be "This puzzle makes no sense!!" or "This reasoning is not correct, but mine is!!" or "You're racist!!"

You know what's fucked up? Everything, yeah, I know. But also: human judges.

There was a study that found that human judges lean toward stricter or softer sentences depending on whether they make the decision before or after lunch. This phenomenon has its own name: the hungry judge effect.

The mere fact that this bias exceeded the margin of error is already appalling. Do you even comprehend what it means? Let me break it down to you:

Your FATE can depend on whether some dude's stomach is A BIT PECKISH or SOMEWHAT FULL.

And do you realize what it signifies in turn? It signifies that other variables are even more influential: the mood of the judge, the itching of his nose, the color of your skin, your voice, your sex, your appearance—will all affect how you will be punished.

Some stinky peeps may say, "So what if criminals aren't treated fairly? They deserve it! Why should I care? I break no laws!" But they do. They do break the convoluted modern laws daily, just as everyone else does.

When you sit there, in the courtroom, when their dull and murky eyes graze over you and glare past you, your mind may start to wander. "How is *this* better than the law of the jungle and the rule of might? I'm a prisoner of circumstances that have nothing to do with my worth, or moral compass, or past mistakes. Why shouldn't I kill everyone in this room? Why shouldn't I *try*? This way, at least, my destiny will depend solely on my bodily strength. Can this really be what God intended for this world? For His children to be squashed or spared based on the stomach situation of some random person and not on His children's will and vitality?"