Mastering Cross-Posting: From Facebook Frustrations to Dual Blogging Excellence | May 23 2026, 14:28

I have perfected the cross-posting from Facebook to my two blog sites [which almost no one visits] – beinginamerica.com and raufaliev.com. When a new post is published on Facebook, a mechanism is triggered to translate the post into English, process attached images, generate descriptions for them, create a title based on the text of the post and descriptions of the images, generate tags from the same basis, record the post in turso db – this is a cloud database, free up to certain limits, create embeddings via openai, record in qdrant cloud – this is also a cloud database, but vector-based, and finally, upload images to wordpress via API, and publish the post in English and Russian via API.

All would be well, but of all the APIs, the silliest one is Facebook’s. Firstly, for pages like mine, transitioned to New Experience, it’s almost impossible to use most of this API. Well, it’s possible, but you have to spend a long time proving to Facebook that you really need it, by showing startup documents, demonstrating the application, etc. Obviously, they are reluctant to deal with something that takes content out of their system. In addition, the token that gives access to the latest messages is relatively short-lived (possibly a few weeks), and it needs to be obtained anew through a browser only. So, any automation requires regular attention, otherwise it breaks.

If you mess up and don’t offload the latest posts through this Facebook Graph API in time, they just disappear from the list of recent ones and that’s it, no more API access to them. The only way is to request an archive download from Facebook. This download is also rather silly – it requires a lot of transformations and removing unnecessary stuff. For example, in the file containing posts, which I process, for some reason there are links that I sent in comments without accompanying text. And the comments are in a separate file!

To assign tags, I had to solve a separate challenge. Here’s the thing: there are about 10,000 posts over all time. That’s a big chunk, and you can’t build tags from it because it doesn’t fit into the contextual window of the LLM. But you need to. So, I did this: a script takes random posts from the 10,000 in such a volume that their total size is just below the specified limit in tokens, and at the end of this block, it adds the prompt “generate the most common tags for me, 30 pieces” (I simplify the prompt used). In the end, I ran this 10 times and got 10 sets of tags with 30 pieces each, generated for different slices of the database. That made 300 tags, some of which are complete duplicates, while others are synonyms and closely related in meaning. All this is fed into the LLM, and we get a list of tags and a hierarchy of tags. Now we have a limited set of tags that reflect the 10,000 posts as closely as possible. Turns out, that in almost 20 years on Facebook, my breakdown is as follows:

Tag Posts

==================================================

#Russia 3412

#Thoughts 3146

#Tech 3105

#Culture 2765

#Hobbies 2726

#AI 1603

#Science 1367

#Software 1358

#Travel 1298

#Learning 1138

#Society 1050

#Nature 958

#Education 915

#Business 902

#Art 894

#Programming 889

#Humor 840

#History 807

#Gadgets 750

#Moscow 713

#USA 614

#Cinema 567

#Webdev 493

#Music 476

#Sports 473

#Mindset 443

#Auto 400

#Books 386

…

and so on. This list includes both tags from the limited list and tags that the LLM appointed to content simply because it didn’t find anything suitable in the limited one.

Tags from the limited list became categories on the site. The rest of the tags + these just became regular wordpress tags.

As for image search. I had two ideas on how to do it. The first – OpenCLIP. It’s pretty straightforward but requires hosting the model somewhere. Easy on my machine, but inconvenient to start it each time, plus I planned to move the migrator to a cheap server on Amazon. It’s also okay to calculate in cloud models, but you have to pay a bit, which is yet another dependency. But the main thing – it works quite well without it. I generate descriptions for images using OpenAI, which is used for translating into English anyway, and then create embeddings using a large model. So far, all search tests are a great success. Especially when there’s text on the image, and it’s a big question whether OpenCLIP would have interpreted it successfully.

In the end:

1) wordpress raufaliev.com – free

2) wordpress beinginamerica.com – free

3) turso db where all posts are stored – free

4) qdrant cloud where embeddings are stored – free

5) openai for translation and image descriptions – not free, but inexpensive (cost $30 for post processing over a year).

I attach two screenshots – how the search by images works, and by texts, as well as the migrator dashboard.

Human Behavior Under Isolation: Lessons from the SPHINX Experiment | May 10 2026, 18:01

In the book Project Hail Mary, Stratt tells Grace that in the USSR there was supposedly an experiment where people were locked up for several months to see what would happen, and that the people almost killed each other, leading to the experiment being halted. That wasn’t the case, but I googled and found there was another experiment – SPHINX in 1999.

There were several groups. In the first, there were four Russians and during the New Year celebrations with alcohol, they beat each other up (10 minutes, blood, they had to be pulled apart). Another group had three guys and a girl (Judith Lapier) and the mission ended because during the New Year’s celebrations, driven by excitement, Judith was attempted to be kissed twice, leading to the mission being terminated.

As VICE reports, the dialogue was “We should try kissing, I haven’t smoked for six months. Then we can kiss after the mission and compare. Let’s experiment now.”

The team included doctors with degrees (Lukyanuk, Karashkin, Murashov) and Haider Hobikhozhin, who essentially was a randomly included technician with secondary education, taking the place of the Japanese man to the right in the photo and who was first in the second photo. Who beat up or kissed whom is now somewhat forgotten.

Vadim Gushchin, a coordinator from IMBP, after the scandal stated that the fight was “friendly,” and that Lapier “ruined the mission, the atmosphere, by refusing to be kissed.”

At the Canadian Space Agency, Lapier was told that such behavior is normal for Russians and that public complaints would be considered taboo in the culture of the host country.

Navigating Simple English in “Project Hail Mary” | May 10 2026, 15:30

I’ve read about a quarter of Project Hail Mary so far. The English is very simple, easy to read, captivating; the movie so far follows the book closely, but still, it makes reading quite interesting. However, I generally find it hard to read fiction because I keep getting distracted to google stuff. I reached the phrase “..I used the bathroom (or “head” I guess, because I was on the ship)…” and it got me thinking, it’s interesting to learn that the toilet is called differently on a ship not just in Russian. And why “head”? Turns out that “galley” in Danish and German is “head”. Interestingly, galleys are also found on airplanes, and historically, galleys were used only by sailors; officers did not use them.

The text is very childish, and understandably so – the main character is a physics teacher at a school after all. All these motherfluffer and dang it, gosh darn it, fudge, holy moly, for cripes’ sake instead of for Christ’s sake, there’s even bull-puckey instead of bullshit. “To go wee” is how they say “to pee” in the book. I recall, the day before yesterday we entered a mattress store, and the consultant, while discussing the topic “if one of you goes to the toilet, the other won’t even notice that the first one got up” – well, because the mattresses are so soft – actively used the verb “to pee”. So what? 🙂

Update: when the physics teacher encounters an alien ship on page 120, the chapter ends with holy fucking shit! That’s what all the rest was leading to;)

Occasionally, there are quite funny expressions that can even be used in life 🙂 For instance, the main character asks, “Who pooped in your Rice Krispies?” which is the idiom “to poop in someone’s cereal” – “who messed up your meal”.

In conclusion, if you’re choosing your first book to read in English – this one is at the top of my list. Even something seemingly simple like “Harry Potter” is more sophisticated, in my opinion. Here, there’s a lot of dialogue, school level but almost slang-free vocabulary, and a pretty interesting plot. Plus, it’s real science fiction, where the author educates the reader about the scientific method, how the world works, etc., all from the viewpoint of the hero, a physics teacher, who shares various facts and thoughts on how physics works, relating it to the plot in his interactions with other characters or thoughts to himself (rather than directly to the reader). It’s middle school level so far, but maybe it’ll get more complex later on.

Disappointed by Project Hail Mary: A Missed Opportunity for Smart Sci-Fi | May 04 2026, 16:40

We went to see Project Hail Mary yesterday. Honestly, neither Nada nor I liked it at all. Maybe we’ve just grown out of the age group that likes such movies. Comic book style. The alien, in my opinion, appeared on a budget. It feels like it’s at the level of the early 2000s – when you could have made exactly the same thing. Remember the movie Arrival, where they tried to come up with something unusual? Really, was this stone monstrosity worth 200M? The whole plot is full of cliches and banalities. In one minute, they made it so that the human and the alien began to understand each other perfectly via some program that a science teacher created in a day. As if the astronaut and a piece of rock just start chatting like buddies… 5 kilometers of iron chain the thickness of an index finger is about 10 tons of metal… did they have that much on the ship?

The directors became successful in the realm of Lego movies and Cloudy with a Chance of Meatballs, along with a couple of comedy series and 21 Jump Street. I was expecting a movie in the spirit of Interstellar or The Martian, smart adult sci-fi that is not afraid of scientific details, and instead I got a children’s fairy tale with an ugly alien.

I googled it, and it turns out that a ton of people are thrilled with it and it’s making a lot of money.. Probably, people are nostalgic for such amidst post-irony, satires, various narratives, Lanthimos, and “Battle after Battle”. As children, everyone watched “Flight of the Navigator” and “Short Circuit”, so adults are nostalgizing over simple goodness. Perhaps it’s just a family movie for watching with the kids. Then it might be okay.

okay, going back to watching the second season of Succession. It’s considerably better.

Exploring Word Clusters in Religious Texts from Gutenberg’s Library | May 02 2026, 03:28

It’s interesting that if you take 8000 books from the Gutenberg library and construct a graph for each based on word connections to see how “friendly” words are—if word A often appears with B, and B with C, then how often does A appear with C? There’s a metric for this—the average clustering coefficient. Then, simply sort the books by decreasing this coefficient, about 70 percent of the top will be religious books—bibles, the Book of Mormon, the Quran. Well, some of them are duplicates in a sense, because a Bible in different formats remains the Bible. But clearly, its different parts are grouped together, meaning, they definitely share commonality in these triangular words.

But what unites all the books in this top— is that they were written many years ago or, as in the case of The Night Land, written relatively recently in the same style as many years ago.

By the way, among these books shines An Introductorie for to Lerne to Read, To Pronounce, and to Speke French Trewly. This is a French language textbook, written in English during the Tudor times (around the 1530s). Soverayn lorde kyng Henry the Eight. It was written by Gilles Du Guez—a French teacher at the English court. This particular textbook was compiled for Princess Mary (the future Queen Mary I, known as “Bloody Mary”), the daughter of Henry VIII. Check out a page from the textbook. Very cool English 🙂 …ye must pronounce it letyng your lippes jointe close, so that there be but a lyttell hole in the middes.

So, I delved into this textbook. It mentions a fruit called “openarses.” As you understand, this is “open arses” in English. In Tudor England, they called a medlar an openarse. If you Google what a medlar looks like, you’ll have no questions why it’s called openarses 😉

In the anatomical section (MEMBRES LONGYNG TO MANNES BODY), the author mentions next to the eyes and ears “the nether beerde” (literally— “the lower beard”).

Harnessing Chat Data for Semantic Q&A Search | April 30 2026, 04:05

In one evening, I created a simple utility that extracts the Natural Language Processing chat for a year and a half – there are 65,000 messages, and converts it into question-answer pairs with semantic search available. Clicking on a search result (on the left) opens the dialogue in the chat. The messages that are responses to the question are highlighted. And at the top, the original phrasing of the question is highlighted as well.

How it works: the system assumes that people mainly reply to messages that are relatively close in the past. If several replies are made to one message, then it is likely useful and caught the interest of others in the chat. The system takes messages starting from the one many have replied to, ending with the last in the reply-to chain – and among such messages, it selects those that have at least 3 reply-tos to the original question. In essence, it cuts a piece from the chat starting with a popular question so that after the bottom cut, most likely, irrelevant content follows. Such blocks can overlap each other – for example, if someone asked a question while others were replying to something else.

So, if user A asked what the weather was like, and they received answers like “good,” “bad,” “rain,” and there were five messages without a reply-to, and then someone replied to “rain” with the question “why rain”, and five more people replied to this question, then the first question about the weather makes it into the system – the piece ends with 13 messages.

Afterwards, these pieces are summarized into question-answer pairs.

It turns out quite cool.

P.S. In the screenshot, the search query has nothing to do with the search result because I foolishly took the screenshot after I changed the query but before I hit send.

Peripheral Vision: Unveiling Optical Illusions in News Apps | April 29 2026, 17:56

I’m trying to figure out if it’s just me or do other people experience this too 🙂 if you look anywhere except at the word “Omurbekova”, the line highlighted in red in the second screenshot (which is actually white) is distinctly visible in your peripheral vision. But as soon as you shift your gaze directly to it, the line disappears. That is, it’s only visible peripherally. Share your experiences 🙂

Cherenkov Radiation: The Ethereal Blue Glow of Nuclear Decay | April 26 2026, 23:52

In the picture – Cherenkov radiation. This is me in 2009 at the Chernobyl Nuclear Power Plant, standing in the hall with the nuclear reactor. The water in the photo is for slowing down neutrons and cooling spent fuel rods. The glow occurs when electrons are ejected from the fuel at a speed exceeding the phase velocity of light in water. Kind of like a sound barrier, but for light. The intensity of this glow can roughly indicate how “fresh” the fuel is in the pool. The brighter and denser the blue, the more active the decay processes are. Interestingly, Cherenkov radiation is the reason why there is no absolute darkness at great depths of the ocean floor.