Month: February 2026

Harnessing Productivity: Personal Techniques That May Just Work for You | February 17 2026, 22:21

I formulated for myself how I manage to get a lot done (actually, I don’t). It’s not a fact that it will work for others. But still, here are the points:

1. Do what you like. You need to do what your heart is in at the moment. If you force yourself, efficiency drops tenfold.

2. Sports anger as a catalyst. You need to treat failure not as a tragedy, but as a personal insult from the task. Anger is the quickest way to enter a state of hyper-focus, turning “didn’t work out” into “oh really, now watch me”.

3. Seamless switching. As soon as energy in one task has waned or the task is done — leap to the next funnel that beckons right now. It might not always be work-related. For instance, I might play the piano, draw, program, write a book, or do work.

4. Completing. Take a chunk that you can chew and bring it to a plus-minus norm. Don’t drop it midway, while there’s still momentum. Finish and refine – that’s a task for the next “high”.

5. If you can’t break through a wall — don’t smash your head against it. Mark the point of stopping, say “I’ll remember you” and retreat to return with a different tool or a different mood. The main thing is to keep this “open gestalt” in active memory and not tuck it away for too long.

6. An external promise is sacred. If you promised a deadline by Monday, personal comfort (like sleep on Saturday) is sacrificed. This pain teaches you to filter promises in the future. Your word must have physical weight.

7. The plus one principle. Always do a bit more than expected of you. How much more is a question of context, resources, desire, but the delta should be tangible.

8. The principle of useful output. Any product of activity should be in a form that can be delivered. The English word for this is ‘deliverable’. Simply getting to grips with something is not a product. But understanding it and documenting it in Confluence is a product. A letter, an article, code – anything.

9. “You gotta, Fedya, you gotta.” Perform mandatory ceremonies and necessary bureaucracy as an inevitable evil that simply has to be done anyway. Need to pass some stupid training every six months? Allocate an hour for it and suffer through.

10. Have the best tools for the task. If you don’t have them, strive to possess them and learn to use them.

Three more principles, which seem unteachable but are very helpful:

0. Don’t get irritated and don’t irritate others.

1. The ability to instantly separate the important from the junk, and the urgent from the hustle. This is an intuition that only develops with years “in the field”. And total curiosity – the ability to find excitement in any topic. This applies to everything – including who to talk to and when to go to the store.

2. If you’re bored — it means you just haven’t dug deep enough. Interest is a matter of immersion scale and having the right people, books, YouTube videos, etc. Eventually, there simply are no topics that seem boring.

3. Lifetime learning principle. Any project is a legal excuse to become smarter at someone else’s expense. Look for what ignites you in routine and what you’ve long wanted to learn. See a task that would be more elegantly solved with a script in Haskell, a language you’ve never seen before? Then today, we are learning Haskell. True, enthusiasm should not bury the deadline. You need to deliver results, even if the experiment completely fails. Promise foundation first, then decorate with the new skill.

These principles have a downside. For instance, I progress very slowly in playing the piano because good progress requires two other principles that don’t “get along” with my principles above:

1. The “one more lap” principle. If you sat down, and got tired after an hour, you need to spend two more, and then you can get up.

2. The “clenched teeth – go” principle. If you’ve taken on learning something, do it regularly, preferably at the same time, and if necessary, through “I don’t want to”.

Understanding Fever: A Physiological Defense Mechanism | February 17 2026, 09:00

I’ve only slightly (hopefully) gotten sick here and realized during the process that many people around me take pills for a minor onset of fever, considering it normal.

I’m sharing my understanding of the process, which should be very close to scientific. When an infection penetrates the body, foreign bacteria or viruses enter the bloodstream, which the immune system attacks. During the attack, signaling molecules are produced, the purpose of which is to declare a general alert throughout the body. Specifically, cytokines are produced, which also inform the brain (hypothalamus) that action is needed. Pyrogens (fever-inducing agents) include cytokines and external pathogens. The hypothalamus activates a fever through the synthesis of prostaglandins. Why: at a temperature of 38.5°C, the immune system becomes more active, antibodies are produced in larger amounts, microbial reproduction slows down, and some viruses do not reproduce.

If you consume, for example, Ibuprofen, it blocks the enzyme (COX) that creates these prostaglandins. Meaning, the pyrogens are still in the blood, but the brain “can’t hear” them and doesn’t raise the temperature.

There are only two cases when you should reduce fever: if you truly feel awful, have a severe headache, vomiting, etc. Unnecessary stress does not help the body. And if the temperature exceeds 39°C. At that point, the harm from high temperature outweighs the benefits. Even then, there are so many “buts” that a doctor should make the decision. For example, if the heart is problematic, these are special cases.

Oh, here’s something else interesting. Why when the temperature is high you feel “cold” and want to cover up. In the hypothalamus itself, there’s something like a thermostat, normally set to 36.6°C. When pyrogens arrive, it raises the temperature through prostaglandins, but since it’s the brain, it immediately cranks up its own “normal temperature” in its thermostat to, say, 38.5°C. As a result, a body temperature of 37°C suddenly feels low, and it feels like “it’s cold around, need to cover up.” Covering up is passive thermal insulation, and generally, it helps to more quickly raise the temperature to the target level. Later, when the temperature reaches 38.5°C, the chills may disappear (unless the hypothalamus further raises its thermostat). And when the temperature plateaus, around 38.5°C, covering up is harmful.

When the temperature starts to drop back, the internal thermostat switches to 36.6°C, and to cool down faster, the body produces sweat. So, if you’re sweating, it’s a sign of recovery.

(Well, what else is there to do at four in the morning, when because of all this, I can’t sleep)

Exploring LLMs and AI: Connecting Neural Processors to Natural Language Learning | February 15 2026, 15:41

Some thoughts on LLMs and artificial intelligence in general. And in the end about neuromorphic processors and Intel Loihi.

As you all know, fundamentally LLMs operate on the principle of “propose the likely next word using the context from the previous N words,” and then the word enters the context, and the process repeats all over again for the next word. Well, and the context is also processed considering the importance of words.

Now let’s think about how children were taught languages in primitive societies. There were no alphabets, nor grammar. But the grammar itself, according to estimates, was quite complex—based on observations of the small languages of small peoples. Simple grammar is modern when the language has spread to millions and billions.

That is, a child’s brain had to reconstruct grammar in its neurons simply from the flow of speech from those around and through testing the understanding of what was said. It’s likely that the child was corrected if they spoke incorrectly, but somehow this grammar and sound extraction had to settle in the brain—and here the same mechanism as in LLMs is used: which words/sounds go next in what context is determined by latent and uninterpretable rules, which each person in childhood creates in their brain in their own way. That is, roughly speaking, it trains the ML model every time from scratch on the flow of speech from those around. A child does not know what a “case” is, but feels what ending is statistically more likely in a given context.

Actually, modern cognitive science (Karl Friston’s theory) asserts that the brain is literally a “prediction machine.” We constantly generate hypotheses about the next sound or word and correct them when they don’t match (prediction error).

The peculiarity of LLMs is that for them, teachers are texts and images, but for a child’s brain, it’s the living world around, and if all the texts they hear were digitized, their volume wouldn’t even be enough to train a very weak model. LLM sees the word “apple” next to the word “red.” A child sees an apple, feels its smell, taste, weight, and simultaneously hears the sound. This “stitching” of different sensory channels allows building neural connections thousands of times faster than on plain text. That is, modern LLMs take a brute force approach—simply observing the speech of billions, not just their immediate environment. A good question is how the human brain manages to learn from a relatively small dataset. However, it’s a big question whether this dataset is small—for example, lip movements, facial expressions, context provide a lot for building this neural network in the biological brain.

About the context: unlike LLMs, a child understands the speaker’s intention. If mom looks at a cup and says “hot,” the child’s brain limits the search space of meanings to one cup. And if he didn’t understand, he’ll get burned and remember.

One might assume, of course, that the brain already has a ready network at birth. It’s true, but science can’t yet explain it properly. Our entire genetic program has about 20,000 genes encoding proteins, and these 20,000 are responsible for everything—where and how the lungs, heart, bones, blood should be built, and they themselves are of mind-boggling complexity, and somewhere among 3 billion nucleotides and 20,000 genes this information must be recorded.

Apparently, genes encode not a map but an algorithm of self-assembly. Essentially, the architecture of the neural network is built dynamically, and this process begins long before birth. Then it is calibrated by all the signals received by the unborn child, and by the time of birth, there is already a somewhat tuned network in the brain.

It’s likely that the child’s brain is millions of neural networks of different “architectures” that evolve and merge in the learning process. Unlike LLMs, here learning and usage are strictly separated in time. But most importantly—the brain, although the most energy-consuming in the body, consumes very little energy in absolute terms, especially compared to the current “candidates for replacements in hardware.”

In the last few years, there has been active development in the field of neuromorphic systems (for example, the old IBM TrueNorth processor and the actively developing Intel Loihi). In conventional AI, neurons transmit numbers (0.15, 0.88…). In neuromorphic systems, they transmit “spikes” (impulses)—as in the living brain (and the architecture is called Spiking Neural Network – SNN). A few years ago, Intel released Loihi 2. Fully programmable. Neurons on Loihi can change their connections (synapses) right during operation. Supports plasticity—the very biological mechanism when the connection between neurons is strengthened if they often “fire” together. But the main thing—it consumes very little.

In this architecture, the model can continue learning “on the fly” right during operation, without forgetting old data (Continual Learning). Besides that—extreme energy efficiency.

Loihi 2 cannot multiply matrices as modern GPUs do, so completely new software has to be written for them (and this is moving very slowly). No PyTorch or TensorFlow—for Loihi there is only the Lava framework available today. And 1 million neurons from Loihi 2 is very little for LLMs. Therefore, Intel creates systems like Hala Point—it’s an array of 1152 Loihi 2 processors. It contains up to 1.15 billion neurons. Theoretically, in terms of performance per watt, such a system can surpass traditional GPUs by 10–50 times when working with AI models.

Experimental LLMs are already being launched on Loihi 2 (for example, models with 370 million parameters). They are not yet going to replace ChatGPT in the cloud, but theoretically, they are the future for “smart” robots and gadgets that need to understand human speech while running off a small battery.

We’ll observe. It might turn out to be a dud, or it could be another major revolution.

From Camels to Bishops: The Fascinating Evolution of Chess Pieces | February 14 2026, 16:24

It all started with a question – why does the elephant ♗ have this notch? And in general, where is the elephant, and where is the bishop, and is this notch about the elephant or the bishop? Anyway, listen to what I dug up, there’s a lot of interesting stuff here.

Chess originates from India. There, this figure was initially called a camel. And their elephant was what we call a rook – which if you think about it, a rook is basically a boat – or in English, rook, which if you think about it in Persian, it means chariot.

The name “Tura”, which we often hear in colloquial speech, is a pure import from Europe. In French – tour. In Italian – torre. In Latin – turris. All of these mean the same thing: tower. When chess arrived in Europe, knights and monks didn’t really understand what a “battle chariot” was (they were out of fashion by then), but they knew very well what a siege tower was.

So, returning to the elephant and the notch.

The short answer – to distinguish it from a pawn. But there’s a long answer.

When chess came to Europe, the Indian camel was switched to the Catholic bishop, and thus the piece was named bishop. The notch supposedly symbolizes a miter – the high headgear of clergymen. That’s precisely why in English the piece is called bishop. Though to me, it’s just a mouth from the Muppet show.

Interestingly, in French, it’s le fou – the jester. In German, it’s Läufer – runner. In Greek – officer (Αξιωματικός). Why officer? I don’t know, but I dug up that in Chinese chess, xiangqi (象棋), the “elephant” piece is indicated and pronounced as xiàng (象). This character indeed means “elephant.” However, in Chinese history, there was a high state office called xiàng (相), usually translated as “chancellor,” “prime minister,” or “chief minister.” This is a different character, although the pronunciation coincides. Probably, the officer comes from here too.

The chess knight is almost a horse in all languages, only in English and a few others, it’s a knight (although, in German, for example, it’s Springer – jumper, and in Sicily – donkey).

So, in German, there is a jumper and a runner. And a little horse in German is actually a king.

I also learned that there are ready-made solutions for ANY chess endgame in which there are seven or fewer pieces on the board, regardless of the position, the composition of the remaining pieces, or possible moves. This information, known as endgame tables, currently occupies 18.4 terabytes.

from the comments: “The most interesting thing is that this week a multi-year work was completed, and there is now a ready solution for any position with 8 pieces or fewer (7 pieces was already about 12 years ago, but there’s a very big difference)”

Interactive Text Enhancer: A Tool for Embedding Clarifications | February 12 2026, 16:11

I whipped up this thing in just an hour. Do you think anyone besides me needs it?

Here’s the idea. Take any text – a Wikipedia article, for example. Highlight any segment, say something unclear. The LLM gives us an explanation, and instantly inserts a box right in the text which you can click to open the explanation. In this explanation, there might be something unclear too. We highlight it with the mouse from this explanation, and a box appears there too. This continues until everything is clear. All the boxes remain in the text, so you can always return to them. So, if the idea was unclear to me, maybe it will be to others, and then a ready link with explanations will come in very handy. The result can be shared with colleagues.

For explanations, not just the fragment is used, but also the context. For example, otherwise, the highlighted word Terrier would yield text about a dog breed, not about the search system.

Chris Pratt’s Race Against AI in “Mercy”: A Cinematic Journey | February 10 2026, 16:24

We went to see the movie Mercy with Chris Pratt yesterday. Bekmambetov! His “screenlife” format has finally been expanded into a $50 million blockbuster and stuffed into IMAX. The guy really did well. First, he made six Yolki movies, and then, bam – he broke out and even started to produce something decent. (We were alone in the theater in super comfy motorized chairs. Empty halls — that’s pretty much the norm for the last many years. I don’t know how cinemas even break even. Even the bar was closed, it only works on weekends when more than two people show up to a hall)

So, the plot. The near future. The justice system is maximally optimized: instead of jurors and years of appeals — an impartial AI. The main character (Chris Pratt) is accused of brutally murdering his own wife. The evidence against him is significant, and society demands blood.

He is placed in a high-tech chair and given 90 minutes. This window” for defense — the time in which he must convince the algorithm of his innocence. If after an hour and a half the guilt probability” scale doesn’t drop below a critical threshold — he will be executed right there. Everything happens in real time, the movie runs for 90 minutes.

In the era of neural networks, this seems very timely. Screenlife here is ideal: we see the evidence and the world through the system’s eyes via cameras and browsers. Chris Pratt and Rebecca Ferguson on screen — always a plus.

However, what causes doubt is the attempt to crossbreed a hedgehog with a snake. Screenlife is good for its chamber feel, but here they sell us IMAX 3D, explosions, and chases, although 95% of the time the hero just sits in a chair.

Classic cinema for streaming. Not bad. On the couch with pizza on a Friday night — it’ll be great, there’s a solid detective story. Your brain might explode from the overload of details. Big question whether it’s worth paying for an IMAX ticket to watch Pratt watching a monitor… Who knows. There are some action scenes here and there, and they’re pretty good, but only occasionally.

Overall, detective fans should like it. From the plot, it’s clear they won’t fry the guy in the chair at the end of the movie, the question is how he’ll manage to wriggle out of it.