Tag: Engineering

Exploring the Mystical Connection Between π² and g in Defining a Meter | March 01 2026, 17:11

It turns out that π² ≈ g is not some mystical coincidence. When the first scientists contemplated the definition of the meter, there was one elegant proposal: to make the meter equal to the length of a pendulum that takes exactly one second to swing from one side to the other.

For a mathematical pendulum, the period of oscillation is calculated by the formula: T = 2π √(L / g). If we take the length L = 1 meter and set the full period T = 2 seconds (so that it takes exactly one second for each half swing), the equation implies: g = π² (m/s²).

The definition of the meter was later changed: it was tied to one ten-millionth of the distance from the equator to the North Pole along the meridian passing through Paris. However, this geodetic definition was inspired by the earlier idea with the pendulum. Notably, both approaches match up with an accuracy of 1%. Essentially, since the old “pendulum” definition was the main candidate for a long time, values were adjusted so that the new meter was convenient and close to the measurements customary at that time.

It is also interesting that the number of seconds in a year roughly corresponds to the number of pi * 10^7. Earth’s orbital speed is about v = 30 km/s. The distance from the Sun to Earth is approximately r = 150,000,000 km. Thus, over a year, Earth travels a path of about d = 2 * π * r. Then, the orbital period equals T = d/v = π * 2 * r/v = π * 10⁷ seconds.

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.

Exploring Algorithmic Stylization in Plotter Art: A CMYK Fractal Journey | February 01 2026, 04:18

Now that I have a plotter, I am fully experimenting with ways of algorithmic image stylization. To achieve what is attached, a Minimum Spanning Tree algorithm was used. Essentially, it converts an image into stochastic rasterization – that is, where it’s darker, there are more dots, and then connects the dots with lines so that all points are connected in a single network, the total length of all lines is minimal, and there are no closed loops (meaning it’s precisely a “tree” with branches, not a “web”).

And this is what I do with each of the CMYK channels, then combine the result into a color picture. On this picture, there seem to be no other colors except for these four CMYK ones, but in reality, there is a bit because some smoothing has crept in.

Printing such on a plotter, of course, is difficult, I will be waiting forever, but I am getting the hang of it, I have already printed the first color picture (it turned out so-so. Well, the first pancake is always lumpy. Comments below)

Navigating the Future: Embracing Earth’s Magnetic Field as a GPS Alternative | January 10 2026, 17:41

I learned today that there is and is actively used a technology for navigation using the Earth’s magnetic field. It is used as a replacement or an extension of GPS.

For example, there is the Scandinavian ferry Express 5 of Bornholmslinjen, which insures against GPS problems (which do happen) by using MagNav navigation. Unlike GPS, the Earth’s magnetic field cannot be jammed or spoofed—it simply exists. The ferry follows the same route, and generally, navigation could even be achieved through household fishing sonars.

But there are a few startups that use this technology for indoor navigation, where GPS signals cannot reach. It’s claimed that the navigation accuracy is within 1 meter. That’s more interesting.

GiPStech, Oriient, Mapsted.

The basis of this technology is a process called magnetic fingerprinting. Engineers or mapping robots walk through a building with a smartphone, recording unique distortions of the magnetic field at every point. These distortions are created by the steel frame of the building, rebar in the walls, and large electrical equipment. A database is formed where each coordinate (x, y, z) corresponds to its unique magnetic field vector (intensity, inclination, deviation).

The collected data is uploaded to the cloud platform of the provider company. There, they undergo noise cleaning and are “stitched” together with the digital floor plan. When a user walks through a shopping center, their smartphone reads data from the built-in magnetometer in real-time. Special software (SDK) compares the current readings with those stored in the database. For accuracy to be within 1–2 meters, the system relies not only on magnets. It uses sensor fusion—combining data from the magnetic field with inertial sensors (accelerometer counts steps, gyroscope determines turns) and sometimes Wi-Fi/Bluetooth signals for rough localization.

This technology is certainly being actively implemented for drones. The main technical difficulty there is dealing with their own interference and considering that the magnetic field changes, requiring constant map updates. Electrics, engines create strong magnetic fields, which “drown out” the natural background of the Earth. However, various filtering algorithms (including neural networks) are used, which in real-time “subtract” motor interference from the overall sensor readings. From what I understand, at high altitudes (kilometers), the magnetic field is more “smooth”, therefore the accuracy is lower (about 1–5 km). But if several drones fly together and exchange signals, overall they can provide very good accuracy each. Additionally, a group of drones can measure the gradient (rate of change) of the magnetic field in space, tying location not to absolute values, but to relative ones. Essentially, using a group of drones turns the navigation system from a set of individual receivers into a distributed phased array antenna, capable of filtering global interferences and working with much weaker useful signals. Considering that small drones capable of staying airborne for long periods can be released into the air by the hundreds (and cost pennies), this is a quite promising area for military.

There’s an interesting startup, Zerokey. They release QUANTUM RTLS 2.0. This device provides spatial accuracy to 1.5mm. It’s used in production, for example. Their video shows a “watch” on a worker’s hand that monitors the correctness of assembling something on a table. Here, the principle is ultrasonic, and it’s understandable that these “watches” are paired with stationary sensors and further multilateration.

A Decade at EPAM: Thriving Through Change and Challenge | January 05 2026, 13:43

10 years at EPAM.

I would have never thought that I would enjoy working in the same place for an entire decade. What’s the secret? At EPAM, I am always evolving: projects change one after another, never letting me get bored.

I am currently on a project at a giant company: over 100 thousand employees and revenue of 30 billion dollars. Before this, it was the automotive industry — a behemoth with a staff of 175 thousand and a turnover of 150 billion. Somewhere around, there was a contract with a company of 80 thousand employees and 35 billion in revenue. True scale and genuinely serious challenges. And earlier, there were cosmetics brands, biotech, and the oil sector. In total, more than 20 projects of various calibers. Despite having over 100% workload every single day. And it seems that this year, I had more vacation than usual, yet still less than I could have taken. I traveled to Costa Rica, Mexico, Seattle, Antalya.

The point is, at each new place you learn something, sometimes from scratch. And that’s freaking awesome. It gives much more energy than if I had been “rooted” in any of these corporations for all 10 years. Perhaps, from a purely financial standpoint, people who stayed in one place at these companies earned more than me, but money isn’t the priority if it means sacrificing interest and enthusiasm. Living life at a job from which you are utterly exhausted is a questionable pleasure.

Last year at EPAM was maximally intense, and I sincerely hope that 2026 will not slow down.

The Unintended Consequences of Misguided Incentives | January 04 2026, 13:30

About KPIs. In English, there’s a concept called perverse incentive, “a harmful stimulus.” It occurs when you try to quash evil, but the methods become the perfect fertilizer for it. There’s a saying, “When a measure becomes a target, it ceases to be a good measure” (Marilyn Strathern based on Goodhart’s Law).

A classic example is the “Cobra Effect.” In colonial India, the British decided to reduce the snake population and offered a reward for every head. The plan seemed as reliable as a Swiss watch until Indians began breeding cobras on farms for the “harvest.” When the authorities realized they were being duped and cancelled the payments, the farmers simply released the now-useless snakes into the wild. As a result, there were many more cobras than before the program started 🙂

In a similar way, the French in Hanoi battled rats by paying money for severed tails. The city became overrun with lively yet tailless rats: the Vietnamese cut off the “currency” and released the creatures to breed further, to not lose a stable income.

In the 19th century, archaeologists searching for dinosaur bones and ancient fossils paid locals for every piece found. As a result, resourceful diggers intentionally shattered whole, priceless skeletons into small pieces to earn more by submitting them separately. Science wept, but the KPI for “number of finds” soared. A similar tragedy occurred with the Dead Sea Scrolls: Bedouins cut the found scrolls into small pieces to sell each fragment separately.

In the USA, this malady struck infrastructure. When building the Transcontinental Railroad, the government paid Union Pacific subsidies for every mile laid. In Nebraska, engineers, in a single corrupt impulse, drew a huge loop—the Oxbow Route. The extra 9 miles of detour made no sense for logistics but brought the builders hundreds of thousands of dollars “out of thin air.”

But if the “loop” in Nebraska was just theft, then the mistakes of U.S. Secretary of Defense Robert McNamara were a tragedy. An aficionado of numbers and mathematical models, he tried to manage the Vietnam War like a Ford assembly line.

When General Edward Lansdale timidly noted that McNamara’s formulas lacked the variable “the spirit and will of the Vietnamese people,” the secretary noted it in pencil in his notebook. And then erased it. He said that if something cannot be measured, it’s unimportant. The main metric became the body count. Officers onsite, eager to curry favor, began labeling everyone indiscriminately as “enemies,” painting an illusion of imminent victory in Washington, while the actual situation spiraled into the abyss.

In science, there’s a radical principle similar to Occam’s Razor— “Newton’s Flaming Laser Sword” (also known as “Alder’s Razor”). Its essence: if something cannot be tested by experiment (or measurement), it’s not even worthy of discussion.

It sounds reasonable for physics, but in life, it’s a direct path to what sociologist Daniel Yankelovich called the degradation of perception. He described this as a descent through four steps:

1. First, we measure only what is easy to measure.

2. Then we ignore what is difficult to measure or requires qualitative assessment.

3. The third step—we decide that what cannot be measured is not so important.

4. And the final step—we declare that what cannot be measured actually does not exist.

And at that moment, we become blind. We view the world through the keyhole of metrics, while in the room behind the door, cobras are bred, dinosaur bones are broken, and wars are lost.

Exploring ASML’s Advanced Chip-Making Equipment with Veritasium | January 02 2026, 00:47

Veritasium released a very cool report yesterday from ASML about the equipment used to print chips for your little phones, cameras, and laptops.

For those who aren’t familiar with the process. First, a monocrystal is grown from ultra-pure silicon and cut into thin wafers, then multiple layers of thin dielectrics, conductors, and semiconductors are repeatedly applied to the wafer surface, each time shaping the necessary areas using photolithography, etching, and ion doping, eventually creating billions of transistors and connecting metallic paths; finally, the wafer is tested, cut into individual crystals, and packaged into casings, making them into finished microchips.

This process had a limitation – the width of the paths and the distance to the next one are limited by the wavelength of the light used, and reducing it is difficult because there’s nothing to focus such a beam with – lenses simply absorb/reflect everything. In EUV lithography (extreme ultraviolet), the wavelength is 13.5 nm. This is virtually soft X-ray radiation.

The video explains details about the ASML machine costing 400 million dollars. Instead of refracting lenses, highly complex systems of reflecting mirrors are used. These mirrors are the smoothest surfaces ever created by humanity. If the mirror of this machine were enlarged to the size of the Earth, the largest bump on it would not be thicker than a playing card. To enable the mirrors to reflect X-rays, up to 76 alternating layers of tungsten and carbon, each less than a nanometer thick, are applied. All this is done by Zeiss. In addition, this mirror has a controlled curvature—it is constantly adjusted by robots with precision up to picoradians. The precision of the mirror control is so high that if a laser were mounted on it, directed at the Moon, the system could choose on which exact side of a 10-cent coin lying on the moon’s surface to hit with the beam.

But. We don’t have a “light bulb” that emits light in the EUV range.

To generate this light, a laser “shoots” at a droplet of molten tin the size of a white blood cell, traveling at 250 km/h. The first pulse flattens the droplet into a disc, the second and third turn this “disc” into plasma – and all this occurs within just 20 microseconds. When hit by the laser, the droplet heats up to 220,000 Kelvin — approximately 40 times hotter than the surface of the Sun. This plasma emits that very necessary light. And it does so 50,000 times a second. They say it’s been brought up to 100,000. Imagine, at a hundred thousand laser shots per second, it never misses a single one. All this happens in a deep vacuum. To clean the mirrors from tin particles, the chamber is constantly blown with hydrogen at a speed of 360 km/h — faster than a Category 5 hurricane. This process is described by the same formula (Taylor-von Neumann) that describes a nuclear explosion or supernova explosion.

The machine layers the chip with an error margin of no more than five atoms, while the matrix swings back and forth with an overload of 20G.

A single High-NA machine is transported in 250 containers on 25 trucks and seven Boeing 747 aircraft.

Link to the video – in the comments. Or search on YouTube on the channel veritasium.

Designing 3D Volleyball Training Tools on the Fly | January 01 2026, 21:21

What I did on the plane to/from vacation and sometimes in between: 3D visualization and editing volleyball schemes for Nadya (she’s a coach). This court in the attached image freely rotates, players can be placed on it, and the ball and player paths are shown – all in 3D.

The ball’s trajectory is calculated so that it does not cross the net when moving from A to B (Bezier formula). Players can take several poses – right now there are hastily made poses for serve, attack, block, pass/receive. Interestingly, in the code: I had to write a bit of “volleyball brains”. The system itself calculates the ball’s trajectory through Bezier curves so that it always passes over the net. Moreover, the height of the launch depends on the type of action: for an attack, the ball “launches” from a higher point than for a pass. I also added auto-rotation: the 3D model itself turns its face to where, according to the scheme, it needs to pass or run.

The longest and most difficult task was creating the 3D model of a female volleyball player. To generate a realistic volleyball player, I used the tripo3D service. It gave me a model in a neutral pose (for free). Theoretically, you can then use Blender and the Rigify plugin to attach an armature to it and move its arms and legs, which would recalculate the model.

However, in reality, this approach does not work well: the AI-generated model contains a large number of geometric errors, which the renderer forgives but Rigify does not. They can be roughly divided into two types — incorrect polygon normals and issues with non-manifold geometry, which are significantly more challenging to fix. Inside the body, there may be “floating” clusters of polygons or intersecting surfaces. When Rigify tries to calculate weights (which bone affects which part of the skin), this internal noise confuses the algorithm, and as a result, the weights are distributed chaotically (for example, moving the arm might start pulling the mesh on the stomach). Plus, the model is slightly asymmetrical.

Non-manifold is a geometry error where the topology of an object ceases to be correct in terms of a three-dimensional body: edges may belong to more than two polygons, polygons may only touch at vertices or edges without a common volume, and “hanging” surfaces or zero thickness may appear inside the model. Such geometry formally does not describe a closed volume, causing problems with rigging and deformations. Moreover, the model needs to be simplified because millions of polygons are not needed for rendering in real-time in a browser.

I fixed these using MashLab, additionally refining by hand (“with a file”). In the end, the model turns out slightly different from the original almost everywhere. The original model had “skin” in the form of textures – the face, shirt, and shorts had to be colored. How to transfer all this to a simplified model? For that, there’s a special operation in Blender called Baking. This also involves some tricks. In the end, it didn’t transfer perfectly, but perfection isn’t necessary yet.

Next, we attach the armature to the “joints”, and after about three hours of figuring out why everything does not work as it should, it finally worked. I made four poses, and now each circle (player) can be told which pose it is in.

I’ll also need to make dynamic changes to the uniform colors – that shouldn’t be difficult. There’s also an idea to transfer poses from photographs – this is more complicated, but generally feasible. Using MediaPipe/AlphaPose, you can detect key points in 2D, then some models like HMR/HybrIK can “lift” flat coordinates into 3D space, outputting relative joint rotation angles. The resulting data can be attempted to be projected onto a Rigify skeleton. Since the proportions of the generated volleyball player and the person in the photo may not match, that’s exactly why Inverse Kinematics (IK) is used. This part is quite complex, but overall it’s not strictly necessary – just interesting to figure out and make something functional.

Video in the comments

From Freezer to Fridge: A DIY Cooling Hack | December 19 2025, 00:56

Today I sold a refrigerator. It has a story. The essence of it is that it’s not a refrigerator, although it looks like one. It’s a freezer. And it freezes on average to minus 18 degrees. I bought it second-hand, thinking it was a refrigerator. The buyer also came today thinking it was a refrigerator.

And here I realize that minus 18 degrees is not at all what I need.

Well, I am a Solution Architect. I didn’t want to dig into it, I just drove to Lowe’s and bought a simple blinker. It turns on and off according to schedule whatever is plugged into it. I stuck a radio thermometer inside (I had one) and adjusted the blinking frequency (20 minutes) so that the internal temperature was on average +4 degrees Celsius. The radio thermometer showed that the temperature fluctuations were very small – nominally plus or minus 0.5 degrees from +4, even less. And so it worked for me for some months until I realized that I just didn’t need it.

Sold it today with the adapter. It’s gone to the people.