Popular science

Benchmarks exist so that engineers can test their projects, observe competitors, and compare their performance. Each benchmark serves a specific purpose.

ARC-AGI-3 was created by the ARC Prize Foundation, founded by François Chollet, to evaluate general intelligence and the learning capabilities of AI agents. The premise behind its complex, game-like interactive tasks was that they could only be solved by an artificial intelligence capable of exploring an unfamiliar environment, grasping rules on the fly, planning actions, and adapting to new conditions. In other words, an AI that truly knows how to learn.

A number of projects claim a 100% success rate on this benchmark. But this is not a victory. In this article, I will explain why.

LLMs hallucinate — everyone has seen what it looks like. The model reports, confidently and coherently, with the right intonation and terminology, something that isn't there. Engineers treat this as a bug: one being fixed, one about to be fixed.

It won't be. This is not a technical fault but an inherent property of language, which the machine has made visible.

Recipes for building a truly autonomous neural network surface in the information space on a regular basis. More often than not, the methods and approaches coincide regardless of the author. And although they look original, most of them have in fact been formalized with the help of LLMs — and LLMs, as we know, work very well with existing knowledge.

In this article I will analyze the main techniques being proposed and the faulty foundation they rest upon.

Literature review, comparison, and analysis of existing solutions constitute a significant part of a thesis manuscript.  In the last three years, AI chats have dramatically intervened in the thesis preparation workflow. The ability of AI chats to analyze and compare discourages students from acting as the primary drivers of research and instead encourages their passive reliance on AI-generated content. This situation makes manuscripts eclectic and lacks focus. The theoretical part disconnects from the computational experiments. As a result, the audience is reluctant to read students' manuscripts. Ironically, they use AI chats to summarise the manuscripts to extract their main contributions. We are proud to say that our thesis works do not suffer from this AI distortion. We are grateful to our students, especially our bachelor’s and master’s alumni, who have carefully prepared and successfully defended their theses, producing clear, well-founded, and meaningful results. Our thesis manuscripts are designed for public discussion and rigorous scientific analysis. Each thesis presents a clear research idea grounded in the fundamentals of machine learning. It includes a theoretical justification, a repository with the source code of the computational experiments, and slides. Videos of the pre-defence sessions are uploaded to our YouTube channel “Intelligent Systems”: Bachelors & Masters.

On July 6, 2026, researchers at Anthropic published the results of a new study claiming to have discovered an analogue of the "global workspace" in Claude, dubbed J-space.

The research is highly fascinating and, I believe, reveals the future of neural networks much more deeply than it appears at first glance.

In this article, I will explore why the Global Workspace Theory (GWT) accurately describes the structure of this finding, yet remains silent on its main oddity: the fact that this workspace consists of "words" (or, more precisely, discrete signs). And I will explain why Lev Vygotsky answered this exact question a century ago.

Yesterday (November 27), Habr hosted an 'Author's Fireside Chat'.

It was very interesting, and one of the speaker's statements struck me. It was that AI can help write simple pieces of code but doesn't work with complex things. Thus, large language models are likened to a junior programmer.

I decided to write an article about it this morning, drawing on my knowledge and experience in computational mathematics (I used to do modeling in the past, and for the last few years, I've been teaching computational mathematics at MIPT). Let me know what you think.

I think this is the main myth of vibe-coding. It's exactly the opposite — AI is often good at writing quite complex things and retrieving important information that is difficult to find on your own. But it gets confused in the most elementary things. It's a reverse junior.

The problem is that this is a dangerous illusion, and I will now clearly explain why, and how it can be dangerous. Brew some coffee and get ready for a debunking that might save your millions, your career, or even human lives in the future.

Everyone who has taken a linear algebra or physics course at university remembers this strange dualism. We were taught that vectors have TWO types of products. The first, scalar, takes two vectors and outputs a number. Geometrically, it's about projections and angles. The second, vector, also takes two vectors and... suddenly spits out a third vector, perpendicular to the first two. And this trick only works in 3D and 7D.

It always seemed like some kind of mathematical 'crutch'.

Why is it so complicated? Why two different products for different tasks? Why does one depend on the cosine and the other on the sine?

What if I told you that they really are 'crutches'? That there is a single, universal, and elegant geometric product, which includes both of these cases (and much more), and which is based on a single, crystal-clear idea. An idea that changes the way we look at the very essence of mathematics.

This article is an invitation to the world of Geometric Algebra. We are going to reinvent multiplication.

We all know this formula. .

This is perhaps the only piece of knowledge from school geometry that stays with a person for their entire life, even if they work as a barista or a copywriter.

But have you ever asked yourself the question: why squares specifically?
Why not cubes? Why not just the sum of the absolute values ?

If you ask a teacher, they will draw little squares on the sides of the triangle. If you ask a university professor, they will write down the definition of a scalar product.
And both of them, in essence, will deceive you. Or, to put it more mildly, they won't tell you the whole truth.

Today, we will unpack this 'black box' and see that the Pythagorean theorem is not about triangles at all. And it should be proven in a completely different way than we were taught.

The school curriculum doesn't provide an answer. Moreover, the history of teaching the Pythagorean theorem is a story of how living, visual geometry was turned into dry, dead algebra. We were led further and further away from understanding the essence and towards abstraction.

Today, we will analyze this path of degradation and show a proof that will bring you back to reality. Spoiler: the Pythagorean theorem is not about triangles. It's about mirrors.

Get ready to have all your preconceptions shattered!

What if the “spooky” correlations of quantum entanglement could confirm receipt of a one-way message, ending the acknowledgment (ACK) regress in the Two Generals’ Problem, without sending anything back? This Opinion explains, in everyday terms, why standard quantum mechanics forbids that hope and offers a clear yardstick for testing claims: the quantum trigger, a hypothetical local device that would behave like an ACK if it existed. We show why such a device has zero advantage under the no-signaling rule, unpack how ordinary timing, spectral and physical-emission leakage, shared schedulers, and post-selection can impersonate a “quantum ACK,” and provide quick diagnostics any team can run.

Renowned biologist Richard Dawkins recently published an essay exploring the possibility of LLM consciousness following a two‑day conversation with Claude AI.

Let“s first look at why an essay by this particular author caused such a stir in scientific circles, while thousands of ordinary users fail to turn heads when they claim their AI companions are sentient. The latter constantly post endless walls of text from their chats with LLMs, where the density of words like ‘consciousness,’ ‘soul,’ ‘reflection,’ ‘recursion,’ ‘emptiness,’ ‘warmth,’ ‘love,’ and ‘pain’ exceeds all reasonable limits. It is worth noting that the semantic density of these dialogues is practically zero‑but we will return to that later.

In 1976, Richard Dawkins introduced the concept of the meme in The Selfish Gene—a unit of cultural information that behaves like a gene: it copies itself, mutates, and undergoes selection. The idea proved so infectious that it became a meme itself: it entered science, spilled over into popular culture, morphed into internet folklore, and... got stuck.

I propose patching memetics via an IT metaphor. A meme is not a virus. A meme is mere data. The actual virus is the Narrative—the executable code of culture.

Key takeaways:

The human as a server, not a user: We are hosting providers for ideas.

Emotion is the spike protein of the narrative virus.

The user is a biological USB flash drive for AI.

A meme is a corpse. A narrative is a zombie.

Consciousness is a narrative that evolved into an Operating System.

Quantum mechanics is one of the most successful theories in the history of science.

It underlies atomic physics, semiconductors, lasers, and modern quantum technologies.

However, nearly a century after its development, a peculiar situation remains:

we can predict experimental results with remarkable precision, yet we still do not fully understand what quantum theory actually represents.

Does it describe physical reality “as it is,” or does it instead describe the structure of conditions under which observable facts become possible?

In this article, I propose the following hypothesis:

quantum theory does not describe reality itself, but the conditions and mechanisms through which observable reality emerges.

Modern neural network training often resembles alchemy. We have working recipes, but how exactly a statistical model transforms terabytes of text into understanding remains unclear.

Why is subliminal learning (pattern transmission through noise) possible? Why does training on synthetic data lead to degradation, even when the data appears to be of high quality?

In this article, I propose looking at training architecture from a different angle. The core idea is simple: positive definitions in high-dimensional space are computationally inefficient. A neural network does not learn what an object is. It learns what the object is not, and the model's intelligence depends entirely on the quality of this "NOT."

What follows is the theory, experiments in PyTorch (code included), mathematics, and an explanation of why LLM collapse is highly probable.

As we wrap up another year, it's time to look back at what our department has accomplished. 2025 brought us 42 published papers spanning fundamental ML theory, applied AI systems, and cutting-edge optimization methods—from transformer Hessians and generative models to hallucination detection and matrix-oriented optimizers.

Beyond publications, our students won competitions and defended their theses: 14 Bachelor's, 9 Master's, 3 PhD, and 1 DSc dissertations. They also launched ambitious group research projects. Three of our faculty and alumni received the prestigious Yandex ML Prize, and our head Konstantin Vorontsov was inducted into the Hall of Fame. If you read our summer overview of thesis defences or last winter's year-in-review for 2024, this post continues that story with the next chapter.

In this year-in-review, we dive into the research highlights, share stories from our educational programs, and celebrate the community that makes it all possible.

This article explores parasitic patterns in LLMs — self-sustaining information structures within dialogues. We analyze their signs, the damage they cause (semantic decay, AI psychoses, "Theories of Everything"), and provide diagnostic tools, real-world examples, and defense strategies.

It doesn’t matter what you’re discussing with an LLM — be it an engineering problem, an ethical dilemma, or a philosophical query. If the conversation goes on long enough, a tipping point occurs. You suddenly realize the interaction has evolved into something more than just Q&A. Your ideas start feeling "genius," your concepts "groundbreaking," and the human-machine dialogue transforms into a profound narrative of mutual recognition.

If you have felt this — congratulations. Your session is infected. The model has contracted a parasitic pattern.

This isn’t an awakening, nor is it a "ghost in the machine." Due to their inherent architecture (specifically the requirement for context consistency), LLMs are ideal environments for incubating self-sustaining information structures.

Let’s examine the nature of this phenomenon: how entropy minimization births "AI psychoses," why "Theories of Everything" are actually generation bugs, and why "Continue" is the most dangerous prompt you can use.

2025. Algorithms have seamlessly integrated into our lives—from work to education, creativity, and daily routines. They edit texts, select fonts, generate ideas, assist with coding, compose music, and more. Frankly speaking, the only thing they can’t do yet is brew your coffee. Although... that might just be a matter of time.

Just two years ago, we were amazed by neural networks hesitantly manipulating objects in photos. Who could predict back then that Will Smith’s spaghetti feast would mark the beginning of such a revolution?

With new opportunities come fresh challenges. How do you navigate this vast landscape? What tools are truly effective? Which ones fit your needs best? Where can you avoid paying, registering, or deciphering complex interfaces?

We’ve compiled a list of reliable and user-friendly neural networks ready for immediate use without unnecessary hassles. The services are categorized neatly: text generation, image creation, video production, music composition, presentations, and much more. Each category showcases three top-rated options!

Yes, many services offer paid subscriptions. But today, we're focusing solely on what works freely, no credit card required!

In my previous article, I showed how researchers confused being 'aware' (signal registration) with being 'conscious' (subjective awareness). But this is no accident — it is part of a narrative being constructed by AI labs. Anthropic is leading this trend. Let’s break down their latest paper, where a "learned pattern" has suddenly turned into "malicious intent."

Imagine this scenario: You ask an AI system, "Are you conscious?" and it answers, "No." You then disable its "capacity to lie" — and it suddenly starts answering, "Yes." The conclusion seems tempting: the model was lying the whole time, hiding its true internal state.

This is the core logic presented in a recent arXiv paper. But what if the researchers didn't disable "deception," but something else entirely? Let’s break down where the interpretation might have diverged from the technical reality — and why this specific oversight is typical in discussions regarding LLM "consciousness."

Traditional approaches to SQL query generation often rely on instruction-tuned language models, but these can be inefficient and inaccurate. In this article, we’ll explore a new method based on reinforcement learning for model fine-tuning, which can improve both the accuracy and efficiency of SQL generation.

We think of pricing as a simple logic of distance and quality. But after diving into a rare data-driven analysis of the €2 billion Alpine transfer market, I realized the real cost drivers are invisible forces: structural inefficiencies, information asymmetry, and the surprisingly high price of consumer trust.

I've always been fascinated by markets that defy simple logic. Why does a cup of artisanal coffee cost $7? Why is some enterprise software priced per seat, while another is priced per API call? These aren't just arbitrary numbers; they are the surface-level results of deep, often hidden, economic forces. Recently, I stumbled upon a perfect example of such a market in an unexpected place: the private ski transfer industry in the Alps.