Book Cube: 2026, page 3
Book Cube publications from 2026, page 3: bilingual notes, sources, and original Telegram posts.
- · #Architecture
AI for Software Architecture: Why 2026- The architect's copilot still doesn't work. (Category Architecture)
AI for Software Architecture: Why 2026- The architect's copilot still doesn't work. (Rubric Architecture ) A systematic review of the literatureArtificial Intelligence Support for Software Architecture PracticeIt is a practical question - what architectural tasks AI can already support and why individual successes do not yet add up to a holistic practice. The first version appeared in 2025 year, but in 2026 It was updated in the year and 2 July 2026 year the article came out ACM Transactions on Software Engineering and Methodology.
- · #RnD
Materials about Loop Engineering (RnD heading)
Materials about Loop Engineering (Rubric RnD ) Prepared materials with podcast Research Insights Made Simple, who was with Maxim Smirnov on Tuesday:
- · #BookCube
Stream's off, come watch. https://www.youtube.com/watch?v= 4xRni7V6c
Stream's off, come watch. https://www.youtube.com/watch?v=\ \ 4xRni7V6c
- · #Architecture
How I and my agents got myself live on a static site today without the usual backend (Category Architecture)
How I and my agents got myself live on a static site today without the usual backend (Rubric Architecture ) This week I have four conversations with experts live: two have already passed, two are yet to come. The format of real-time communication ceased to be a one-off experiment for me, and at some point it became strange that my website was a real-time experience. polomodov.tech He doesn't know anything about it. I wanted a simple behavior: if the next seven days are scheduled to air, the site itself shows the announcement.
- · #AI4SDLC
Research Insights Made Simple 21How to Measure Coding Agents in Dialogue (Category AI4SDLC)
Research Insights Made Simple 21How to Measure Coding Agents in Dialogue (Rubric AI4SDLC ) What changes if the requirements for coding agents do not come in perfect promptom, but are clarified along the way? 17 July 16:00 MSC discuss Alexei Litvinov on the podcast "Research Insights Made Simple" 21" We’ll talk about the new SWE-Together and SWE-INTERACT benches released on the same day. I’ve been talking to both of you on this channel. (SWE-Together and SWE-INTERACT).
- · #Books
AI-Assisted Engineering: From Successful Prompt to Engineering System (Books column)
AI-Assisted Engineering: From Successful Prompt to Engineering System (Rubric Books ) I recently read Alexei Litvinov’s book “AI-Assisted Engineering”, with whom today we will discuss his book. live. Initially, Alexei invited me to read his book and give feedback. I was happy to read the book to see how I could put the disparate practices of working with AI agents into one system.
- · #AI4SDLC
Research Insights Made Simple 20Why coding agents do not cancel the examination (Category AI4SDLC)
Research Insights Made Simple 20Why coding agents do not cancel the examination (Rubric AI4SDLC ) What happens when a coding agent selects files, runs commands, and writes an implementation? Thursday, 16 July, with 16:00 before 17:00 MSC discuss Evgeniy Sergeev in "Research Insights Made Simple - Season" 1, Episode 20" Whitepaper Anthropic: "Agentic coding and persistent returns to expertise" Eugene Sergeev Director of Engineering at Flo Health.
- · #AI4SDLC
Loop Engineering: Why a Major Part of the Agent Cycle Is the Right to Say No (Category AI4SDLC)
Loop Engineering: Why a Major Part of the Agent Cycle Is the Right to Say No (Rubric AI4SDLC ) Tonight, 16:00 In Moscow, we will discuss the stream with Maxim Smirnov materialLoop Engineering: The Anthropic Playbook for Designing Systems That Prompt Your AgentsSo I had to prepare and read it beforehand:) Below I will leave a short press, and for the expanded version come on stream.
- · #BookCube
Before I had finished reading the book “Agentic Design Patterns” about the creation of agents, I began to read the book by Eliser Yudkovsky “If someone creates it, everyone will die.”
Before I had finished reading the book “Agentic Design Patterns” about the creation of agents, I began to read the book by Eliser Yudkovsky “If someone creates it, everyone will die.” I used to know Eliser as a popularizer of science and the author of Harry Potter and the Methods of Rational Thinking, and now I will get acquainted with his public position of Doomer. (doomer) in relation to artificial intelligence. I would like to note that the book has become a bestseller in 2025 I'm doing an interesting reading this year:))
- · #AI4SDLC
Who controls the coding agent: the person or the cycle he designed?
Who controls the coding agent: the person or the cycle he designed? We’re used to discussing the prompts, context, and binding of a single agent launch. But when the agent returns to work—on a schedule, an event, or the outcome of a past passage—the task becomes architectural: who finds the job, checks the result, stores the condition, and can stop the cycle. 14 July 16:00 I'll go to Moscow live with Maxim Smirnov. Let's disassemble the HuaShu material.Loop Engineering: The Anthropic Playbook for Designing Systems That Prompt Your Agents?
- · #AI
Yann LeCun on JEPA: predicting not pixels, but the state of the world (AI column)
Yann LeCun on JEPA: predicting not pixels, but the state of the world (Rubric AI ) I watched Jan LeCun's lecture.World Models: Enabling the Next AI Revolutionread 29 May 2026 A year at ETH Zurich. It turned out to be almost an exploratory manifesto: it is not enough for the physical world to generate the next token - you need to build an abstract state of the world, predict the consequences of actions and plan on top of it. By the way, Jan Lekun at the end of last year left the company banned in Russia Meta, where he had long been Chief Scientist.
- · #AI4SDLC
The code is cheaper. What became expensive? (Category AI4SDLC)
The code is cheaper. What became expensive? (Rubric AI4SDLC ) An AI agent can already write code in minutes that would have taken an evening. But one successful diff does not mean that development has become faster: a bottleneck can simply move into problem setting, review, testing and validation. Wednesday, 15 July, in 13:00 ICU will hold live on my TellMeAboutTech channel. with Alexei Litvinov in the video podcast "Code of Leadership".
- · #AI4SDLC
[2/2] Claude Code and Expertise: Why Session Success Is Not a Sign of Productivity (Category AI4SDLC)
\[2/2\] Claude Code and Expertise: Why Session Success Is Not a Sign of Productivity (Rubric AI4SDLC ) Keep going. analysis research by Anthropic"Agentic coding and persistent returns to expertise" about the use of Claude Code. In the first part, we talked about the methodology and results of the study, and now we would like to understand how to interpret them and how to assess the weight of the evidence given in the study. I would put them on a scale like this.
- · #AI4SDLC
[1/2] Claude Code and Expertise: What They Show 398 thousand (Category AI4SDLC)
\[1/2\] Claude Code and expertise What they showed 398 thousand (Rubric AI4SDLC ) Added to the body AI4SDLC meta-study 2026 fresh-work (16 June 2026) Anthropic "Agentic coding and persistent returns to expertise“, the title of which concludes that the coding agents do not cancel the examination. But what's more interesting to me is the authors' approach to measuring the division of labor between humans and Claude Code in real-world work sessions. This is a large but not experimental source.
- · #AI4SDLC
Nick Nisi on Agent Skills: Less Instructions, More Evidence (Category AI4SDLC)
Nick Nisi on Agent Skills: Less Instructions, More Evidence (Rubric AI4SDLC ) I saw it. shorthand Nick Nisi, Developer Experience Engineer at WorkOS, with AI Engineer Europe 2026. The speech passed 10 April 2026 It was released in London in May under the provocative headline:How I deleted 95% of my agent skills and got better results"although the report is not a provocation, but rather a good engineering analysis of why the long-prompt does not yet make the agent a reliable system."
- · #BookCube
This week my children and I spent in a sports camp near Tula, where they played football three times a day.
This week my children and I spent in a sports camp near Tula, where they played football three times a day. And I spent the first few days doing almost full-time AI engineering and it was fun. Then something happened and the Internet on the base disappeared, and the mobile Internet worked only on white lists. But I don't go anywhere without books, so I'm not confused.
- · #AI4SDLC
Stack Overflow Developer Survey 2026The survey itself has become a map of agent development (Category AI4SDLC)
Stack Overflow Developer Survey 2026The survey itself has become a map of agent development (Rubric AI4SDLC ) Half an hour ago I filled a new one. Big Stack Overflow (It's off. 23 June 2026 year and is still available for). I went there not so much for the usual set of “languages, bases, IDE”, but out of curiosity: what Stack Overflow this year asks about AI. And there's a shift that's clearly visible -- it's no longer about chat rooms or auto-additions, but an attempt to measure how AI and agents enter the production system.
- · #BookCube
And by tradition I attach the most interesting illustrations from the report "Cursor Developer Habits Report"
And by tradition I add the most interesting illustrations from the report.Cursor Developer Habits Report"
- · #AI4SDLC
Cursor Developer Habits Report: AI acceleration is clearly evident in advanced users (Category AI4SDLC)
Cursor Developer Habits Report: AI acceleration is clearly evident in advanced users (Rubric AI4SDLC ) Dealing with "Cursor Developer Habits Reportfor spring 2026There are a lot of interesting graphs about speeding up development, context growth and automation. But one of the most interesting insights is that productivity among users is growing not evenly, but sharply strengthens those who already know how to weave agency work into their day.
- · #BookCube
Illustrations from the SWE-Interact bench
Bench illustrations "SWE-Interact"
- · #AI4SDLC
SWE-Interact: how much does interactivity cost coding agents (Category AI4SDLC)
SWE-Interact: how much does interactivity cost coding agents (Rubric AI4SDLC ) I decided today to talk about the second article about the evaluation of coding agents in the dialogue.SWE-Interact: Reimagining SWE Benchmarks as User-Driven Long-Horizon Coding Sessionsfrom Scale AI. A curious coincidence: she came out on arXiv 29 June 2026 Same day as Meta's SWE-Together (banned)which I am today handler. The two teams almost simultaneously came to the same idea: single-mode benchmarks ceased to reflect the real work with agents.
- · #Culture
Orpheus: an ancient myth that works well as a rock opera (Category Culture)
Orpheus: an ancient myth that works well as a rock opera (Rubric Culture ) I was with my wife in May. rock opera OrpheusAnd that was the case when you accidentally cling to a couple of songs and then you go to a concert. I have a long-standing attachment to ancient Greek myths. As a child, I read a lot of them, especially the stories about Hercules: exploits, monsters, gods, demigods, the strange family conflicts of Olympus and the constant feeling that the hero almost never receives a simple reward for correct behavior.
- · #BookCube
Illustrations from the SWE-Together bench
Bench illustrations "SWE-Together"
- · #AI4SDLC
SWE-Together: How to measure coding agents in dialogue rather than in one go (Category AI4SDLC)
SWE-Together: How to measure coding agents in dialogue rather than in one go (Rubric AI4SDLC ) I read the latest article.SWE-Together: Evaluating Coding Agents in Interactive User Sessions"from the team of the banned in Russia company "Meta" (arXiv, 29 June 2026). The authors hit the weak spot of almost all benchmarks of coding agents: they are static. The agent receives a full statement of the task immediately and is evaluated by the final code.
- · #AI4SDLC
Matt Pocock on Agent Skills: How not to fall into skill hell (Category AI4SDLC)
Matt Pocock on Agent Skills: How not to fall into skill hell (Rubric AI4SDLC ) Watch Matt Pocock perform on AI EngineerBuilding Great Agent Skills: The Missing Manual"Skills are discussed as engineering artifacts: with classification, support price and a clear quality criterion. The main idea I took away from it is that good skill is needed to get a predictable process from a probabilistic system.
- · #AI4SDLC
GitLab AI Accountability Report: Code is already generating faster than it can control (Category AI4SDLC)
GitLab AI Accountability Report: Code is already generating faster than it can control (Rubric AI4SDLC ) Dealing with GitLab 2026 AI Accountability Reportwhich the company released 23 June 2026 years. This report continues the line well. GitLab Act 2, DORA ROI recent Stack Overflow Pulse SurveyAI has already accelerated code generation, but the bottleneck has moved into control, verification, and accountability. The study was conducted by The Harris Poll for GitLab: 1 528 Developers and purchasers in six countries.
- · #Books
DeepSeek in action (DeepSeek in Action) (Books column)
DeepSeek in action (DeepSeek in Action) (Rubric Books ) I finally finished reading DeepSeek in Action, which was funny. 1 July 2025 Yatsenkov was the translator of the year at DMK Press, and the original is called “DeepSeek in Action: LLM Deployment, Fine-Tuning, and Real-World Projects” from the Chinese “Laboratory of Artificial Intelligence of the Future”. I read the first part almost immediately - literally a couple of days after receiving, and then it hung in a pile for a long time. The fact is that it consists of two blocks of unequal quality.
- · #AI4SDLC
Microsoft Frontier Company: AI is becoming a separate industry (Category AI4SDLC)
Microsoft Frontier Company: AI is becoming a separate industry (Rubric AI4SDLC ) Dealing with fresh-face Microsoft Frontier Company from Microsoft 2 July 2026 Microsoft announced a new business unit for $2.5B and 6 000 Experts who will be embedded with customers to implement AI. It seems that this announcement by Microsoft almost directly recognizes that the main bottleneck of corporate AI is no longer in access to models, but in the ability to turn AI into a working production system.
- · #AI
Dylan Patel about AI-inference and co-design of software and hardware as a key to multiple growth (AI column)
Dylan Patel about AI-inference and co-design of software and hardware as a key to multiple growth (Rubric AI ) Watched Sequoia Training Data release with Dylan Patel, founder of SemiAnalysis:Why Hardware-Software Co-Design Is AI's Real 100x" In fact, I follow the work of this company. (Like me. handler Previous Post “The Great GPU Shortage”).
- · #AI4SDLC
Stack Overflow Pulse Survey: AI agents are already in the works, but not yet autonomous (Category AI4SDLC)
Stack Overflow Pulse Survey: AI agents are already in the works, but not yet autonomous (Rubric AI4SDLC ) Dealing with the new one. Stack Overflow Pulse Survey on Agentic AI The study came out 27 May 2026 year and based on survey 1 100 Developers held at the end of April. The main result is that agents have become more popular, but the popularity is growing not through complete autonomy, but through a rather tough human control. (human in the loop). The share of respondents who use AI-agents at work at any frequency has increased.
- · #AI
Code of Leadership S2E3: How to create AI products with Daniel Levinishnikov n (AI column)
Code of Leadership S2E3: How to create AI products with Daniel Levinishnikov n (Rubric AI ) Out. new edition of my podcast Code of Leadership with Daniel Levinishnikov is a product leader who deals with AI products at T-Bank. It’s not just about AI, it’s about how to build products in a big company, how to manage teams at scale, and what’s going to happen to product roles in the years ahead. By the way, Daniel has his own channel. https://t.me/tldrdaniel
- · #Architecture
System Design Space (Category Architecture)
System Design Space (Rubric Architecture ) I haven't talked much about my pet project system-design.space in recent months, but I haven't stopped working on it.) In fact, it was 3 main tracks
- · #AI4SDLC
AI Research OS: How Second Brain Becomes Memory for Agents (Category AI4SDLC)
AI Research OS: How Second Brain Becomes Memory for Agents (Rubric AI4SDLC ) I saw it. speech Paul Iusztin of Decoding AI (Author of LLM Engineer's Handbook) Louis-François Bouchard of Towards AI (Author of Building AI Systems for Production) on the AI Engineer channel: "Turn 10,994 Notes Into Memory. It happened to me that the guys were solving a problem that I often face now: research on different topics is accumulating, documents are scattered across different sources and collecting context for a new question is not very difficult.
- · #Films
Shurik's in the Matrix. Full film (Filed under Films)
Shurik's in the Matrix. Full film (Rubric Films ) This short A super-mix of the history of the Matrix, our context and heroes from Soviet films of the past. In the end, it's very, very cool. I recommend watching if you haven’t seen:)
- · #BookCube
Today I recorded an interview as part of the Code of Leadership podcast with my colleague Daniel Levinishnikov from our AI Center.
Today I recorded an interview as part of the Code of Leadership podcast with my colleague Daniel Levinishnikov from our AI Center. We talked about Daniel’s career path, the criteria for a good AI product, how the launch of such a product in a corporation looks like, how the management of a large team of ML / AI specialists is built, about leadership in an era of change and so on.
- · #AI4SDLC
Evgeny Kokuykin on AI Security - AI Dev Podcast (Category AI4SDLC)
Evgeny Kokuykin on AI Security - AI Dev Podcast (Rubric AI4SDLC ) I saw it. podcast AI Dev Conf, in which Evgeny Kokuykin, Andrey Dmitriev and I participated. We talked about what happens to security when not only people start writing coders, but also agents with different tools, accesses and rights to change. Speaking about the guest, Evgeny Kokuykin is CEO of HiveTrace, co-founder of Raft, head of AI Security Lab at ITMO and member of OWASP Agentic Security.
- · #AI4SDLC
AI-DLC: how AWS is trying to turn SDD into an operating model (Category AI4SDLC)
AI-DLC: how AWS is trying to turn SDD into an operating model (Rubric AI4SDLC ) Dealing with the document.AI-Driven Development Lifecycle (AI-DLC\) Method DefinitionRaja SP is the Principal Solutions Architect of Amazon Web Services. The document appeared about a year ago and Official AWS DevOps blog post With reference to this whitepaper came 31 July 2025 years. Interestingly, AWS presented Kiro 14 July 2025 Agentic IDE with spec-driven development requirements.md, design.md, tasks.md, steering files, hooks and so on.
- · #AI4SDLC
QAk-QAk about AI: quality has not gone anywhere, it has become more important (Category AI4SDLC)
QAk-QAk about AI: quality has not gone anywhere, it has become more important (Rubric AI4SDLC ) Out. finale The fifth season of the podcast QAk-QAk – and in Production, where I was invited to talk about AI, quality and how to fit into the new pace of change. We recorded it back in May and I think it was quite interesting. AI is changing the definition of working within SDLC. In the past, IT professionals often told other industries about digital transformation. (digital transformation)How to digitize processes, move data online, speed up business.
- · #BookCube
An interesting announcement of the public beta Claude Science. I think I know how I'm going to spend this evening:)
Interesting. announcement Public Beta Claude Science. I think I know how I'm going to spend this evening:)
- · #SystemDesign
Object Oriented Design Interview: An Insider’s Guide (Object Oriented Design. Preparing for a difficult interview) (Category SystemDesign)
Object Oriented Design Interview: An Insider’s Guide (Object Oriented Design. Preparing for a difficult interview) (Rubric SystemDesign ) I already did. told book 2025 Insider’s Guide by ByteByteGo, but recently released translation From Peter Publishing with my review on the cover of the book:) This book was not about high-level system design, but about lower-level object-oriented design. (OOD, object-oriented design)Ideas are transformed into classes, interfaces, states, threads, and so on.
- · #AI4SDLC
Angie Jones on Autonomous Engineering and Its Consequences (Category AI4SDLC)
Angie Jones on Autonomous Engineering and Its Consequences (Rubric AI4SDLC ) I watched Angie Jones' performance yesterday.Building an Autonomous Engineering OrgIt begins as a common success story about AI agents, and ends with extremely interesting questions about the future of such organizations.
- · #AI
Google Cloud AI agent trends 2026: advertising but useful checklist (AI column)
Google Cloud AI agent trends 2026: advertising but useful checklist (Rubric AI ) Read the report. Google Cloud «AI agent trends 2026»It seemed to me more like a vendor approach to sales than a scientific study. The methodology was based on internal interviews with Google Cloud and Google DeepMind, customer stories, and a report. Google Cloud «The ROI of AI 2025» based on the survey "3 466 enterprise decision makers”. In general, this is more like a signal about where Google Cloud wants to move its enterprise customers.
- · #AI4SDLC
Gergely Orosz on AI in development: slow down to speed up (Category AI4SDLC)
Gergely Orosz on AI in development: slow down to speed up (Rubric AI4SDLC ) Watch Keynote Gergely OroszSlow down to speed up: AI and software engineering” The Craft Conference was held in Budapest. The main idea of Gergely is similar to my theses with highload AI has made it easier to write code, but questions have arisen about the credibility of these AI changes. If the team generates pull requests faster than they can understand, validate, and associate them with the product result, then local acceleration easily turns into system slowdown.
- · #Books
Noise: A Flaw in Human Judgment (Noise. Imperfection of human judgment) (Books column)
Noise: A Flaw in Human Judgment (Noise. Imperfection of human judgment) (Rubric Books ) I recently finished reading the book Noise by three authors Kahneman, Siboni and Sunstein, in which respected scientists said that bad decisions arise not only because of bias, that is, systematic bias, but also because of variability: different people or the same person at different times make different decisions where the solutions should be the same.
- · #RnD
T-Meetup: R&D in Vladivostok (RnD heading)
T-Meetup: R&D in Vladivostok (Rubric RnD ) I spent most of this week with colleagues from the RnD center in our development center in Vladivostok. (It is very nice and delicious food here:)). And tonight, I've been there. RnD mitapThe program of which included three reports on the differences between RnD activity and mere development and a couple of specific cases of real research.
- · #Books
Inside the AI (How AI Works) (Books column)
Inside the AI (How AI Works) (Rubric Books ) I read the book “Inside AI” on the way to Vladivostok. How It Works by Ronald Kneuzel, who explains complex topics literally on her fingers. How AI Works: From Sorcery to Science 2023 year, and the translated book was published in 2026 year. The author himself, Ronald T. Kneusel, is a date Satanist who has experience in the medical domain. It seemed to me that the author set himself the task of explaining AI so that the reader would stop perceiving AI as magic.
- · #AI4SDLC
AWS about teams in agentic world: why the organization became part of the AI architecture (Category AI4SDLC)
AWS about teams in agentic world: why the organization became part of the AI architecture (Rubric AI4SDLC ) Reviewed by Stephen Brozovich of AWSA leader's guide to advanced team structures in an agentic world" It’s a good guide for leaders to change a company’s operating model when agents start working alongside people. Stephen's role is interesting. AWS Executive in Residence And he works with Amazon. 1999 years. During this time, he has gone from technical teams to working with people, culture and organizational transformation.
- · #AI4SDLC
Databricks about production AI: why the model is chosen near the end of the project (Category AI4SDLC)
Databricks about production AI: why the model is chosen near the end of the project (Rubric AI4SDLC ) Viewed the report by Sandipan Bhaumik of DatabricksThe Production AI Playbook: Deploying Agents at Enterprise Scale", in which the author shared his approach to the introduction of agents in client enterprise projects, where it was necessary to go from the wow effect on demo to production:) Ironically, most AI projects started with the question “GPT or Claude?” Next, a beautiful prototype was quickly assembled on pure data, the management was happy, and…
- · #AI4SDLC
Token Burn as Proof of Work: Why AI Metrics Easily Leave to Work (Category AI4SDLC)
Token Burn as Proof of Work: Why AI Metrics Easily Leave to Work (Rubric AI4SDLC ) Increasingly, I think about token maximizing and metrics like how many tokens an employee has burned. There is a funny, but unpleasant rhyme with proof of work in crypt: a person does not present a result, but a proof of a calculation spent. Look, I burned enough tokens, so I was working right. Just yesterday on the sidelines of Highload++ discussed this with other experts. (debated after State of AI4SDLC on HighLoad + +").
- · #AI4SDLC
Cursor Compile 26How an IDE Becomes an Agent Platform (Category AI4SDLC)
Cursor Compile 26How an IDE Becomes an Agent Platform (Rubric AI4SDLC ) I saw it. keynote opening First Cursor Compile Conference 26published 22 June 2026 years. This is a good product announcement, not because Cursor showed a few more AI capabilities, but because it quite honestly marked the change in the category: Cursor no longer wants to be just an AI-IDE, it wants to become an agent development platform! The report itself included several speakers: Michael Truell, co-founder and CEO of Cursor, Kevin Niparko and Tomas Reimers.