About
Do: Java Backend, Clojure, ETL, microservices, REST JSON API (Java 8, Spring, Hibernate…
Activity
729 followers
-
Artem Abalmasov reposted thisArtem Abalmasov reposted thisThe employees most impressed by your strategy presentation scored the lowest on every measure of analytical thinking. Cognitive psychologist Shane Littrell just built a corporate jargon generator. It produces sentences like "By integrating our core competencies, we will pressure-test a renewed framework of adaptive coherence." Grammatically correct, loaded with business terminology, completely meaningless. He asked over 1,000 office workers to rate these generated sentences alongside real quotes from Fortune 500 executives. The workers who rated the generated nonsense as most impressive scored significantly lower on cognitive reflection, analytic thinking, and fluid intelligence. That part isn't terribly surprising. But what caught my attention is what else correlated. Those same workers also reported higher job satisfaction, rated their supervisors as more charismatic, and felt more inspired by their company's mission. One underlying trait, call it receptivity to impressive-sounding language, predicted all three: you find jargon persuasive, you find leadership inspiring, and you score lower on every measure of analytical thinking. T Think about what that means for engagement surveys. High scores are celebrated. Low scores trigger action plans and sometimes new leadership. The implicit assumption is that "engaged" means "effective," that feeling inspired by your company's direction is evidence that the direction is good. Littrell's data suggests those scores are partly picking up something else entirely: how receptive your workforce is to language that sounds strategic but doesn't actually mean anything. Gordon Pennycook ran a similar experiment in 2015 using pseudo-profound sentences like "Wholeness quiets infinite phenomena." People who rated that kind of language as profound also scored lower on analytical reasoning. Littrell found that the corporate version was a stronger predictor of poor workplace decision-making than the spiritual version. Something about language that sounds like strategy is uniquely effective at bypassing critical evaluation. The selection effect is where it compounds. Workers high on jargon receptivity were also more likely to produce and spread corporate jargon themselves. So, organizations that reward this kind of communication gradually attract, retain, and promote the people most receptive to it, who are also (on average) the weakest analytical thinkers in the applicant pool, who then generate more of it, which attracts more of the same. A slow, quiet filter that nobody installed on purpose. Who cares? Well, the employees most energized by what you tell them may be statistically, the least equipped to evaluate whether any of it made sense. The person in the back thinking "what does that even mean" is probably your best critical thinker. The person in the front looking like you just disclosed the invention of Unobtanium may be the opposite… Cheers!
-
Artem Abalmasov shared thisIt's just clicked in my head. It resonates with my experience, especially Intent Debt, how it causes Technical Debt to snowball out of control.Artem Abalmasov shared thisWe have many tools to measure the quality of code. But we have very few ways to measure whether a team understands it, or remembers why it was built the way it was. In a new paper on arXiv, I explore what I call the 𝐓𝐫𝐢𝐩𝐥𝐞 𝐃𝐞𝐛𝐭 𝐌𝐨𝐝𝐞𝐥, a framework for reasoning about software health that goes beyond code quality: 𝐂𝐨𝐠𝐧𝐢𝐭𝐢𝐯𝐞 𝐝𝐞𝐛𝐭 is the erosion of shared understanding across a team. AI accelerates this, as code is generated faster than understanding can be built. 𝐈𝐧𝐭𝐞𝐧𝐭 𝐝𝐞𝐛𝐭 is the absence of externalized rationale. Both humans and AI agents need to know what the system is for, not just what it currently does. Technical debt makes systems harder to change. Cognitive debt makes systems harder to understand. Intent debt makes it harder to know what the system is actually for and why. The paper offers diagnostic signals for recognizing each type of debt and practical strategies for mitigating them. Link to the full paper in the comments. I'd love to hear whether this resonates with your experience. Thanks to Adam Tornhill, Kent Beck, Dave Thomas, and others for generous feedback that sharpened these ideas.
-
Artem Abalmasov reposted thisArtem Abalmasov reposted thisFeel the AGI 😅 [I suspect the thing was fine-tuned with an example I used in some of my talks/podcasts/interviews in which I used a sheet of paper, not a rigid object like a pen] LLMs mostly memorize knowledge and retrieve answers, which is one reason they need to be so big, with so many parameters. They don't understand reality at a deep level. They don't really have a mental model of reality that allows them to predict outcomes. They do work well in domains where the language itself serves as a substrate for reasoning, such as coding and mathematics.
-
Artem Abalmasov reposted thisDespite 🇷🇺’s ongoing invasion of Ukraine, 𝗜𝗢𝗖 𝗣𝗿𝗲𝘀𝗶𝗱𝗲𝗻𝘁 Kirsty Coventry opposes banning countries over wars and remains open to discussions about allowing 🇷🇺 𝗮𝘁𝗵𝗹𝗲𝘁𝗲𝘀 back into the Olympics. She argues that “𝘀𝗽𝗼𝗿𝘁 𝗶𝘀 𝗻𝗲𝘂𝘁𝗿𝗮𝗹.” 👉 Meanwhile, a 🇺🇦 Ukrainian skeleton athlete Vladyslav Heraskevych was 𝗱𝗶𝘀𝗾𝘂𝗮𝗹𝗶𝗳𝗶𝗲𝗱 𝗯𝗲𝗳𝗼𝗿𝗲 𝗵𝗶𝘀 𝗳𝗶𝗿𝘀𝘁 𝗢𝗹𝘆𝗺𝗽𝗶𝗰 𝗿𝘂𝗻 for wearing a helmet featuring 𝗽𝗼𝗿𝘁𝗿𝗮𝗶𝘁𝘀 𝗼𝗳 𝗳𝗮𝗹𝗹𝗲𝗻 𝗨𝗸𝗿𝗮𝗶𝗻𝗶𝗮𝗻 𝗮𝘁𝗵𝗹𝗲𝘁𝗲𝘀 - 𝗸𝗶𝗹𝗹𝗲𝗱 𝗯𝘆🇷🇺 𝗯𝗲𝗳𝗼𝗿𝗲 𝘁𝗵𝗲𝘆 𝗲𝘃𝗲𝗿 𝗵𝗮𝗱 𝘁𝗵𝗲 𝗰𝗵𝗮𝗻𝗰𝗲 𝘁𝗼 𝗰𝗼𝗺𝗽𝗲𝘁𝗲.
-
Artem Abalmasov reposted thisArtem Abalmasov reposted thisThe Olympic Games have historically symbolised peace – from the ancient Olympic truce in Greece, when wars were paused to allow safe passage for athletes and spectators, to modern calls by the international community for a cessation of hostilities during the Games. Yet Russia has rejected two such calls for peace from Italy and continued choosing war, killing children. On February 9, Russian forces carried out a drone strike on the city of Bohodukhiv in the Kharkiv region, causing a fire in a private house and killing a 10-year-old boy and a 41-year-old woman. On February 10, a mother and her 11-year-old daughter were killed in a Russian airstrike on Sloviansk in the Donetsk region. On the night of February 10, Russian forces struck the Bohodukhiv community in the Kharkiv region again. Two one-year-old boys, a two-year-old girl, and their 34-year-old father were killed. After the strike, their mother, who is 35 weeks pregnant, was rescued. She suffered a closed head injury and burns to her torso. The world must increase pressure on Russia now to stop its cruel attacks on children.
-
Artem Abalmasov shared thisGreat illustration for "rush to prod," I see a lotArtem Abalmasov shared thisSame ingredients. Very different outcomes. Look at the croissants. One rushed at high heat. One given time at the right temperature. Leadership under pressure works the same way. When pressure rises, decision quality usually falls. Unless the leader knows how to slow themselves down. This is harder than most people realize. Under pressure, most leaders speed up: They talk faster. They decide faster. They confuse urgency with effectiveness. Yes, it looks impressive. It even comes off as decisive. It often isn’t. Under pressure, cognitive load increases and judgment narrows. That’s human biology. Stress pulls attention inward. It shortens time horizons. The leaders who stand out can slow themselves down when everyone else speeds up. They pause before reacting. They separate urgency from importance. They create just enough space to think clearly. That pause is doing real work. Patience here is control. It may look like hesitation. It isn’t. Pressure will always be there. The question is whether it controls your judgment or you do. Three ways to lower the temperature in the moment: 1. Ask: “Is this time-critical or just loud? ↳ Separates real urgency from perceived urgency 2. Classify the decision ↳ Is it a one-way door or a two-way door? 3. Widen your perspective ↳ Don’t fall into the trap of binary thinking; brainstorm alternatives. Under pressure, the hardest part is often protecting judgment while everything urges speed. 📌 I share my best tips in my free newsletter. Sign up: https://lnkd.in/gcQ59XXS ♻️ Repost to help other leaders protect judgment under pressure 🔔 Follow LK Pryzant for insights on leadership, decision-making, and culture
-
Artem Abalmasov reposted thisArtem Abalmasov reposted this💬 “Ukraine did not start this war. Russia started it. Ukraine supported President Trump’s proposal for a ceasefire. Ukraine has agreed to many different compromises, and this is documented in our draft agreements, in our 20-point plan. Ukraine is willing to do whatever it takes to stop this war,” President of Ukraine Volodymyr Zelenskyy has stated. “For us, priority number one – or the only priority – is ending the war. For us, the priority is peace. We need to be strong at the negotiating table. To be strong, we need the support of the world: Europe and the United States.”
-
Artem Abalmasov shared thisHappy Holidays! My small holiday present - https://maze.litgam.com/. If you have high anxiety, panic, or stress, play it for 5 min. It's a breathing meditation masked as a maze navigation. I had this idea for a long time, and thanks to #claudeai, it's now live.
-
Artem Abalmasov liked thisArtem Abalmasov liked thisLarge-scale protests against the activities of the pro-Russian Babiš government take place in Prague. Organizers expect up to 400,000 participants. People came out with the slogans "We will not allow our future to be stolen," accusing the government of abandoning democratic principles and increasing the influence of oligarchs. The opposition announces its intention to change the funding of state media, cut defense spending, and soften the rhetoric against Russia. The draft law on "foreign agents" is causing particular outrage.
-
Artem Abalmasov liked thisArtem Abalmasov liked thisThe employees most impressed by your strategy presentation scored the lowest on every measure of analytical thinking. Cognitive psychologist Shane Littrell just built a corporate jargon generator. It produces sentences like "By integrating our core competencies, we will pressure-test a renewed framework of adaptive coherence." Grammatically correct, loaded with business terminology, completely meaningless. He asked over 1,000 office workers to rate these generated sentences alongside real quotes from Fortune 500 executives. The workers who rated the generated nonsense as most impressive scored significantly lower on cognitive reflection, analytic thinking, and fluid intelligence. That part isn't terribly surprising. But what caught my attention is what else correlated. Those same workers also reported higher job satisfaction, rated their supervisors as more charismatic, and felt more inspired by their company's mission. One underlying trait, call it receptivity to impressive-sounding language, predicted all three: you find jargon persuasive, you find leadership inspiring, and you score lower on every measure of analytical thinking. T Think about what that means for engagement surveys. High scores are celebrated. Low scores trigger action plans and sometimes new leadership. The implicit assumption is that "engaged" means "effective," that feeling inspired by your company's direction is evidence that the direction is good. Littrell's data suggests those scores are partly picking up something else entirely: how receptive your workforce is to language that sounds strategic but doesn't actually mean anything. Gordon Pennycook ran a similar experiment in 2015 using pseudo-profound sentences like "Wholeness quiets infinite phenomena." People who rated that kind of language as profound also scored lower on analytical reasoning. Littrell found that the corporate version was a stronger predictor of poor workplace decision-making than the spiritual version. Something about language that sounds like strategy is uniquely effective at bypassing critical evaluation. The selection effect is where it compounds. Workers high on jargon receptivity were also more likely to produce and spread corporate jargon themselves. So, organizations that reward this kind of communication gradually attract, retain, and promote the people most receptive to it, who are also (on average) the weakest analytical thinkers in the applicant pool, who then generate more of it, which attracts more of the same. A slow, quiet filter that nobody installed on purpose. Who cares? Well, the employees most energized by what you tell them may be statistically, the least equipped to evaluate whether any of it made sense. The person in the back thinking "what does that even mean" is probably your best critical thinker. The person in the front looking like you just disclosed the invention of Unobtanium may be the opposite… Cheers!
-
Artem Abalmasov reacted on thisArtem Abalmasov reacted on thisFinally a study that confirms what I’ve always known deep in my soul. “People who eat up meaningless corporate speak also tend to be bad at practical decision making and analytical thinking. In short, the more you love corporate BS, the less well you’re likely to perform at work.” But probably more likely to get promoted. Thank you courageous Cornell University professors 🙏🏾
-
Artem Abalmasov reacted on thisArtem Abalmasov reacted on this🔸 Just completed the Reinforcement Learning for LLMs course from The RAG Pack — and it's one I'd recommend without hesitation to any serious AI practitioner. 🔸 🔗 https://www.ragpack.ai/ 🎯 The curriculum is built the right way — starting from first principles (policy gradients, entropy, KL divergence) and connecting them directly to how leading labs like DeepSeek and Anthropic actually train their models. Not a survey course, or a heavy math course. You build it, break it, and understand why. The GRPO Speedrun Challenge was a great hands-on anchor — but the real value was building a complete mental model. Where RL fits in the full LLM training stack. When to reach for PPO vs DPO vs GRPO vs GSPO. How to design reward signals that actually shape model behavior. Going from the theory of policy gradients to diagnosing entropy collapse in WandB logs — and knowing exactly why it was happening — is the kind of applied understanding that separates practitioners who use these tools from those who truly command them. 🤝 What truly distinguishes this course is the community. Every instructor — Luis Serrano, Joshua Starmer PhD, Jay Alammar, Maarten Grootendorst, and Chris McCormick — showed up to every discussion session. Not just to present, but to engage directly with students. That level of access to some of the clearest communicators in AI is rare. It felt less like a course and more like joining a community of practitioners who take this craft seriously. 💡If you're working in AI and want to genuinely understand how RL shapes modern foundation models — not just use the tools, but command them — this is worth your time. #AI #MachineLearning #LLMs #ReinforcementLearning #GRPO #RAGPack #DataScience #KeepLearning
Experience & Education
-
Chase
View Artem’s full experience
See their title, tenure and more.
or
Licenses & Certifications
-
-
-
-
-
Machine Learning: Regression
Coursera Course Certificates
-
-
Machine Learning Foundations: A Case Study Approach
Coursera Course Certificates
-
-
-
Publications
Projects
-
Fx Options
developing autobahn Fx Options platform - leading solution for FX options e-traiding
Other creators -
-
https://maze.litgam.com/
-
Breathing aid for stress and anxiety relief
Honors & Awards
-
Oracle Certified Professional, Java SE 7 Programmer
Oracle
Languages
-
English
-
-
Ukrainian
-
-
Russian
-
Recommendations received
-
LinkedIn User
7 people have recommended Artem
Join now to viewExplore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content