Ivan Nosov

Writing

Everything I have published on rebuilding knowledge work with AI, newest first: 62 pieces so far. Posts react to a source or a week; long reads are the essays.

Engraving of an eighteenth-century print shop: three compositors at slanted type cases set lines of type by hand, printed sheets hang drying overhead.
Compositors at the type case. Encyclopédie, Imprimerie en lettres, Pl. 1, drawn by Goussier, engraved by Bénard; top scene of the plate. Musée Carnavalet, Paris Musées. CC0.

  1. No. 62

    The tasks worked. The job did not.

    Epoch AI tested six models on its own work. Where they failed is where most companies run on unwritten rules.

    Engraving of an eighteenth-century pin workshop: workers at wheels, benches and a forge, each doing one step of pin making, with the tools drawn in numbered figures below.
    Plate Épinglier, Pl. II, 1763. Credit

  2. No. 61

    OpenAI's 722 maths papers and the shape of jobs

    This week mathematics showed how the shape of many jobs will change, and it happened faster than anyone planned for.

    Business Insider article headline about an NYU professor saying OpenAI's release of maths papers wiped out early-career researchers' projects, with the byline and a quote from Tristan Buckmaster highlighted.
    Exhibit Business Insider, 10 October 2026. Source

  3. No. 60

    The Italian Hackathon League final at OGR Torino

    An unusual start to the day: the final of the Italian Hackathon League, at OGR Torino during Wave.

  4. No. 59

    FAIR Barcelona 2026: keynote recap

    Last Thursday I opened my speech at FAIR Barcelona with a question I have been solving for a while: what would it take to build JARVIS for HR?

  5. No. 58

    Everyone becomes a bit of an engineer

    Almost everyone who gets serious about AI eventually becomes a bit of an engineer.

    Editorial illustration: a wrench cut out of a printed page on an orange ground, headline AI makes you a bit of an engineer.
    Exhibit Visual from my LinkedIn post. Source

  6. No. 57

    Twenty supplier offers: where the change AI brings is hard to reach

    Fundamental change that AI brings might sound simple, yet it is hard to reach.

  7. No. 56

    The more I delegate to AI, the more distilled knowledge is worth

    The more work I delegate to AI, the more valuable the knowledge I have taken the time to distil becomes.

    One-page diagram titled Graph engineering: a process drawn as a network of nodes from frame and evidence check to storyboard, integration and release, with distilled knowledge feeding every node.
    Exhibit Visual from my LinkedIn post. Source

  8. No. 55

    Databricks as an AI workbench for knowledge work

    Databricks became an end-to-end AI workbench for knowledge work. More than 100 knowledge workers now use coding-class agents on it against challenges of HR, Finance, Commercial, and Marketing.

  9. No. 54

    OpenAI limits Cursor’s access to its models

    Now, this is Big. OpenAI is trying to curb Cursor's growth by limiting access to its models.

  10. No. 53

    A watermark on AI text only hurts if AI help is something to hide

    Anthropic turned on invisible text watermarking for Claude this month, a way to solve for EU AI Act requirement in force since August 2.

    Illustration of a fingerprint under a red scanning beam, with the words subject identified and critical.
    Exhibit Visual from my LinkedIn post. Source

  11. No. 52

    Every correction to an AI answer is a process rule you paid for

    You know the ritual. The first AI's output is too long. The wrong data was used.

    Carousel cover titled Every correction is process documentation, with counts of minutes of correction, messages and saved instructions.
    Exhibit Visual from my LinkedIn post. Source

  12. No. 51

    Microsoft's biggest obstacle to AI ROI may be Microsoft Office

    Copilot entered the world’s dominant productivity suite with a distribution advantage nobody else could replicate.

  13. No. 50

    Digital transformation now sits in every leadership job

    Digital transformation now sits in every leadership job, whether the job description admits it or not.

  14. No. 49

    Model rankings shift; my workflows stay provider-neutral

    ChatGPT 5.6 is fantastic, Grok 4.5 is the working horse. In tokenomics time - Antropic loses the edge.

  15. No. 48

    Start with the output and let it pull the data

    The dominant narrative in the feed repeats the same rigid prerequisite: you cannot adopt AI until your data is perfectly clean.

    Quote card: You do not find the leaks by inspecting dry pipes in the dark. You turn the pump on, apply pressure, and watch where it breaks. Do we actually need clean data to adopt the AI?
    Exhibit Visual from my LinkedIn post. Source

  16. No. 47

    When processing time collapses, the bottleneck moves to thinking

    Giac Soliman had me on the Range podcast on AI adoption in comp and HR - with Pay Equity as center showcase.

  17. No. 46

    10% × 10% × 10%: layered checks that make AI errors rare

    10% x 10% x 10% = 0.1% error rate. That is the promise behind stable AI work.

  18. No. 45

    What 40+ AI reports from March to June 2026 agree on

    I ran an analysis on 40+ AI reports and papers released March-June 2026.

  19. No. 44

    Two weeks testing Genie Code as a governed AI workspace

    I spent two weeks testing Genie Code for the job Databricks undersells.

  20. No. 43

    Personal context is the only durable asset in AI

    Consistent maintenance of your personal context is the only durable asset in the AI space.

  21. No. 42

    AI adoption is loud, but much of it still feels like extra work

    AI adoption is finally loud, but a lot of it still feels like extra work, not extra leverage, because most of us are paying a rework tax on generic output, then pretending that is the future of work.

  22. No. 40

    Anthropic releases Claude Opus 4.7

    Three times the vision resolution of 4.6. More literal instruction following, which means existing prompts may need adjustment.

  23. No. 39

    Two Chinese AI labs as an experiment in organisational design

    Two Chinese AI labs are running a live experiment on organizational design.

  24. No. 38

    Anthropic passes OpenAI in annualised revenue

    Anthropic just passed OpenAI in annualized revenue. $30B vs $25B ARR as of April 2026.

    Line chart of annualised run-rate revenue, 2023 to 2026: Anthropic crosses OpenAI in April 2026, 30 billion dollars against 25 billion.
    Exhibit Visual from my LinkedIn post. Source

  25. No. 37

    The play most companies run with AI: hand it to IT

    Most companies I talk to are running the same play with AI - hand it to IT, set adoption KPIs, measure how many people opened Copilot this week.

  26. No. 36

    Distillation, licences and trust in the AI model ecosystem

    The AI model ecosystem right now. Moonshot AI used 24,000 fake accounts and 3.4 million interactions to distill capabilities from Claude into their Kimi model.

  27. No. 35

    AI and the Solow paradox

    You can see the computer age everywhere but in the productivity statistics" Robert Solow wrote that in 1987.

  28. No. 34

    When leaders build with AI themselves

    The AI adoption that actually stuck around me started with nobody's permission and no training program.

    Diagram titled Two Paths to AI Strategy, comparing a leader who delegates AI with a leader who builds with it.
    Exhibit Visual from my LinkedIn post. Source

  29. No. 33

    Microsoft Copilot Cowork at $99 a month

    Microsoft just launched Copilot Cowork bundled in at $99/month - AI product built on someone else's engine - then removed the three features that make it work.

    Carousel cover titled Copilot Cowork, The AI Masquerade: Microsoft licensed Claude’s engine and shipped it without the capabilities that make it work.
    Exhibit Visual from my LinkedIn post. Source

  30. No. 32

    Software engineering hiring and the Jevons paradox

    Software engineer job postings are up 11% year-over-year in the most AI-automated profession on the planet - a textbook case of Jevons Paradox, where making a resource cheaper to use increases total...

  31. No. 30

    Most AI training programmes teach people to type

    Most AI training programs are teaching people to type. The capability gap is in what they should be writing.

  32. No. 29

    Anthropic’s architecture for running HR operations through an LLM

    Anthropic just shipped a clean architecture for running HR operations through an LLM.

  33. No. 28

    Years of Excel skill, quietly made optional

    Spent years getting really good at Excel. Like, genuinely good - it was one of the things that set me apart early in career.

  34. No. 27

    HBR: AI does not reduce work, it intensifies it

    HBR just published a warning: "AI Doesn't Reduce Work — It Intensifies It" I read it as a confirmation that it's working.

    Chart titled The Cognitive Compounding Effect: organisational capability against knowledge democratisation, with team compounding rising fastest.
    Exhibit Visual from my LinkedIn post. Source

  35. No. 26

    A 200-line markdown file and $285 billion off the software market

    Anthropic published an open-source plugin for Claude - structured prompts, first-year law school content, some workflow logic.

    Collage of news headlines on the software stock selloff after Anthropic released a Claude plugin, from Fortune, Reuters, CNN and TheStreet.
    Exhibit Visual from my LinkedIn post. Source

  36. No. 25

    Software developers rewired how they work in under two years

    HR hasn't even started. HR sits on the most AI-compatible raw material in the company - policies, frameworks, guidelines, job architectures, regulatory docs.

  37. No. 24

    Knowledge Compounds

    What happened after we launched a company-wide AI agent Launched our first AI agent inside the performance management system.

  38. No. 23

    Claude Code is for you, not only for coders

    Not just for Coders. The most common question: "This is great, but how do I actually set it up?

    Carousel cover: Claude Code is for you. You do not need to code to use it; setup takes 15 minutes.
    Exhibit Visual from my LinkedIn post. Source

  39. No. 22

    Pay equity analysis in two hours instead of €50K

    Consultants charge €50K for pay equity analysis. I built it with AI in 2 hours.

    Carousel cover: I built a €50K framework in 2 hours, on EU Pay Transparency Directive compliance with Claude Code.
    Exhibit Visual from my LinkedIn post. Source

  40. No. 21

    From Late-Night Deck Crunching to Strategic Thinking

    How Claude Code transformed my workflow from late-night deck slavery to strategic thinking Large part of my job involves creating presentations, documents, and compelling stories. of time spent on...

    Carousel cover: I quit PowerPoint slavery. How Claude Code in VS Code turned me from midnight deck-cruncher to strategic thinker.
    Exhibit Visual from my LinkedIn post. Source

  41. No. 20

    2025: the year HR started coding

    2025: The Year HR Started Coding Birds aren't supposed to swim. HR people aren't supposed to build.

  42. No. 19

    Employees are assets: true or false?

    Had a moment of surprising freeze during an Accounting Basics review today.

  43. No. 18

    Karpathy’s march of nines

    One concept from Andrej Karpathy's latest interview that I take as a large takeaway is The "March of Nines".

  44. No. 17

    Your AI just created more work and called it productivity

    Congratulations! Your AI Just Created More Work (And Called It Productivity) Let's talk about the elephant in the server room: we're drowning in AI-generated nonsense, and we're calling it innovation.

  45. No. 16

    Rumelt’s Good Strategy/Bad Strategy and the Nvidia example

    Years before the AI gold rush made it a household name, Richard Rumelt’s classic, "Good Strategy/Bad Strategy," pinpointed Nvidia as a textbook example of brilliant strategy.

  46. No. 15

    The barrier to AI transformation is not technology or budget

    It's imagination and desire. I recently spent a bit of time building something that helps to solve initial steps in Skills Management complexity.

  47. No. 14

    What sits behind “AI layoffs”

    The phrase "AI Layoffs" sounds terrifying, doesn't it? When Salesforce CEO Marc Benioff recently said he cut 4,000 jobs because he "needed less heads", it made me wonder what's really going on behind...

  48. No. 13

    Is Google the dark horse of generative AI?

    I've been exploring their latest models and had a shot at a number of deep researches with Gemini 2.5Pro, and their video generation Veo.

  49. No. 12

    Retrieval as an answer to the knowledge crisis

    RAG AI: The Savior for Knowledge Crisis Retrieval-Augmented Generation (RAG) brings one Simple Promises: Transform your existing knowledge into your favorite AI's thoughts.

  50. No. 11

    The 95% AI failure rate, and making the 5% count

    I'll Make Those 5% Count and Scale It Forward Fresh MIT NANDA research dropped in July with an uncomfortable truth: 95% of business GenAI initiatives deliver zero ROI.

  51. No. 10

    ChatGPT 5: why the launch fell flat

    ChatGPT 5: The Hype Fell Flat - Here’s Why Is ChatGPT 5 the game-changer we were sold, or just a shiny letdown?

  52. No. 9

    Fifteen minutes to build what HR has been overthinking for years

    15 Minutes to Build as HR What You've Been Overthinking for Years I've heard countless variations of "I wish we had a tool that could..." followed by resigned acceptance that IT priorities don't...

  53. No. 8

    Three and a half charts for a total rewards career

    The 3.5 Charts That Will Save Your Total Rewards Career (And Your Sanity) After years of drowning in compensation analytics, I've made a simple discovery: 90% of our questions can be answered with...

  54. No. 7

    Why centres of excellence keep failing

    Centers of Excellence solutions consistently fail - why is that? I think I found some ground for the approach I've been trying to implement with MIT's help.

  55. No. 6

    Why the average employee can never get average rewards

    The Zero-Sum Confession: Why Your "Average" Employee Can Never Get Average Rewards There's a moment in every merit cycle when the math stops being friendly.

  56. No. 5

    The most useful tool I have found for decision making

    The most powerful tool I've discovered for improving decision making? Transparency.

  57. No. 4

    Nobody actually wants data

    They want permission with a spreadsheet attached. Continuing my 15-year compensation confessions...

  58. No. 3

    Nobody plans to become a compensation professional

    It's something that happens to you, like getting caught in the rain or developing strong opinions about Excel formulas (the look of horror when seeing VLOOKUP instead of XLOOKUP in 2025... 😱).