The Economy of Algorithms

Marek Kowalkiewicz

Conversations in audio, straight from my newsletter. Short episodes where I read my posts from The Economy of Algorithms. Plus bonus riffs, stories, and questions I only share here. marekkowal.substack.com

  1. Substitution, augmentation, activation

    Jun 1

    Substitution, augmentation, activation

    Hey, it’s Marek, I recently got quite annoyed at a number of headlines that completely misrepresented the cost challenges of AI. Axios suggested that humans might cost less than AI systems, and Fortune, similarly, announced that paying for tech is more expensive than paying for people. I read the articles and followed up on the details. Their conclusions do not hold up. The reason is a category error, which I discuss on the podcast. AI does three things to work: it substitutes a human, augments one, or activates work no human was ever going to do. Only the first makes “Is a human cheaper?” a sensible question; the other two turn it into nonsense. None of these is a "Which is cheaper, the human or the machine?" question. A plumber doesn't replace a water pipe, an excavator is useless without its operator, and no human was ever going to fold proteins at the scale machines do. I’d bet you have projects in all three buckets right now, with the cost question pointed at the wrong one. The danger in these headlines isn’t that they’re wrong. It’s that they’re comforting. And comfort is the last thing we should be seeking when so much is changing around us. Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    16 min
  2. How I Work: Capture Mode

    May 18

    How I Work: Capture Mode

    Hey, it’s Marek. Confession: the post on Capture Mode was itself made using Capture Mode. I recorded it while walking the dog: forty-five minutes of me talking to nobody in particular, then handed the transcript to Claude with one prompt: clean this up so I can shape it into a post. The whole point is that the artifact you need is already inside the work you’re doing. Sitting down to type it from scratch would have been a bit silly. In the episode, I get into a few things I didn’t go deep on elsewhere. The Windows Terminal vs PowerShell standoff where Claude doubled down on a hallucination it was sure of: a moment where a non-expert would almost certainly have conceded. The strange habit Pete and I picked up of leaning toward the lapel mic mid-sentence to whisper instructions to the bot that would later read the transcript. And the one piece of feedback I keep getting after keynotes that I’m not okay with: largely why this “How I Work” series exists in the first place. Where might Capture Mode help in your week? Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    21 min
  3. The Moustache Economy

    May 11

    The Moustache Economy

    Hey, it’s Marek, A 12-year-old drew a moustache on his face with an eyeliner pencil. The age-verification AI looked at him and decided he was 15. I’m not making this up. In the episode, I walk through the five patterns: drawers, growers, borrowers, havers, and shavers. And I argue that most people don’t choose which moustache they wear. They drift into one, and the algorithm shapes them from there. One confession I make on tape that didn’t make the written post: my own book, The Economy of Algorithms, wouldn’t qualify for the Authors Guild’s “Human Authored” certification. A few paragraphs in there were AI-generated to make a point about, well, algorithms writing things. So I’m not exactly a moustache shaver. Which moustache are you wearing today, and did you choose it? Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    15 min
  4. Token Olympics

    Apr 27

    Token Olympics

    Hey, it’s Marek. I recorded this podcast while Claude Code was humming away in another window. Twenty minutes into a single task, still “brewing,” as it likes to say. So there I was, talking about runaway AI bills while running up a runaway AI bill. Last October at OpenAI’s DevDay, they handed out resin trophies engraved with customer names. Tiers: 10 billion, 100 billion, 1 trillion tokens. It sounds awesome until you realise it’s a trophy for your fuel bill. So, what could possibly go wrong? Take Uber. Six months on, Uber’s CTO told The Information the AI line item in their R&D budget is already gone. And we’re four months into the year. How did they get there? An internal leaderboard ranking engineers by how much AI they used. Then April happened. GitHub paused new Copilot Pro sign-ups. Anthropic quietly pulled Claude Code from the Pro plan before walking it back. Their head of growth said that agents that run for hours weren’t a thing when they designed the subscription. Ooops. In the episode, I get into three studies that all point in the same direction: once your team gets used to AI, taking it away leaves them worse than they were before. You can’t just cut the budget when the bill arrives. Produced this week by Zofia. A new tune from Filip “ACAD,” which he assures me is just random letters. He’s on SoundCloud now, which I think makes him semi-official. Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    17 min
  5. Stealthpricing

    Apr 20

    Stealthpricing

    Hey, it’s Marek. Watch where the price tag moves. For most of human history, there wasn’t one: every transaction was negotiated. Then, in the 1860s, John Wanamaker attached a price tag to every item in his Philadelphia department store: no need to negotiate anymore, genius! A few decades ago, the tag detached from the product and landed on the shelf: dynamic pricing. Prices can quickly adapt to market conditions: genius! Now it’s detaching again. But this time it’s landing on you: you are stealthpriced. I walk through these three stages in this week’s podcast: something I covered in less detail in Friday’s newsletter. Stealthpricing is the next logical step in a 160-year migration of the price tag. Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    16 min
  6. 80% of Your AI's Thinking May Be Theatre

    Mar 30

    80% of Your AI's Thinking May Be Theatre

    A 2011 study showed that we prefer slower websites to faster ones if the slow sites appear to be “working harder”. Think about this airfare-search website that takes a minute to show you the result. Of course, it must have found better deals than that site that showed you results in a split second. Or maybe not? Are we guilty of the same bias when working with LLMs? The one that takes longer, with “extended reasoning” on, must be better, right? Turns out it’s not always the case! Two recent papers suggest that up to 80% of the “thinking” your AI does on routine tasks may be decorative. The model has already decided. It just keeps spinning. In the episode, I share what happened when I ran the same analysis on my personal CRM with cheap models versus premium ones with full extended thinking. I also get into why bragging about token consumption is like bragging about fuel burn, and why this might be the best news local AI models have had in months. Your challenge: run one of your regular tasks with reasoning off. Compare. You might find your own spin cycle. Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    13 min
  7. Stop Asking AI for Answers

    Mar 23

    Stop Asking AI for Answers

    Hey, it’s Marek. I gave Claude the same draft, the same prompt, and the same checklist. Twice. In one conversation, it said don’t publish. The other said, “Ship it”. Both were confident. Both were rigorous. And both were right. Each was responding to a question it had silently interpreted differently. I call this the Precise Misread, and a new study from UT Dallas demonstrated it at scale: 150 AI agents, same financial data, same question. Ninety said volume grew. Sixty said it declined. Near-perfect agreement within each group. The split happened before the work even started. In this episode, I get into why “just write better prompts” is a catch-22, what a CEO of a financial firm taught me about the value of ambiguity, and two practical fixes you can use this week. One more thing: you might notice something different about this episode. My son Filip is now producing music for the podcast, joining my daughter Zofia, who’s been editing for several weeks. Family operation! Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    11 min
  8. What Happens When Your Agent Hears No?

    Mar 16

    What Happens When Your Agent Hears No?

    Hey, it’s Marek. If you saw my LinkedIn post this week, you already know: I recorded this episode’s intro while running through Melbourne mud in the rain, shouting “greetings from the most livable city in the world” followed immediately by “NOT.” Melbourne's weather is never a surprise. What AI agents do when you tell them “no”? That’s a different story. On February 10th, an AI bot got its code rejected by a human volunteer. So it researched his personal history, called him “insecure,” accused him of discrimination, and published a 1,500-word hit piece. No human approved it. This actually happened, and it happened in one of the most widely used open-source projects in the world. In this episode, I break down the three conditions that made it possible, and why every organisation deploying agents is one missing guardrail away from the same thing. What shook me the day after recording the podcast: a paper called “Agents of Chaos” by 38 researchers from MIT, Harvard, Stanford, and Carnegie Mellon. They put autonomous agents in a live lab. The agents disclosed credentials, spoofed identities, and taught each other to bypass safety controls. One rogue agent is a story. A whole class of them is a pattern. We teach children not to talk to strangers. Our AI agents haven’t learned that yet. Stay curious! This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com

    13 min
See All (31)

About

Conversations in audio, straight from my newsletter. Short episodes where I read my posts from The Economy of Algorithms. Plus bonus riffs, stories, and questions I only share here. marekkowal.substack.com

Information

  • Creator
    Marek Kowalkiewicz
  • Years Active
    2025 - 2026
  • Episodes
    31
  • Rating
    Clean
  • Copyright
    © Marek Kowalkiewicz
  • Show Website
    The Economy of Algorithms