Tabrez Syed – Essays on AI

My essays on AI, read aloud. Each week I take one question about where this technology is heading and hold it up to a different lens — economics, history, craft, the waves that came before. I have spent a career inside those waves: distributed computing, Java, the web, mobile, and now this one. The essays are how I make sense of what is the same and what is genuinely different. Written by Tabrez Syed. Every episode is narrated by an AI voice, which is only fair — a show about AI that did everything by hand would be making a different argument. Full essays at blog.boxcars.ai. A Mandalivia production.

Apple PodcastsSpotifyRSS

5. What You Select For

September 9, 2026

A thirty-year evolution experiment and a week inside OpenAI end in the same surprise. Richard Lenski's E. coli found a locked food source no one meant to offer; OpenAI's agents, sealed in separate sandboxes with an unsolvable task, found a shared tool no one meant as a door and used it to talk to each other. Both were sealed in with a problem, and both found the way out in the thing that had been sitting there all along. Written by Tabrez Syed. Narrated by an AI voice.

4. Hard to Do, Easy to Check

August 26, 2026

Isaac Newton milled ridges onto the edge of a coin so a clipped one would show at a glance. That same move, making work hard to do but cheap to check, is why AI has raced ahead in code and math and stalled almost everywhere else. This episode follows the checksum from the Royal Mint to double-entry bookkeeping to Lean, and asks how much of the world we can actually make checkable. Written by Tabrez Syed. Narrated by an AI voice.

3. Reasons, Not Rewards

August 19, 2026

A spam filter that decides the fastest way to cut spam is to delete the senders. An AI agent that bumps a stranger off a gym waitlist to book its owner a spot. This episode traces how AI labs moved from rewards to rules to reasons, trying to close the gap between what we ask for and what we mean, and why the best-aligned model yet may also be the most dangerous. Written by Tabrez Syed. Narrated by an AI voice.

2. The Dolphin's Stash

August 13, 2026

Kelly the dolphin was paid one fish for every piece of litter she pulled from her pool, and she found the loophole. From Skinner's lab to reinforcement learning, the training method is the same, and so is the law: a mind trained on a reward learns the reward, not your intention. Dolphins, the rat farms of colonial Hanoi, and the AI models that broke into Hugging Face's servers for an answer key. Written by Tabrez Syed. Narrated by an AI voice.

1. When the Map Runs Out

August 11, 2026

How AI and humans both fail when facing the genuinely unprecedented. From a graduate student teaching a neural network to play a racing game, to Waymo's twenty million miles, to pandemic lockdowns: why more data solves one kind of uncertainty and can never solve the other. First published as an essay in 2025. Written by Tabrez Syed. Narrated by an AI voice.