Skip to content
You Have Died from Excessive Chatbotting episode artwork

Aug 10, 2026 46 min

You Have Died from Excessive Chatbotting

from Unqualified Advice Hosted by PFL LLC

The Mirror and the Force Hello dear show notes readers! Substack spent the week scanning everyone's prose for robot fingerprints with Pangram, the AI detector now built into the platform. We discuss the uncomfortable truth that a word-order detector can tell you how something was written but not if idea is any good. Dan's warning: the incentives now point toward machine versus machine, writers slopping up their prose so other robots can't tell a robot helped, and readers throwing out very good essays on principle. AI should make you think harder, not think for you. I walk through the agent I talk to on my commute, Number One, and Dan counters with his council of five advisors that occasionally blows up an essay he was sure about. We talk about when companies skip the thinking part entirely: congratulations, you just hired one million terrible employees. The back half gets personal, as we ask what happens when the machine reads you? Personality tests, a suspiciously accurate inference of my IPIP scores, an app that accused Dan of camouflaging, and the mirror problem: tune this thing to your feedback carelessly and you've built a digital echo chamber with a population of one. Dan closes by comparing the whole technology to the Force, and honestly, it holds up better than it should. Use the force and go build something this week. Cheers, Sean Tools & Platforms Mentioned Pangram — the AI-detection tool Substack now runs Substack — and its new AI detection feature Hemingway Editor — the sharp-and-short editor we both swear by (and I swore to return to) Claude — the app, the audio feature, and the Fable model that guessed my personality scores Deep Personality — Andrew Wilkinson's hour-long assessment app GitHub — home of the Bridge repo and agent "number one" Firebase / Vercel — the "two years ago I couldn't have done this" website stack The Oregon Trail — you have died of dysentery (2026 remix) Companies Discussed Anthropic OpenAI Substack Pangram Adobe Dell IBM MrBeast (the playbook) Links & References Substack bets readers want to pay for human-written content (Axios, Jul 23, 2026) — the Pangram rollout that kicked off the episode Third-party evaluations of Pangram (Pangram Labs) — the studies behind "cream of the crop" The Prometheus Dispatch — Dan's Substack, where this week's machine-versus-machine essay lives MrBeast's 100-videos advice (X, 2020) — the bar we're about fifty episodes from clearing NASA: Long-range transport of Canadian wildfire smoke into the US, July 14–20, 2026 — cause of death #1 Taylor Farms lettuce recall, July 17, 2026 (Marler Blog) — cause of death #2 Colorado air quality: blame wildfire smoke (CPR, Jul 7, 2026) — Dan's half-on-fire home state Wildfires force 250,000+ evacuations in France and Spain (CNN, Jul 25, 2026) — my half-on-fire counterpoint Andrew Wilkinson on building Deep Personality (X) — "$5,000 of Claude credits replacing tens of thousands in psych evals" 16Personalities: INTJ, "The Architect" — Dan, apparently IPIP — International Personality Item Pool — the more "scientifically just" alternative Dan's consulting contact — unqualifiedadvicepodcast.com under Hosts Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 Substack really did unleash Pangram this week Dan said Substack rolled out AI-detection scanning of writing on the platform via Pangram. Confirmed: the partnership launched July 21, 2026 — readers can scan posts, notes, replies, and comments over 100 words, creators can pre-scan drafts and dispute results (Axios, Jul 23). Dan even correctly hedged that they didn't retro-scan the whole corpus — it only covers content published after launch. Nailed it, hedges and all. 🟢 Pangram is the cream of the crop, by a good mile Dan's characterization holds up in the third-party record: a June 2026 Vrije Universiteit Brussel evaluation found only Pangram produced satisfactory results (97.5% detection on fully-AI text, 0% false positives on ESL essays) while GPTZero, Turnitin, and Copyleaks whiffed, and a University of Chicago study concluded Pangram "dominates the other detectors across all thresholds" with an essentially-zero false-positive rate. 🟡 The Beast handbook says record 100 episodes, then come talk to me Sean remembered MrBeast's advice as coming from "the Beast handbook... playbook that came out a couple of years ago." The advice is real and Sean's paraphrase is dead on — MrBeast: "make 100 videos and improve something small with each video... by the end, either they're successful and don't need my help, or they never make 100 videos to begin with." But it comes from MrBeast's long-running social posts and interviews, not the leaked 2024 production handbook (that was an internal onboarding doc for employees). Right advice, wrong container — and Sean flagged his own fuzziness, so half credit and roughly fifty episodes to go. 🟢 Both sides of the Atlantic are half on fire Dan said Colorado is "half on fire" (with a fire in the neighborhood next door); Sean countered that France is too. Hyperbole licensed on both counts: by early July, Colorado had air-quality advisories on 22 of its 64 counties, the Aspen Acres Fire had burned tens of thousands of acres, and Denver briefly ranked worst air quality in the US (CPR, Jul 7). The week of recording, wildfires in France and Spain forced 250,000+ evacuations amid a record heat wave (CNN, Jul 25). 🟢 Death by wildfire smoke inhalation and diarrhea lettuce — same week, both real Dan's Oregon Trail death screen checks out on both causes: NASA tracked long-range transport of Canadian wildfire smoke into the US July 14–20, 2026, and Taylor Farms recalled its shredded iceberg lettuce July 17 amid an outbreak with over 4,000 CDC-confirmed cases. The meme was documentary. (Pedantry corner: the lettuce outbreak is Cyclospora, not E. coli — the diarrhea is real either way.) 🟢 Andrew Wilkinson's Deep Personality app Dan described deeppersonality.app accurately: created by Andrew Wilkinson, roughly an hour of questions, and the output includes a compact prompt approximating how you'd score across a battery of historical personality instruments. All confirmed — Wilkinson says he vibe-coded it with about $5,000 of Claude credits to replace "tens of thousands of dollars of psychological evaluations." 🟡 Every Bible was written by a scribe, then Gutenberg, then a thirty years war Dan's causal thread — print breaks the scribal monopoly, upheaval follows — is a respectable historian's argument (press → Reformation → wars of religion). But the Thirty Years' War began in 1618, about 170 years after Gutenberg's press, with a Reformation and several other wars in between. Directionally great, chronologically compressed. 🟢 INTJs are the architects (and the Big Five is the scientific one) Sean tossed off that INTJ types are "architects" — 16Personalities does brand INTJ exactly that, and Dan promptly outed himself as one. Sean's accompanying hedge, that Myers-Briggs is "a little horoscopy" while the IPIP/Big Five is "a little more scientifically just," matches the psychometric consensus almost word for word. A+. Final Score: 6🟢 / 2🟡 / 0🔴 When the week's news is "the lettuce is dangerous and everything is on fire," accuracy is easy — the machines fact-checking us have never had less to do. Chapters 00:00 - Cold Open: Did Anthropic Steal a Street Artist's Color Scheme? 02:37 - Substack Unleashes Pangram: AI Detection Comes for Writing 06:29 - Style Isn't Substance: What a Word-Order Detector Can't Score 09:28 - Oregon Trail, 2026 Edition: Wildfire Smoke and Diarrhea Lettuce 11:08 - AI Should Make You Think Harder 13:40 - Sean's Commute Agent: Daily Briefs and Audio Journaling 17:23 - Dan's Council of Five Advisors 20:01 - The Shop-Floor Filter: Killing Bad Ideas Cheaply 21:14 - The Gen X Opportunity: Build a Version of Yourself 23:02 - Leaky Companies and Tacit Knowledge 24:54 - Cleaner Agents: Never Disrespect the Janitor 26:59 - Second Grade vs. Graduate School 29:48 - Voice DNA: Staying You at the Point of Generation 31:38 - Fifty Episodes from Mr. Beast 32:56 - Gutenberg, Genies, and Machines Reading Machines 36:43 - Hire a Transformation Agent, Not Another IT Person 40:28 - What Agents Do You Bring With You? 42:43 - Personality Tests: INTJs, IPIP, and Deep Personality 48:35 - AI Therapy and the Mirror Problem 51:05 - The Force: A Dual-Use Invitation

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

You Have Died from Excessive Chatbotting

0:00 46:27

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of Unqualified Advice?

This episode is 46 minutes long.

When was this Unqualified Advice episode published?

This episode was published on August 10, 2026.

Can I download this Unqualified Advice episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.