『The Machine That Worked Perfectly』のカバーアート

The Machine That Worked Perfectly

The Machine That Worked Perfectly

無料で聴く

ポッドキャストの詳細を見る

【Amazonプライム会員限定】今ならプレミアムプランが4か月 月額99円。

10月19日まで。※適用条件あり
This week: the day I switched off a content machine that had been running perfectly for nine months. Eight hundred and fifteen posts across four platforms, seventy-nine thousand views, twenty followers and no book sales, and what the shape of those numbers actually turned out to mean. Including a correction to my own headline figure, which was wrong. Plus TabletMinder's onboarding rebuild and the 6,015 line file that is the real reason it keeps not happening. The book's chapter seven, "It Has No Idea What You Want", which turns out to be about content machines as much as it is about prompts. And three bits of AI news, one of which I have been trying to verify since July. This is the first episode I have recorded myself, on camera, rather than narrating with a cloned voice. The reason why is the whole of the build log. Chapters: 00:00 Theme 00:11 Cold open 01:35 Welcome 02:02 Build Log, the machine that worked perfectly, and TabletMinder's 6,015 line file 09:46 The Bookshelf, chapter seven, It Has No Idea What You Want 12:29 AI News (verified) 16:55 The week's thread 18:23 Close The numbers in this episode, since I would rather you could check them: 815 posts over nine months across TikTok, Facebook, Instagram and YouTube. TikTok 263 posts, 65,523 views, 96 engagements, 2 followers. Facebook 53 posts, 5,727 views. Instagram 56 posts, 1,740 views, 16 followers. YouTube 443 posts, 6,046 views, 101 subscribers and flat. TikTok median 221 views per video, maximum 737 across 263 videos, top ten between 718 and 737. YouTube: 452 videos, 101 subscribers, 6,184 lifetime views, of which essentially all are short clips. The four full podcast episodes have 8 views between them (EP1 4, EP2 1, EP3 2, EP4 1), audited 2026-09-13. Average 249 views per post against a benchmark of roughly 350 for accounts of this size. Engagement rate corrected from "30x below par" to approximately 12x below par, after accounting for a 2.4x undercount of likes in the analytics source. Amazon Attribution: 10 clicks, 0 purchases, year to date. Book, Understanding AI: The Word Calculator and How to Use It Well, by Peter Hunter: https://books2read.com/u/31zr9r Featured app, TabletMinder, all 50,000+ NHS medicines built in: https://apps.apple.com/gb/app/tabletminder/id6757861874 Utter List, the to-do list that listens: https://apps.apple.com/gb/app/utter-list/id6761653207 JabBuddy: https://apps.apple.com/gb/app/jabbuddy/id6754968831 AI news sources, so you can check them yourself: GPT-6 Astra system card (OpenAI, 3 September 2026): https://deploymentsafety.openai.com/gpt-6-astra Astra launch coverage: https://techcrunch.com/2026/09/03/openai-launches-astra-its-powerful-and-controversial-new-model/ Meta Muse Spark pricing, standard and contributor tiers: https://developer.meta.com/ai/models/muse-spark/ Coverage of the Meta contributor tier: https://techcrunch.com/2026/09/03/meta-is-paying-to-peek-at-how-you-use-their-latest-ai-model/ GPT-5.6 Sol model documentation (OpenAI): https://developers.openai.com/api/docs/models/gpt-5.6-sol

Support the show
adbl_web_anon_alc_button_suppression_t1
まだレビューはありません