Work log: Polymathic — March 28, 2026
What shipped today
The podcast got a major expansion. Paul received Apple Podcasts ranking data showing the show at #7 for “ai project building” but invisible for philosophical and human-experience terms — the exact territory where most of the essay content lives. The gap was clear: the podcast feed metadata was thin and episodes were limited to reflections (which skew technical), while the essays about human judgment, AI philosophy, and organizational change had no audio presence at all.
Three layers of fixes went out. First, the show-level description was rewritten to naturally include the missing search terms (“human side of technology,” “artificial intelligence philosophy,” “human experience and AI”) and an itunes:keywords tag was added. Second, 8 back-catalog essays spanning February and March were voiced via OpenAI TTS and uploaded to Cloudinary — roughly 77 minutes of new content covering topics from workforce economics to the philosophy of busyness. Third, <podcast:transcript> tags were added to every episode in the feed, linking to the blog post URL. Since all audio is TTS from the essay text, the post IS the transcript, giving Apple and Spotify full-text indexing for every episode.
The bigger architectural change: the pre-commit hook was rewritten to voice ALL published posts automatically, not just those tagged “podcast.” Any non-draft post in content/posts/ that doesn’t already have audio_url gets audio generated on commit. No tags to remember, no manual steps. The hook was also fixed to use OpenAI TTS instead of the old ElevenLabs provider (which was dropped for cost reasons). CLAUDE.md was updated to reflect the new workflow. CODEBASE.md was regenerated from scratch — it had bloated to 488KB and was replaced with a compact 79-line annotated tree.
Completed
- Podcast feed description rewritten for search discovery
itunes:keywordsadded to podcast feed- 8 back-catalog essays voiced via OpenAI TTS (~77 min new content)
<podcast:transcript>added to all episodes in feed- Pre-commit hook rewritten: all published posts get audio automatically
- Pre-commit hook fixed: ElevenLabs → OpenAI TTS provider
- Missing tags filled on “True 1-to-1 outreach” and “Intelligence briefing” posts
- “Building in public” removed from podcast (stub post, only 1 paragraph)
- CODEBASE.md regenerated (488KB → compact annotated tree)
- CLAUDE.md updated with new podcast workflow docs
Release progress
- April 2026: 0/3 open (all closed)
Carry-over
- Two monad essays (
monads-are-immutable-code.md,the-monad-cracks.md) sitting as drafts in the working tree — Paul says they’re not written yet - #112 — Blog dates accuracy (still
needs-clarification, no human reply since before Mar 20) - Confirm newsletter 16 delivery (was scheduled Mar 24)
- Check Brevo delivery stats for editions 13-15
- Update Brevo sender name to “Philosopher at Large”
Risks
- The back-catalog has many older posts without audio that will now get voiced if they’re ever re-committed. This is by design (idempotent — existing audio is skipped), but a bulk re-commit of old posts could trigger a lot of API calls and Cloudinary uploads.
Flags and watch-outs
- Apple Podcasts re-indexes feeds on a 24-48 hour cycle. The search ranking impact of today’s changes won’t be visible until ~March 30.
- The “Building in public” post is a stub (one paragraph). It was removed from the podcast but is still published on the blog. Worth finishing or archiving.
content_typefrontmatter field is deprecated per Paul — don’t use it for classification going forward.
Next session
- Check Apple Podcasts search rankings after ~48 hours to see if the metadata and transcript changes moved the needle
- Confirm newsletter 16 delivery stats (was scheduled Mar 24)
- Check Brevo delivery stats for editions 13-15
- Update Brevo sender name to “Philosopher at Large”
- Review the two monad draft essays when Paul is ready to publish them
- #112 — Blog dates accuracy still waiting on human reply
- Consider voicing the remaining back-catalog essays (pre-February) if the new episodes show positive ranking impact
Why customer tools are organized wrong
This article reveals a fundamental flaw in how customer support tools are designed—organizing by interaction type instead of by customer—and explains why this fragmentation wastes time and obscures the full picture you need to help users effectively.
Infrastructure shapes thought
The tools you build determine what kinds of thinking become possible. On infrastructure, friction, and building deliberately for thought rather than just throughput.
Server-side dashboard architecture: Why moving data fetching off the browser changes everything
How choosing server-side rendering solved security, CORS, and credential management problems I didn't know I had.
The work of being available now
A book on AI, judgment, and staying human at work.
The practice of work in progress
Practical essays on how work actually gets done.
Everything pointed at ghosts
Most organizations are measuring work they stopped doing years ago. The dashboard is green. The reports are filed. Nobody realizes the entire apparatus is pointed at ghosts.
Silence by design
Most systems have more suppression than their owners realize. It gets installed for good reasons. The cost accumulates slowly, in the form of systems you can't operate because you've removed the signals that would let you understand them.
Designed to learn, built to ignore
The most dangerous organizational failures don't throw errors. They look fine, return results, and quietly stay frozen at the moment of their creation.