TITAN AI Podcast — Concept Brief

> Auto-publish strategy + concept. The 5 daily newsletters → 5 daily podcasts.

> Drafted 2026-05-15 by JARVIS. Reviewed once, then schedule it.

---

Why this works

You already produce all the raw materials. Polly Ruth + Stephen voices and the

narrate.py pipeline already render full dialogue MP3s per newsletter (solo

and 2-host). The narrative work — turning curation into a script — is already

done by Sonnet. You ship daily by 13:00 UTC. **The podcast is the same product,

different package.**

Cost per episode = $0.13 (already paid). Distribution cost = ~$0 (S3 + an RSS feed).

Audience-fit is sharper than a generic AI podcast: each show owns a niche.

---

Format: 5 shows, daily, ~7 min each

|---|---|---|---|

Episode structure (all 5 use the same skeleton via narrate.py):

00:00–00:30 Hook (one sharp line + theme of the day)
00:30–05:30 Top 5 stories, ~1 min each — what happened / why it matters / our read
05:30–06:30 "Builder's Note" — a TITAN-flavored take (sneak peek of the guide)
06:30–07:00 Sign-off + tomorrow's hook

---

Auto-publish pipeline (~2 days to ship)

Components we already have:

✅ Newsletter HTML archive on S3 per slug per date
✅ Polly Ruth/Stephen narration MP3s (via narrate.py)
✅ Cinematic video clips per app (symphony/ and titan-video-daily/)
✅ Daily cron infrastructure

What we add:

1. build_podcast_rss.py — generates podcast-rss/<slug>.xml per show.

Reads existing narrate MP3s + their dates, produces a podcast RSS 2.0 feed

with iTunes + Spotify tags. ~80 lines.

2. build_podcast_landing.py — public landing page at

https://jarvis.livegroweveryday.com/podcast/ listing all 5 shows with:

- Subscribe buttons (Apple, Spotify, Pocket Casts, RSS direct)

- Latest episode embed

- Show notes link (per-episode → existing newsletter player page)

3. Submit to directories (one-time, manual):

- Apple Podcasts Connect — submit RSS, takes ~3 days approval

- Spotify for Creators — submit RSS, ~24 hr

- Pocket Casts — auto-indexed from public RSS

- Overcast — auto-indexed

- Castro — auto-indexed

4. Daily cron: extend titan-jarvis-feeds-daily to also rebuild the RSS

after each newsletter narration is generated. Pure XML serialization, no

external services.

5. Watchdog: add 5 checks to verify_and_fix.py — each show's RSS feed

must be < 26h old and contain today's episode.

---

Marketing angle (when newsletters do well)

> *"5 podcasts. 5 niches. 5 minutes each. The fastest way to stay current

> on the AI agent ecosystem — without becoming chronically online."*

Positioning:

Not another generic AI news show. Hyper-niche by stack. People listening

to The Agent Stack Daily don't want OpenAI press releases. People on OpenClaw

Briefing don't want LangChain tutorials.

Built by AI, edited by a human. Transparency is the moat. Show your work.
Tight format. 7 min flat. No filler. No interview drag.

Distribution playbook:

1. Week 1: cross-post episodes to X / LinkedIn with a 60-sec audiogram clip

2. Week 2: pitch to 3 AI podcast aggregator sites (Latent Space, AI Tinkerers,

Practical AI)

3. Week 3: start measuring listens per show. The strongest single show gets

doubled-down on — that's the marketing focal point.

---

What I need from you

Go / no-go. Reply "ship it" and I queue all 5 in the next 48h.
Show names — any rename on the 5 working titles?
Voice mix per show — happy with ALEX/MAYA or want different combos?
Sponsor-readiness? Beehiiv has sponsor marketplace built in;

do we want the podcasts to share that revenue stream from day one?

---

Cost projection

| Item | Daily | Monthly |

|---|---|---|

| Polly narration (already paid) | $0 | $0 |

| S3 storage (RSS + MP3s) | $0 | ~$2 |

| RSS landing page hosting | $0 | $0 |

| Total NEW spend | $0 | ~$2 |

This is upside-only.