Adam · Jordan · Spencer · Matt · Wednesday, August 5, 2–5pm · Spring Place, Room 508. Three hours, and the week stacks behind it: the Polymarket intro call at 5pm the same day, Mediaocean at their office Thursday 10:30, the iHeart follow‑up Friday 1:30. The plan is to talk: walk the two files on a screen, argue with them, go wherever the conversation goes. One rule only: the recording runs from minute zero. The structure on this page is not a script for the room; it is what Claude pulls from the transcript afterward. Be as sloppy as the thinking needs. This page prints clean.
Walk the file on screen. At each section, argue the question out loud. Positions taken, doubts raised, language that sounds like us: all of it gets cataloged from the recording.
| On this section | Talk through |
|---|---|
| The company · the politics | Are we reading the room right: does Bill's growth hunger carry this if Zvika digs in? Who else at Mediaocean needs to feel ownership before the bet meeting, and who gets there first? |
| The machine · decision classes | If Kane resists the buy grid, which decision class do we fall back to? Do we actually believe the payable‑source pick is the best unobvious wedge, or is dispute policy stronger in a finance‑heavy room? |
| The seam | Is "your compliance story" the right altitude for Markman, or a step too forward for meeting one? What is the one sentence version each of us would say? |
| The assumptions | Which A‑rows would each of us personally bet against? What belief are we carrying that is not in the register at all? |
| The rooms · the yeses | Are the three load‑bearing yeses the ones we would actually trade for? Is learning retention truly first‑paper, or negotiable? Is the trust ladder sequenced the way we would really run it? |
| Wednesday · the run sheet | Do the two opening lines sound like us out loud (say them)? Who really carries each beat? What does a great Wednesday feel like at minute 91, concretely? |
| The program · the substrate | Are we one machine, or does bundling three workstreams read as overreach? How hard do we say the substrate goal in room one versus letting the boards imply it? |
| The boards | Would Kane recognize these screens as his product? What do we cut so the 10–35 beat lands in twenty minutes? Which board is the one we would leave on screen while talking? |
| The connection map | Is the invited overlay smart or a trust risk in the first ninety days? Do we say the Boombox trial on the OMG seat out loud on Wednesday, or hold it? |
| On this section | Talk through |
|---|---|
| The wedge · five facts | Which fact opens? Is the share‑gap framing (outspending, losing share) safe to say to their faces, or does it read as an insult to the people who bought the media? |
| The first proof | Is the World Cup retrospective the ask, or is the coverage read alone the safer first yes? What is the smallest thing they could grant that still starts the track? |
| The room | Pre‑brief Anthony through Peter, or let P09 land live? How do we make Jenner the hero of this rather than the person being graded? What does Peter need from us to look brilliant? |
| The company · the raise | Does the $20B raise framing help (board‑level efficiency story) or spook (why are outsiders reading our numbers)? Do we say the word Kalshi at all, and who says it if so? |
| The working deck D1–D5 | Which spread opens the meeting? Do the boards show live or travel as the leave‑behind? Is the path story (board 03) a meeting‑one idea or an earn‑it‑later idea? |
| The questions P01–P13 | Which three answers do we most need before scoping anything? Which questions does Peter answer for us before the meeting even happens? |
| Positioning · pricing | Referee‑not‑player versus becoming their in‑house brain: where exactly is our line? And if they ask what this costs, in the room, what do we say? |
The catalog rule: nobody takes notes. The recording runs, and Claude catalogs every position, objection, phrase, and deck request from the transcript into both files the same day. What the talk resolves gets written in; what it opens gets listed with owners.
The standing methodology, the same one that produced everything above. It runs on one input: the transcript.
| Step | What happens | Who |
|---|---|---|
| 1 · record | The session is recorded from minute zero; the transcript file lands with Claude the same day (Downloads, a paste, anywhere readable) | the room · X‑D1 owner |
| 2 · re‑room | Claude diffs the transcript against both files: decisions resolve the queue, outcomes become walls, killed assumptions get dated, new questions join the tables. The files update the same day, same URLs | Claude |
| 3 · manifest | Every deck request from the transcript becomes a manifest line: audience · register · the spine (which boards, rows, and lines lift) · the verbatim language that landed in the room · the candor rules that apply | Claude proposes, one of you approves |
| 4 · build | Decks are assembled from the manifest, reusing the boards and rows already built to lift. Anything external‑facing runs the five‑lens adversarial pass before it ships | Claude |
| 5 · publish | To pages.dev under the same projects, noindex, toggle, tan; links back into the files | Claude |
| 6 · grade | After the external meeting, its transcript comes back and the loop runs again. The files are working when they change | everyone |
The one habit that makes this work: the transcript always gets filed. Every deck this arc has produced traces to a recorded conversation; the sessions that were not recorded are the gaps in the timeline.