Friday 9 October 2026 · private briefing

Next moves

Everything open across CT, CallumAI, ELA and Lore, sorted by what it needs from you. Seven calls to make, five things you're waiting on, two agents still working.

If you only do three things

  1. Send the DDRE paragraph to Mark. Version 3 is ready in the first card. Ten seconds.
  2. Declare the one-pager winner. Nothing was ever picked. B (payment slip) is the ranked favourite.
  3. Say yes to the Lore security check. One harmless request confirms or clears a July finding on prod.

Decide

7 calls
Decide10 seconds

DDRE: send the rewritten paragraph

Take three. White label dropped, half the words, one proof number. Mark wants to show Daniel the art of the possible; this is built to be forwarded as-is.

Send to MarkTweak againPark
read the paragraph
There's a currency deal inside every property deal DDRE closes: buyers bringing dollars or dirhams into sterling, sellers taking proceeds home, all of them treated by their banks like queue numbers. The partnership is simple. DDRE introduces the client; we look after them to the standard DDRE sold them: a named person who answers, rates banks won't give on big money, and the exchange rate locked from the day the deal is agreed to the day it completes. DDRE earns a share of our revenue on every client, for life, long after the keys change hands. Our top introducer has earned over £350,000 from a single introduction. Nothing to build, nothing to pay, and the first introduction could be trading this week.

Hold the exact percentage for the call (premium tier is flat 30, same paper Imran signed).

Decide10 seconds

One-pager: pick the variant

No pick was ever recorded: board ratings empty, memory says "hasn't picked yet". You downloaded A (navy), D (house) and B (payment slip) on 30 Sep; C and E never. The session's own ranking: B, then C, E, A, D.

Declare BRate the boardReopen the session
resume point + loose ends

The session is likely already open in a herdr tab (restored 3 Oct). Check before opening a second copy. Otherwise:

cd /Users/stevanl/dev && claude --resume f1663a29-6629-426f-8038-6028476ba696

Board: cjl2-1dsd-y587.pages.dev. Still open from that thread: Laura's OK on her quote, Paul's EUR account details, and your calls on headline, QR target, £250k threshold, back page.

Decide5 minutes

CallumAI: name and the $3k domain

Real collision: an AI calendar assistant called Callum (getcallum.com, Product Hunt, iOS) plus a pending US trademark for "Callum" (Vortex Software, filed Jul 2026). Domain is a speculator hold from Jul 2025: the $3k ask is pure margin.

Keep name, offer $1,200Persona Callum, different company brandName sprint
why decide now + haggle path

Decide before Meta approves you. The WhatsApp display name goes through review; renaming later means another round of the same pain.

Haggle: open the for-sale page; if there's a Make Offer button, open at $1,200 and let the counter-dance land near $1,500 to $2,000. Buy-now only? GoDaddy Domain Broker (about $120) negotiates for you. Pay $3k only if you're certain and tired.

Decide5 minutes

Video: publish the social cut

The public version already exists: a 0:34 social cut + silent loop, specced by Ingrid's own brief for the partnerships page. The 1:20 and 1:30 films stay as one-to-one WhatsApp sends. Two fixes before it goes live.

Publish after 2 fixesRe-shoot earnings sceneHold
the 2 fixes + logo answer

1. The "£5,997.87 earned" scene is a whale client; on a recruitment page it reads as an earnings claim. Keep the illustrative label working or re-shoot on a typical client.

2. End card says "Earn on every trade they make"; clause 9 of the current agreement can stop commission. Sign the new agreement first or drop "every".

Logo (Ingrid): not sticky. Enlarge the existing wordmark on title and end cards (it's about 55px now). Include the social cut in tonight's font re-render; Phil needs partner-social-loop.mp4 + poster jpg. Preview: draft 4.3 (public board still serves 4.2).

Decidenext step

ELA: fixes landed, documents now lag the model

All four changes are in and verified: 28,642 cells cross-checked with zero mismatches, backups kept, nothing deployed. The honest numbers moved: Base hits first profit in month 8, dips in months 11 and 12, steady from month 13; month-36 cash AED 16.0M (was 19.8M); Cautious now runs out in month 14. The plan, teaser and Jon's note still quote the old figures.

Regenerate the docs, then shareReview the model first
what changed + the Jon plan

Label bug fixed everywhere; a "Start here" front sheet that prints on one A4; engine tabs hidden; sensitivity rebuilt (Base breaks at about AED 885 of ads per paying member; seats row added); your diff-report retention applied to all three cases. Formula count down 32%. Backup: backup-pre-fix-2026-10-09/.

The 15-minute Jon walkthrough still stands: Summary and Assumptions open, what-ifs pre-run, "why not 300,000" anchored to his own July range. Known lag: the pack page description and thumbnail, the 2 Oct report copy, and build-plan.js (now fails loudly on the removed grid, on purpose). Pack: preview.ela.now/pack.

Decideyes / no

Lore: one security check, standalone parked

Lore is alive: 328 commits and 33 PRs merged in three weeks, revamp branch mid-P4a right now, prod running the P3 build behind your founder flag. One July finding looks unfixed in code: about 70 API routes with no login required. Unverified against the live site.

Yes: run the 401 checkFold a fix into P4aLeave it
the standalone decision recipe

Don't decide standalone during the revamp. Two to three weeks after P4b, check three facts: colleagues use it weekly without help (September usage was near zero), a non-CT sample dataset reaches a trustworthy answer in days, and the security finding is closed. All three hold: pick one name (codebase says Lore, bundle says Syno, usesyno.com is currently dead) and put up a one-page waitlist before building signup or billing. Any fail: keep it internal.

Parked housekeeping: 79 worktrees (33 GB, 62 deletable), six dead September PRs to close, two unique local plan files to save before resetting the stale checkout.

Decideyes / no

Partner docs: make the library real

Your instinct was right: ~/dev/partnerships is already the natural home (Wow memo, connector briefs, June plan docs, term sheets), but it has one git commit, from 7 June; everything since is untracked, and the one-pager sources live outside it in .gstack with a deploy script pointing at a folder that doesn't exist. Proposal: move the one-pager finals in, write a CONVENTIONS.md capturing style and reusable patterns (per-connector QR, proof-point rules, the pricing caveats), and commit the lot so new partner docs take minutes.

Approve the tidy-upLater

Waiting on

5 people

On you tonight

3 items

In flight

1 workstream

Closed this week

Mira's passport DHL'd CIL paid Peter's Xero query answered Deyan's fixes in PR #5732, your eyeball test Wow: Peter meeting 19 Oct confirmed Imran agreement sent Shaz pack held on purpose until numbers settle