The teaser, cut from 7 pages to 3.
Two versions of the same 3-page document. Both restore the back story from the original two-pager, and both apply the edit notes I judged made it stronger (full list at the bottom). They differ in one thing only — how much room goes to the argument versus the product. Page 1 is identical in both.
Keeps the full case: the four-part “why now”, both pull-quotes, the three-verb triptych (Talk / Book / Connect as explained cards) and the four things that make a brain grounded, in their voice, citing, and honest enough to refuse. The product appears once, small, on page 3. Choose this if the reader needs convincing before they will care what the screen looks like.
Trades the triptych and the “what a brain is” panel for two large product shots: one question answered by six different minds, and a real expert page with Ask and Book session — $200 visible. Talk and Book get proved rather than described. Choose this if the reader is the kind who scrolls to the screenshots first — it shows there is a working product, not a concept.
| Note | Verdict | Why |
|---|---|---|
| P1 headline — remove “and” | Taken | “Knowledge you can talk to, a human you can still book.” Dropping the conjunction makes the two halves a matched pair instead of a list. Better rhythm, same meaning. |
| P1 subhead — should “Talk to it free” go, if users pay a subscription? | Your call | Left in for now — see the flag below. This is a pricing decision, not a copy one, and I shouldn’t make it silently. |
| P2 — “AI can source knowledge. It can’t provide real human insight… personal nuance is often the driver in learning.” | Part | Took “AI can source the knowledge” — a sharper, more accurate verb than “hand you information”. Left the rest: “can’t provide real human insight” is vaguer than “can’t hand you judgment”, “often” hedges the claim, and learning narrows marnji to education when it’s equally about decisions and advice. |
| P2 — “The position isn’t AI platforms vs human knowledge. The solution is a platform that combines the best of both.” | Left | “Combines the best of both” is a phrase every deck uses, and “the position isn’t…” is a slower way in. The original lands harder in half the words: “The question isn’t AI or people. It’s a platform built for both.” |
| P2 quote — “Access to knowledge has never been greater. More than ever we need lived experience to turn insight into learning…” | Part | Took the opening line — it’s cleaner than what was there. Kept the original’s logic for the second half: you turn information into insight, not insight into learning. Now reads: “Access to knowledge has never been greater. Lived experience is what turns it into insight instead of noise.” |
| P3 — delete “One brain, two of the three verbs…” | Taken | Agreed, and worth naming why: it made the reader do arithmetic about our internal framework. Gone in both versions. |
| P5 — is there a better image for the top of the page? | Taken | Yes, and the problem was bigger than that page. The old shot was mostly empty white space and showed the asking screen on a page arguing that anyone can create a brain — wrong evidence for the claim. Replaced with a tight crop of one question being answered by six different minds. I also recut the cover image, which was worse: two of the four faces were sliced to just hair and a forehead. |
| P7 closing line — end on “Everyone knows something worth knowing…” | Taken | Right instinct. By the last page the product needs no restating; the ellipsis leaves the reader with the human idea rather than a summary. |
Browser labels instead of URLs. The screenshots were captioned marngi.com — the real staging
domain, but a letter off the brand, which reads as a typo in a pitch. Swapped for plain descriptive labels, which is
also honest: we’re not implying a URL that isn’t live yet.
What went, to make 3 pages. The four Connect examples (songs / your AI / Slack / a grandparent’s brain), the standalone “Anyone can create a brain” page, the grounded-answer screenshot, and the team photo. Connect survives as the third verb but no longer gets its own space — it was the most speculative page in the deck (“ideas of what’s possible, not live integrations”), so it was the cheapest to lose. Say the word if you want it back as a fourth page.