PARTLY

Partly true: the 'vibe code a $16m sales guy' agent is a sound reciprocity-email idea on a real, cheap decision model (Jev at $0.042 per million input tokens), but the demo is staged data and it copies one tactic, not a top broker

Facebook · Oct 2026

The claim'Vibe code a $16m sales guy.' Founders are 'plagiarizing' Ryan Serhant's sales system: he keeps sending you valuable things until you respond ('the Give, Give, Give, Give, Ask formula'). Steal it: when you publish an article, an AI agent saves it to a database table, pulls every contact in your CRM that has gone quiet, uses 'any small categorization model' (Jev, shown at '$0.042 per million input tokens, output free, no free tier') to rank each person by how relevant they would find it, and emails the highly relevant ones. Or in reverse: when a new contact lands in the CRM, rank every article and send the best fit. 'You just engaged in a little bit of reciprocity.' 'Follow your boy, I'm gonna keep shitposting stuff like this for free.'

The idea is good and the tool is real; the packaging is louder than the build. The reel shows a Serhant.com page describing a 'Give, Give, Give, Give, Ask' follow-up sequence and a $15,995,000 penthouse listing, which is where the '$16m sales guy' line comes from. I could not independently confirm the formula text, but the screenshot is of Serhant's own site and the principle (lead with value several times before you ask) is standard sales advice. The model named on screen, Jev, is real: TypeSafe launched it in mid-September 2026 as a 'decision model' that returns typed answers with probabilities instead of chat, and TypeSafe's site prices it at $42 per billion input tokens, which is the $0.042 per million the reel shows, with no free tier mentioned. Press from Tom's Hardware and Forbes covered the launch, and AWS and OpenAI shipped competing decision APIs within two weeks, so 'any small categorization model' is accurate. What is staged is the demo. The Airtable base, the 'Sam Porter' CRM contacts with 312-555 phone numbers, the 'Jev scored 212 quiet leads' chat and the drafted emails are mocked-up examples, not a running pipeline with real customers or results. And the agent copies one tactic, send relevant content to quiet contacts, not a $16M broker's whole system of listings, relationships and follow-up. Nothing is sold; the ask is a follow.

What holds up

Reel read frame by frame: a Serhant.com page headed 'The Give, Give, Give, Give, Ask Formula', a $15,995,000 listing, an Airtable content library, a CRM contact record, a 'Jev pricing' page ($0.042 per million input tokens, output free, no free tier), a Claude Desktop chat 'Jev scored 212 quiet leads', and relevance buckets (Very relevant 6, Slightly 41, Not relevant 165).
typesafe.ai fetched Oct 2: 'Jev is TypeSafe's first public System One Model, optimized for automation', priced at '$42 per billion input tokens' (= $0.042 per million); no free tier shown. News feed: TypeSafe launched Jev Sep 16, 2026 (The Rundown), Tom's Hardware Sep 21, Forbes Sep 22; AWS and OpenAI announced competing decision models Sep 29 to Oct 1.
The demo CRM, contacts and emails use placeholder names and 312-555 numbers; this is a mocked walkthrough, not a live run with results.
The Serhant 'Give, Give, Give, Give, Ask' formula appears only in the on-screen screenshot of his site; not independently verified by me. The principle is ordinary value-first follow-up.
Nothing is sold; the call to action is 'follow your boy'.

What doesn’t

'$16m sales guy' is a punchline; the agent reproduces one follow-up tactic.
Demo data is staged; no evidence it was run against a real CRM or produced a reply.
Jev has no free tier and the reel skips the CRM, email and database plumbing that is most of the work.
Mass-emailing 'quiet' contacts needs consent and unsubscribe handling, which the reel ignores.

The catch

Matching your content to the people most likely to care, and sending it before you ask for anything, is a genuinely good use of a cheap classifier. Build it against your real CRM with opt-in rules, and judge it by replies, not by the vibe of the demo.

How to actually do it

Tag each article you publish with a two-line summary and store it in a table (Airtable, Notion, a sheet); export your CRM contacts with their last-touch date and any notes.
Use a small classifier (Jev, or a cheap Claude or GPT model) to score contact-to-article relevance; a prompt that returns a 1 to 10 score with one reason is enough. Start with the top 10 scores, not the top 100.
Send short, personal, one-link emails from your own address, respect unsubscribes, and log replies for 30 days; keep the system only if it starts conversations.

Jev is real and priced as shown, and the reciprocity-ranking pattern is sound. The reel oversells a mocked demo as a cloned top broker, and leaves out the data plumbing and consent rules that decide whether it works.

Confidence: Medium
Posted by: a build-in-public 'vibe coding' creator, ~5k followers; reel sent to Buddy Thu Oct 1, 2026, ~9:31 PM

See the original claim →

We test hype for free. We build the real thing for a living.

Thirty minutes, no pitch — and you'll leave with something useful either way.

Book a call with Todd or start with the free Business Checkup →

The Verdict Weekly

Three verdicts every Friday. Free forever, unsubscribe anytime, no spam — that would be ironic.