The Final Phase of Social Media Just Started
A 16-minute breakdown of why the viral playbook is dead, the safe-community playbook fails silently, and the five pillars that build a brand algorithms cannot touch.
June 11thA 20-minute live-fix session where three creators get their hooks rebuilt from scratch, one frame at a time.
A hook works not because it is interesting but because it opens a loop the viewer's brain physically cannot close without watching further — and every design, caption, and framing decision either protects that loop or collapses it.
When someone scrolls their feed the swipe is a reflex — the brain is on autopilot until something interrupts it. A hook's only job is to plant a question the brain cannot answer without watching. The video demonstrates three failure modes live: missing context, missing emotional stake, and visual clutter. The consistent repair is the same each time — add context immediately, withhold the answer, and ensure the first frame has one clear focal point with readable text in the safe zone. Visual quality is secondary to curiosity architecture.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →
The neurological frame: autopilot scrolling, the swipe-back behaviour, and open loops as the interrupt mechanism.

Creator 1 (Max): hook fails because there is no context. Fix — state who you are and challenge the belief simultaneously.

Caption placement and size — the safe zone rule and why oversized captions destroy credibility.

The strongest hook framework: find what the video already challenges and lead with that contradiction.

Live Photoshop fix of Max's hook: font weight hierarchy, background overlay, safe-zone text positioning.

Using ChatGPT image generation to create a better first-frame visual; before/after side-by-side comparison.

Creator 2 (Stefan): his reels feed and photo feed are completely mismatched, destroying trust on landing.

Why 'Who can relate?' is a dead hook: no context, no question, no target-audience filter.

The contrast: Stefan's cold photography vs client Constanti's 'my photos 10 years ago vs now' — story plus emotional investment.

Creator 3 (Ella): coffee creator, hook score 3/10. Problem: visual clutter and zero curiosity loop.

Live rebuild for Ella: AI first-frame, text hook 'this coffee should NOT taste good', deliberate withholding of the product reveal.

Color picker live demo: red reads cheap, yellow with outer glow wins; blur the key ingredient to deepen the loop.

Recap and CTA: free account review link and next-video bridge.
Every hook that fails does so for the same reason: it either skips the question entirely or accidentally gives the answer away before the viewer is committed.
“You wanna open loops and don't close them as long as possible. Basically keeping them dangling at all times.”
“There's a very fine line from being borderline criminal and getting people's scroll-stopping attention.”
“Challenging a common belief is one of the strongest forms of creating hooks.”
“Who can relate is not a strong question. Because I'm like, I don't know what this is even about so I'm gone.”
“There are a gazillion photographers out there and you're not special — that is until they get to understand your personality.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
Start with a number that sounds like a credential and you have already bought yourself five seconds. The title makes a research claim — a thousand hooks studied — and the opening line cashes it immediately: bad hooks are the single most reason your reels fail. Everything that follows is proof.
Open a question in the viewer's mind as early as possible, withhold the answer as long as possible, and open a new loop the moment you close the previous one.
A viewer must understand who the content is for within half a second of seeing the first frame. If they cannot, the algorithm loses the targeting signal and the viewer swipes.
Find the claim in your content that contradicts what your audience already believes. Lead with the contradiction, not the explanation.
Stack three layers of ambiguity — visual, textual, and subject-level — so that resolving any one of them still leaves the others open.
“If you wanna get your accounts and content reviewed for free, link is in the description.”
Soft lead-in then bridges directly to a next-video CTA — two-step conversion play.
00:01
00:16
00:34
00:56
01:08
01:24
01:32
01:58
02:05
02:18
02:34
02:52
03:02
03:18
03:33
03:42
04:02
04:18
04:35
04:51
05:06
05:21
05:36
05:50
06:05
06:20
06:34
06:48
07:08
07:17
07:33
07:42
07:56
08:15
08:31
08:52
08:54
09:16
09:30
09:48
10:00
10:20
10:30
10:52
11:07
11:22
11:33
11:52
12:11
12:27
12:42
12:58
13:10
13:29
13:43
13:55
14:10
14:29
14:46
15:02
15:16
15:36
15:49
16:05
16:19
16:27
16:50
17:02
17:20
17:35
17:50
17:58
18:22
18:38
19:02
19:10
19:26
19:35
19:50
20:02A 16-minute breakdown of why the viral playbook is dead, the safe-community playbook fails silently, and the five pillars that build a brand algorithms cannot touch.
June 11thA 16-minute breakdown of the curiosity psychology behind viral hooks, built around one three-step formula and five tactical amplifiers.
December 19th 2024A 4-minute breakdown of why on-screen text outperforms spoken hooks -- and the three-step STI framework that turns any hook into a scroll-stopper.
June 5thA 100x Engineers host walks through the exact Claude Projects setup he uses to auto-generate Instagram Reel scripts from 400 of his own past posts.
November 11th 2024A creator gives away the five Claude Skills he built for thumbnails, content research, carousels, motion graphics, and scripts.
July 1stA timer-driven, screen-shared work session where a creator scripts, batch-films, and edits short-form videos in real time — and lets you do it alongside him.
June 16th