A deep read of Tuan Le's What Getting 3 Billion Views Taught Me About Human Psychology, the source layer behind BuildOS's viral-content-for-boring-brands skill.
Tuan Le runs Shortscut and describes the framework as the operating logic behind brand-account campaigns for categories that are not inherently exciting: instant ramen, software, restaurants, and other products that need content mechanics before they can earn attention.
Why this content psychology analysis exists
This is the source analysis behind the BuildOS viral-content-for-boring-brands skill. It supplies the six pre-conscious filters that the skill turns into agent checks: format recognition, curiosity gap, identity layer, credential shortcut, sharer payoff, and story skeleton.
The useful point is not the headline statistic. The useful point is the sequence: viewers do not calmly decide whether to watch. Their brain pattern-matches, trusts, rejects, or leans in before the first two seconds are over.
Core thesis
Virality is less about the content itself than about six predictable shortcuts in the viewer's brain.
The filters fire roughly in this order:
| Timing | Filter | Operating Question |
|---|---|---|
| 0.5s | Format recognition | Have I seen this kind of thing before and liked it? |
| 1.0s | Curiosity gap | Is there a knowledge gap I want closed? |
| 1.5s | Identity layer | Does this relate to who I am or want to be? |
| 2.0s | Credential shortcut | Should I trust this person or scene? |
| Post | Share-as-status | Would sharing this make me look good to someone else? |
| Post | Hook -> Problem -> Story -> Payoff | Am I inside a story I want to finish? |
If a piece fails early, later craft cannot rescue it. A beautifully written post that opens with a product feature still triggers the ad-recognition pattern and loses the viewer before the argument starts.
Principle 1: Format recognition
Tuan's first claim is that the brain likes familiar structures. Original formats often fail because the brain reads novelty as confusion, not opportunity.
His ramen example is the cleanest version: a Korean instant ramen brand was stuck publishing manufacturing and ingredient content. The pivot was not a new product; it was putting the product inside spicy-challenge and reaction formats that viewers already understood. The brand moved from a static product story to a recognizable content structure.
For agents, this becomes a preflight check:
- Name the format the draft is borrowing.
- Confirm the format has recent evidence of working in an adjacent audience.
- Score whether the brand can use that format without distorting its voice.
- Reject formats whose audience-acquisition mechanism is rage, fake confession, humiliation, or manufactured drama.
BuildOS keeps the principle but filters the tactic. Familiarity reduces processing cost; slop-format theft burns trust.
Principle 2: Curiosity gap
Leading with the product triggers the viewer's ad filter. Leading with an unresolved knowledge gap gives the brain a reason to stay.
Tuan's Stan example makes the point: instead of opening with a creator-platform feature, the video opened with a rent question inside an office-tour frame. The software appeared after curiosity was already active.
Agent checks:
- Does the first sentence or frame name the product too early?
- Can the hook be described as disbelief, identity recognition, forbidden knowledge, surprising claim, or in-progress action?
- Does the piece close the loop it opens?
- Is the payoff strong enough to justify the gap?
Curiosity gaps become low-trust when they do not close. The skill rejects hooks that promise drama and deliver a normal feature.
Principle 3: Identity layers
Tuan uses the means-end chain: products have attributes, functional consequences, and psychological values. Most brands stop at the attribute layer, but viewers respond to identity and emotional payoff.
The ladder looks like this:
Layer 1: Attribute
-> what the product has
Layer 2: Functional consequence
-> what the product lets someone do
Layer 3: Psychological value
-> what that says about the person's identity or life
For BuildOS, "agentic workspace" is layer 1. "Keeps context across tools" is layer 2. "I am the architect; the agents execute" is layer 3.
Agent checks:
- Tag every claim in the draft as layer 1, 2, or 3.
- If the lead is layer 1, demote it into evidence.
- Keep asking "so what?" until the answer becomes an identity statement or emotional state.
- Use layer 3 as the message; use layers 1 and 2 as proof.
This is the most important principle for boring-brand content because it moves the content out of product explanation and into viewer relevance.
Principle 4: Credential shortcut
Viewers use authority as a shortcut because they do not have time to evaluate every claim. A credential can be a real number, a professional setting, a known affiliation, or a visible role.
The restaurant example shows the mechanism: generic cinematic food shots underperformed, but a chef in a professional kitchen inside a "can you make me something gourmet?" format gave the viewer an immediate trust cue.
The risk is performative authority. A credential shortcut is honest only when the authority is real.
Agent checks:
- Identify what credential appears in the first two seconds.
- Prefer real numerical proof, real usage, real customer outcomes, and real environment.
- Reject borrowed authority, staged expertise, fake offices, fake lab-coat energy, and implied credentials the brand has not earned.
- Move late credentials earlier if they are true and relevant.
For BuildOS, the honest credential is a founder using BuildOS to brief real coding agents, not a staged "AI expert" posture.
Principle 5: Share-as-status
Tuan's share thesis is simple: people share what makes them look smart, funny, knowledgeable, or ahead of the curve. Sharing is not only recommendation; it is identity performance.
That means every piece needs a one-line sharer test:
"yo check this out - _________________________"
If the blank is hard to fill, the piece probably will not spread. If the blank is only "this made me angry" or "this made me tired," it may attract views but will not compound into the audience BuildOS wants.
Agent checks:
- Write the one-line DM a viewer would use to share the piece.
- Name what sharing the piece lets the sharer say about themselves.
- Prefer humor, surprise, and insight over exhaustion or outrage.
- Reject pieces whose share payoff is only negativity.
This maps directly onto anti-feed content. The sharer payoff is: "I think about my attention and tools differently than people who let the feed run them."
Principle 6: Story skeleton
Tuan's default structure is:
Hook -> Problem -> Story -> Payoff
The hook opens the gap. The problem names the pain. The story gives concrete movement. The payoff closes the loop.
For video, this also means cutting any frame that does not deliver new information and using captions as a second processing pathway. For long-form writing, the equivalent is strong sectioning, bold pull-lines, and concrete receipts instead of abstract explanation.
Agent checks:
- Confirm the draft has all four beats.
- Cut dead-space sections that do not add new information.
- Make the payoff explicit.
- For blogs, translate visual pacing into structure: sections, pull-lines, examples, and receipts.
Anti-patterns
- Format theft of slop formats: stealing rage-bait or fake-confession structures inherits the wrong audience.
- Performative credentials: authority cues that are not literally true.
- Layer-1 identity claims: calling a feature an identity statement without laddering.
- Curiosity gaps that do not close: clickbait trains audiences to distrust the channel.
- Manufactured share-bait: engineering shareability without substance.
- Story skeleton with no story: using the shape without a lived receipt.
- Cutting for cuts' sake: fast edits only help when each cut adds information.
How to apply the six principles
Use the six principles as a diagnostic order — audit any piece of content through the same sequence: format, curiosity, identity, credential, shareability, story. Two of them need an ethics check, because the easiest implementations turn into low-trust slop:
- Lean on: the curiosity gap (that actually closes), identity layers, and the story skeleton — these are pure upside.
- Extend: share-as-status, but tie it to a real identity payoff rather than manufactured outrage.
- Handle with care: format recognition and the credential shortcut. Use familiar formats without stealing low-trust patterns, and use real receipts as credentials rather than borrowed authority.
This is the source layer behind the viral-content-for-boring-brands skill, which runs this exact audit order.