DECLASSIFIED
PUBLIC BRIEFPULSE-BRIEF-001
...
SUBJECTThe YouTube Intelligence Problem
PREPAREDPulse Architecture Group
STATUSActive Development

The smartest people in your field publish on YouTube. You're missing most of it.

One creator publishes 3 hours a week. You follow twenty. That's 60 hours of content. You have maybe 4.

01Situation

In AI, markets, crypto, and geopolitics, the most current expert thinking increasingly arrives on video — not in papers. 500 hours of content are uploaded to YouTube every minute. The platform now reaches 2.7 billion monthly users.

The ideas are brilliant. The format doesn't scale.

You watch one video and miss nine. You forget what someone said last month. Two creators contradict each other and you never notice. A new idea surfaces in an obscure channel — by the time it reaches the people you follow, you're weeks behind.

The signals are there. You just can't extract them at scale.

Sometimes you just have a question. “What did Lex say about open-source models?” The answer is buried 47 minutes into a 2-hour conversation. You scrub the timeline, skim chapters, watch at 2x. Fifteen minutes later you either found it or gave up.

You shouldn't have to watch a video to get what's in it.

Watch-later playlists with 400 items. Notes you never revisit. 2x playback until everyone sounds the same. The problem isn't discipline — no human can synthesize dozens of minds publishing in real time.

What you do now
×Watch what you can, skip the rest
×Forget what you watched last month
×Miss when creators contradict each other
×Never see patterns across channels
×Find out about ideas after they're mainstream
What an analyst would do
→Track every video from every creator you follow
→Map who agrees with whom — and where the debates are
→Flag when someone reverses a position or makes a new claim
→Surface ideas gaining traction before they break out
→Brief you every morning on what changed
02Assessment

The intelligence community solved this decades ago. When raw information overwhelms — intercepts, field reports, satellite feeds — you don't ask analysts to consume faster. You build a system.

Collect.Process.Analyze.Brief.

That cycle — the intelligence cycle — turns overwhelming volume into clear understanding. Bloomberg built the same architecture for financial news — 1.5 million stories a day, distilled into actionable intelligence.

Nobody has applied this to YouTube. Until now.

03The Approach

Pulse is a personal intelligence
system for YouTube.

Point it at the creators you follow. It watches everything they publish, extracts positions, claims, and ideas — maps connections across creators — and delivers briefings you can read in minutes.

Know what changed overnight
Open Pulse. See who published, what they said, and whether it shifts the picture.
See who agrees with whom
On any topic — consensus, debate, and who’s on which side. Instantly.
Ask, don’t watch
Question about what a creator said? Ask Pulse. Get the answer with the exact moment — no scrubbing.
Track claims over time
Predictions, positions, insights — Pulse remembers. See who’s credible by track record, not reputation.
Catch what’s emerging
When creators start discussing something new, Pulse flags it before it goes mainstream.

Watching the creators you follow becomes a choice — not the default mode of engagement.

04Status

Pulse doesn't exist yet.
The architecture does.

10 research bets validated. 22 design decisions documented. Every tradeoff, every correction — public. Not a pitch deck. A working blueprint, built in the open.

10 architectural bets██████████ validated
22 design decisions██████████████████████ documented
0 lines of production░░░░░░░░░░ design first
View on GitHub ↗
DECLASSIFIED