70% of YouTube watch time comes from algorithmic recommendations. Not searches. Not subscriptions. The machine. This episode exposes the actual mechanics of YouTube's recommendation system - the two-tower neural network that tracks 800 signals to control what 2.5 billion people watch.

Frank breaks down the code that destroyed human attention. From the 2012 watch time pivot that broke creator sanity to the rabbit hole engine that radicalizes by design. From the shadow ban that "doesn't exist" to the comment manipulation system that promotes toxicity for engagement.

Key revelations exposed:

• The two-tower architecture: Candidate Generator + Ranking Network

• Why 70% of watch time is algorithmic (90% for heavy users)

• The watch time formula: Session Duration × Frequency × Lifetime Value = Revenue

• How the rabbit hole pushes users toward extremes for engagement

• YouTube Kids: Algorithmic grooming for addiction

• Creator exploitation: Why passion gets punished, formula gets promoted

• The 4,000-hour reprogramming period before monetization

• Shadow banning through "reduced discoverability"

• Comment sections designed for conflict, not discourse

• Notification psychological warfare based on vulnerability patterns

• Why legitimate education gets suppressed alongside misinformation

YouTube made $31.5 billion in 2023. Not from subscriptions. From advertising. Every feature exists to create ad inventory. You're not watching videos. You're being processed into advertising opportunities.

The algorithm doesn't reflect taste. It manufactures it. It's not artificial intelligence. It's artificial instinct.

The brutal truth: YouTube built a machine that controls what humanity watches. And humanity volunteered to be controlled.

BrutalTechTruth: No hype. No BS. Just brutal truth about tech.

Support the show