EveryReviewed by the Liquid Gold team

Titles suggest solid learning usefulness for staying current with AI tools and applied practice, though instructional depth likely varies between tested walkthroughs and reactive live commentary. (Title/profile-level estimate; learning value cannot be confirmed without transcripts.)

• Concrete learning signals: "How to Build..." titles, "Build Your Own Repeatable AI Design Workflow," and week-long tests ("We Tested Claude Opus 5...") promise teachable, experience-based material.
• Expert instruction: Interviews with practitioners (GitHub CEO, Microsoft CTO, Notion cofounder, AI startup founders) indicate first-hand knowledge transfer.
• Variable depth: "LIVE VIBE CHECK," reaction-style, and prediction titles suggest some uploads are commentary rather than structured lessons.
• Category note: This is a mixed channel, so usefulness here is scored as learning value, not watch-worthiness.

Reliability appears mixed: many titles claim first-hand testing and name specific expert sources, but superlative hype framing and speculation are frequent. (Estimate from titles and profile only; claim support cannot be judged without transcripts.)

• Verifiable-style claims: Repeated "We Tested X for a Week" framing implies hands-on demonstrations viewers could compare against their own experience.
• Named sources: Titles cite specific executives, engineers, and researchers (GitHub CEO, Microsoft CTO, Anthropic team members), which typically grounds claims in attributable interviews.
• Hype framing: "INSANE New Models," "IT'S A MONSTER," "IS A BEAST," and "THE GOLD STANDARD" read as promotional superlatives rather than evidenced claims.
• Speculation: Prediction and commentary titles ("Four Predictions for How AI Will Change Software in 2026") are inherently unverifiable outlook pieces.

Promotional density looks moderately high: the profile is an explicit subscription pitch and a meaningful share of titles showcase the channel's own paid products, revenue, and services. (Higher score is a negative funnel signal; this is a title/profile-level estimate, not a count of actual calls to action.)

• Direct subscription pitch: The profile states "The only subscription you need to stay at the edge of AI" and links to every.to for "Ideas, apps, and training."
• First-party product titles: Spiral, Cora, Monologue, Claudie, and "Compound Engineering" uploads promote tools branded or sold by the channel itself.
• Revenue-narrative content: Titles like "Behind Our Largest Subscription Revenue Increase" and "How We Built a Media and Software Empire With AI" market the channel's own business as content.
• Counterweight: A large portion of titles are editorial testing, third-party interviews, and commentary with no visible product hook, keeping this below the top of the scale.

Titles mostly indicate genuinely new material per upload — original tests, builds, and a numbered interview series — with some recycled re-cuts and a heavily templated live-test format. (Estimate from titles only; re-edit duplicates cannot be confirmed without transcripts or curated metadata.)

• Original production: Numbered podcast episodes (Ep. 1–56), in-house builds, and first-person experiments suggest newly produced content for most uploads.
• Recycled re-cuts: Multiple "Best of the Pod" titles repeat interviews already released in full (e.g., Reid Hoffman, Kevin Kelly, Guillermo Rauch, Dwarkesh Patel), adding uploads without new recordings.
• Templated format: Recurring "LIVE VIBE CHECK" and "We Tested..." uploads follow a repeatable format, though each appears to cover a distinct new model or product release.
• Limit: Whether re-cuts contain new commentary or are near-duplicates is not determinable from titles alone.

Newest video 3 weeks ago · 105 uploads in the last 12 months