YouTube digest 2026-06-02 (6 worth watching)

[JustinPerea]

Bookmark Queue Append

https://www.youtube.com/watch?v=hqHC6Z_lXyo
https://www.youtube.com/watch?v=BM2JX9hqsVQ
https://www.youtube.com/watch?v=u-rJwPPU3QA
https://www.youtube.com/watch?v=zqUdtL5l9yM
https://www.youtube.com/watch?v=oNRrjbKiF04
https://www.youtube.com/watch?v=KiJcZlqJKuE

Copy block above into ~/Documents/Obsidian Vault/Bookmark Queue.md, then run /process-queue.

Preview

20 days of compute vs 7 hours: rethinking what state-of-the-art means — Bertrand Charpentier, Pruna — AI Engineer

TL;DR: Pruna's chief scientist reframes "state of the art" as a Pareto front between quality and efficiency — and shows the headline model is the wrong choice for ~40% of real use cases.

What if the network was the sandbox? — Remy Guercio, Tailscale — AI Engineer

TL;DR: Tailscale's Aperture moves credentials out of the agent runtime entirely — the agent holds a placeholder; the network injects the real key at the WireGuard layer.

How to talk to statues — Joe Reeve, ElevenLabs — AI Engineer

TL;DR: A two-hour Sunday vibe-coded multimodal voice app — phone + OpenAI deep research + ElevenLabs voice design — went to 1.5M impressions and inbound from museums, auction houses, and travel platforms.

Full digest

daily/2026-06-02-youtube.md

Transcripts unavailable today (IpBlocked on transcript API; SSL cert chain failure on yt-dlp fallback) — same egress-proxy issue as recent days. Summaries lean on publisher-provided descriptions and chapter markers. The AI Engineer talks had unusually rich descriptions, so the WATCH summaries hold up well.

https://claude.ai/code/session_01JuTV563DMrDEsPWWmLwJcw

---

Generated by Claude Code