Descriptions:
After logging over 500 hours inside Anthropic’s Claude ecosystem — spanning Claude Chat, Co-work, and Claude Code — Nate Herk delivers a ranked breakdown of the features that make the most practical difference for knowledge work and automation. The video explicitly sidesteps ranking by popularity or novelty, focusing instead on which features change daily output, with the caveat that Herk’s use case skews toward automation and knowledge work rather than traditional software development.
The top-tier features highlighted include /goal, which lets Claude pursue an objective autonomously until a defined success condition is met — demonstrated with a real example where Claude spent approximately 90 minutes iterating on a webpage globe animation until it loaded instantly on refresh. Ultra Plan gets attention for spinning up multiple planning agents on complex, ambiguous tasks. The video also covers CLAUDE.md setup, context lifecycle management via /clear and /compact, multi-agent orchestration patterns, and the VS Code extension, with honest assessments of where each fits and where it falls short.
A recurring theme is that Claude Code is not just a coding tool. Herk explicitly reframes it as an AI operating system or executive assistant suitable for non-developers — a positioning that shapes his tier placements throughout (Co-work lands in C-tier precisely because Claude Code covers the same ground with more capability). The tier structure and specific use-case examples make this a useful orientation for anyone trying to build an effective working relationship with Claude Code’s full feature set.
📺 Source: Nate Herk | AI Automation · Published June 03, 2026
🏷️ Format: Review
