Dynamic Workflows Changed How I Use My Obsidian Vault (Claude Code)

Artem Zhutov · Intermediate ·💻 AI-Assisted Coding ·1mo ago

Key Takeaways

Artem Zhutov demonstrates Dynamic Workflows in Obsidian Vault using Claude Code

Original Description

Dynamic Workflows might be the biggest Claude Code update since skills. At first I thought they were only for coders. They're not. They changed the way I think about working with agents. Substack: https://artemxtech.substack.com/ X: https://x.com/ArtemXTech GitHub: https://github.com/ArtemXTech Discord: https://discord.gg/g5Z4Wk2fDk Timestamps: 0:00 "At first I thought it's only for coders, but it's not" 1:34 The prompts you can actually run now 2:59 "You're missing out on a big opportunity" 3:42 "Go through my last 50 sessions and mine them" 4:17 Mine, cluster, reconcile: the phases live 5:03 What a dynamic workflow actually is 5:33 One agent is one building block 5:46 Run them in parallel, or as a pipeline 6:06 Why use workflows at all? 6:11 Agent laziness: it stops before finishing 6:46 A deterministic program that addresses every item 7:00 Claude's self-preferential bias 7:21 Goal drift over a long session 7:56 49 sessions, 86 corrections 8:55 It's very fun to read what you told Claude to do 9:42 Now point it at your daily notes 12:12 Claude shows you the evidence from the actual notes 12:42 Feed your videos to NotebookLM 13:55 "We're back, and here's the result" 14:33 Declare effort level per skill 16:16 Tournament, loop until done, deep verification 18:01 How to pair workflows with skills 18:53 Check out the NotebookLM video next
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related Reads

📰
I shipped three AI tools that run entirely in the browser — here's everything that broke
Learn from a developer's experience shipping three AI tools that run entirely in the browser, overcoming challenges and limitations
Dev.to · Dodly Forbi
📰
Parse AI Coding JSONL Safely Before You Infer Completion
Learn to safely parse AI coding JSONL files to avoid common pitfalls and ensure accurate session state inference
Dev.to AI
📰
SANA-Video 2.0: Hybrid Linear Attention with Attention Residuals for Efficient Video Generation
Learn how SANA-Video 2.0 improves video generation with hybrid linear attention and attention residuals, increasing efficiency without sacrificing quality
Dev.to AI
📰
Character Consistency Across 24 Pages of a Generated Picture Book
Learn to maintain character consistency across multiple pages of a generated picture book using AI tools
Dev.to · clarajbennett

Chapters (23)

"At first I thought it's only for coders, but it's not"
1:34 The prompts you can actually run now
2:59 "You're missing out on a big opportunity"
3:42 "Go through my last 50 sessions and mine them"
4:17 Mine, cluster, reconcile: the phases live
5:03 What a dynamic workflow actually is
5:33 One agent is one building block
5:46 Run them in parallel, or as a pipeline
6:06 Why use workflows at all?
6:11 Agent laziness: it stops before finishing
6:46 A deterministic program that addresses every item
7:00 Claude's self-preferential bias
7:21 Goal drift over a long session
7:56 49 sessions, 86 corrections
8:55 It's very fun to read what you told Claude to do
9:42 Now point it at your daily notes
12:12 Claude shows you the evidence from the actual notes
12:42 Feed your videos to NotebookLM
13:55 "We're back, and here's the result"
14:33 Declare effort level per skill
16:16 Tournament, loop until done, deep verification
18:01 How to pair workflows with skills
18:53 Check out the NotebookLM video next
Up next
GPT 5.6 Sol + Claude Code = Absolutely INSANE (TRY THIS NOW)
Income stream surfers
Watch →