Vibe Check: Claude Cowork Is Claude Code for the Rest of Us (1h 33m)
- Release date: 2026-01-13
- Listen on Spotify: Open episode
- Episode description:
Anthropic just dropped Claude Cowork—essentially Claude Code for everyone, not just engineers—and we got to chat about it with a product engineer at Anthropic who helped build it.In this live Vibe Check, Dan Shipper and Kieran Klaassen explore the new interface together, testing what works (and what doesn't) in real time. Anthropic’s Felix Rieseberg joins midway through to explain the philosophy behind Cowork's design: why it separates "Tasks" from "Chats," how the queue system lets you send messages while the agent is working, and what "agent-native" architecture means in practice. They also dig into Skills—Claude's prompt system that lets you customize how it works—and the Chrome connector for browser automation.This is a raw, unfiltered first look at what might be the future of how knowledge workers interact with AI: async workflows instead of turn-by-turn chat.If you found this episode interesting, please like, subscribe, comment, and share!Want even more?Check out Dan's guide to building agent-native applications: https://every.to/guides/agent-nativeTo hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribeFollow him on X: https://twitter.com/danshipper00:01:00 - What is Claude Cowork00:02:36 - First demo: competitor analysis00:03:33 - Email drafting that sounds like me00:06:18 - Calendar audit running for an hour00:07:39 - Book taxonomy demo00:08:42 - PostHog analytics via Chrome browsing00:14:36 - Chat vs Code vs Cowork: when to use what00:31:06 - Felix from Anthropic joins00:36:39 - Why they built it in a week and a half00:37:57 - Design decision: why a separate tab00:43:57 - Skills as the primary hackable surface00:49:36 - Agent-native architecture principles00:56:57 - The origin story of skills at Anthropic01:03:00 - Our final rating
Summary
- 🚀 Claude Cowork Launch: Anthropic’s rapid sprint delivers Claude Code-like async agents for non-coders, running long tasks on users’ computers via browser control.
- 💻 Key Demos & Capabilities: Handles competitor analysis, email drafting, calendar audits, analytics scraping, book taxonomies, and research with persistent execution and queuing.
- 🗣️ Felix’s Insights: Product lead emphasizes experimental tab for fast iteration, skills for hackability, and future of generalized ‘I want’ interfaces over specialized UIs.
- 🔧 Skills & Agent-Native: Skills enable personalization; agent-native apps prioritize parity, granularity, and composability for emergent capabilities.
- ⭐ Yellow-Green Rating: Promising idea (green) for paradigm shift in async work, but janky execution (yellow) with UX rough edges; poised for quick improvements.
Insights
How is Claude Cowork democratizing agentic AI for non-technical users?
Time: 0:11 – 5:31
Category: AI in Everyday LifeAnswer: Claude Cowork extends Claude Code’s long-running, async agent capabilities to non-coders via a friendly UI, allowing tasks like website analysis, email drafting, calendar audits, and data crunching on the user’s computer without coding knowledge. This shifts users from quick chat responses to handing off complex work and returning later, accelerating everyday productivity. It bridges the gap for non-technical teams in growth, consulting, and research. (Start at 0:11)
Why does Claude Cowork’s async task queue represent a paradigm shift in AI interaction?
Time: 2:31 – 7:25
Category: AI in Everyday LifeAnswer: Unlike standard chat where responses block new inputs, Cowork allows queuing tasks while the agent runs long loops on your computer, mimicking developer workflows for non-coders. This enables parallel, extended operations like hours-long browsing or analysis without interruption. It fosters a ‘hand-off and review’ mindset, essential for deeper work. (Start at 2:31)
Why ship experimental features like Cowork in a separate tab?
Time: 15:16 – 39:35
Category: AI Governance & LawsAnswer: A dedicated tab signals it’s a ‘construction site’ for rapid iteration, separating polished chat from bleeding-edge agentic experiments running locally. This intentional opt-in gathers user feedback to refine UX, like clearer computer vs. cloud distinctions. It enables weekly shipping, learning from users over isolated development. (Start at 15:16)
What role do skills play in making Cowork hackable and personalized?
Time: 20:13 – 45:02
Category: AI-Driven Innovation EconomyAnswer: Skills are Markdown files with prompts, guidelines, or scripts that auto-load into Cowork, enabling custom behaviors like Swiss design, 3D printing, or writing styles without rebuilding tools. Users create them via deep research in Claude, turning personal workflows into reusable agent instructions. This composability unlocks emergent uses, from VST plugins to copyediting. (Start at 20:13)
How are agent-native architectures redefining app development?
Time: 21:34 – 52:27
Category: AI-Driven Innovation EconomyAnswer: Apps like Cowork and Proof replace deterministic code with agents wired to UI, where buttons trigger prompts and tools ensure UI-agent parity. This allows non-coders to build via ‘vibe coding,’ combining low-level tools for emergent capabilities. Principles like granularity and composability future-proof apps against model improvements. (Start at 21:34)
What persistent challenges remain in agentic computer use?
Time: 28:24 – 77:33
Category: AI Bias & FairnessAnswer: Agents struggle with apps like Google Docs due to complex DOMs, failing at precise edits despite skills. Permissions, skipping user questions, and unclear local/cloud execution confuse UX. Improving browser control and UI feedback is key for reliable async work. (Start at 28:24)
Will AI interfaces converge to fewer, generalized ‘I want’ boxes?
Time: 38:14 – 47:25
Category: AI in Everyday LifeAnswer: Felix predicts fewer specialized UIs as models generalize, akin to Google’s unified search bar evolving from subproducts. Chat-like inputs persist, with agents handling tasks dynamically rather than hyper-specialized tabs. Power users stick to familiar tools like Excel despite niche alternatives. (Start at 38:14)