Uncovered Bookmarks #21: Coworking at Airtree
Onboarding the team to Claude Cowork
Why is Cowork such an unlock for Anthropic? The clean UX removes the learning curve. Compared to using the terminal for the first time, it’s immediately accessible.
This week, we onboarded every Airtree team into Claude Cowork - Finance, Legal, Marketing, Investor Relations - and started building Skills together.
Some quotes from the team:
“honestly blown away by how easy this is to use. The biggest takeaway for me: it's going to be game-changing for our work. We have SO many repetitive tasks”
“The biggest unlock for me was seeing how it can read through any file type, organize everything in standard format, and do all the research in parallel before updating a CSV in one go. The batch processing is 🔥”
“Cowork is really user-friendly — it uses common sense and plain English to create these automation tools. No technical complexity, just describe what you want.”
“Biggest takeaway from the Cowork session: building skills (reusable instruction files) that Claude follows every time. Huge potential for month-end close”
In the first hour, we collectively automated a day of work each week.
There’s plenty we need to figure out to ensure:
We focus on outcomes, not just process automation
Teams build cross-functionally rather than in silos
Internal tools are tested and maintained
But we’re having a lot of fun on the journey together.
Links
X got a little existential this week:
“Since the industrial revolution, technology’s trajectory has been one of humans becoming more like machines. Modern work has quietly trained us to act like machinery with repeatable tasks, narrow lanes, endless throughput, constant responsiveness, and performance measured by the hour. You can make a good living being a dependable machine…and slowly forget you’re a person.
When AI starts doing the most machine-like parts of the work shockingly well, where does that leave us? Instead of making humanity more like a machine, I wonder if it does the opposite. Could the actual machines remove the “machineness” from the human? What if we’re living in a time of re-enchantment, of re-learning what it means to be human?
….
AI can imitate outputs. It can predict patterns. It can accelerate execution. But it can’t carry moral weight. It can’t love. It can’t repent. It can’t take responsibility. It can’t look another human in the eye and choose courage over approval. It can’t suffer with someone. It can’t forgive. It can’t be faithful. And it can’t answer the question underneath every technological leap, “What is this for?””
“The craftsmen who built churches worked for decades on stone carvings placed so high that no human eye would ever see them clearly. They carved them anyway. They carved for God. Those calling for new aesthetics without calling for the thing that has always produced great aesthetics are asking for the fruit without the root.”
Arc Institute continues to publish exciting new research:
Solar panels in the mountains:
Just because you can ship 1000x more, doesn’t mean you should. Prototype a lot, kill 9 of 10 ideas:
Gorgeous:
This wave keeps building:
Jared Sleeper’s account is one to follow for those wondering what to do with public SaaS:
