“It’s not the manager’s job to prevent risks. It’s the manager’s job to make it safe for others to take them.”
— Ed Catmull, Creativity, Inc.
Today’s newsletter is brought to you by Stripe.
Watch four talks on AI pricing and usage-based models. Leaders from Replit, Fin, Snowflake, Clay, and Growth Unhinged discuss metering, variable spending, revisiting pricing, and forecasting AI costs and returns. Recorded at Monetize 2026, Stripe and Metronome’s pricing and billing conference.
Every Pixar film you love was, at some point, bad.
That is not an insult. It is the central idea of “Creativity, Inc.,” the book by Ed Catmull, who co-founded Pixar and ran it, and later Walt Disney Animation Studios, for more than three decades. Early versions of the films were confused, flat or broken. What made Pixar different was not that its first drafts were good. It was what the company did with drafts that were not.
When did you last show someone your work before it was ready? Most of us wait until it is polished, because showing it early feels like exposure.
A book about the people, not the hits
We expected a book about creativity: how Pixar finds ideas and makes hits. What we got was mostly about running an organisation. That surprised us, and then it made sense.
Stories, ideas and taste cannot be measured the way sales or uptime can. You cannot put a number on whether a scene works. As we read him, that is why Catmull put his attention on the things he could shape: the people who make the work, and the process they use to make it better.
The principle he returns to again and again is that the team matters more than the idea. In the publisher’s summary of the book’s lessons: “Give a good idea to a mediocre team and they will screw it up. But give a mediocre idea to a great team and they will either fix it or come up with something better.”
The Braintrust
Pixar’s best-known habit is the Braintrust. At points along the way, a director shows the current cut of a film in progress to a group of fellow directors, storytellers and studio leaders. They watch it and then say, candidly, what is not working.
The design has one rule that matters more than any other. The Braintrust has no authority. It cannot order changes. The director listens, takes the notes away, and decides what to do with them.
That rule is what makes candor possible. Nobody is defending territory, because nobody can take the film away. The notes are offered as help, not as a verdict. Catmull’s point, as we read it, is that honesty this direct does not happen by itself in any group of people. It has to be built into how the company works, so it does not depend on one person’s courage on one day.
The Beast and the Ugly Baby
Catmull has two images for the pressures inside a creative company. The Beast is the constant demand for output: schedules, deadlines, the next release. The Ugly Baby is every new idea in its early form, awkward and easy to dismiss.
The Beast is always hungry, and it will happily eat the Ugly Baby. Catmull’s job, as he saw it, was to protect the new and fragile long enough for it to grow up.
He applies the same thinking to mistakes. Problems will happen, so the goal is not to avoid them but to find them early and fix them together. Again from the publisher’s summary of the book: “It’s not the manager’s job to prevent risks. It’s the manager’s job to make it safe for others to take them.”
The founder behind the book
Catmull rarely talks about himself in the book, but his resolve shows through. He is not a loud founder. He comes across as someone with a quiet, strong core and deep passion underneath. Of all the founders whose stories we have read, he is one of the easiest to relate to.
There is also a surprising portrait of Steve Jobs, who owned Pixar for years. At Apple, Jobs was famous for imposing his will. At Pixar, by Catmull’s account, Jobs agreed never to sit in on Braintrust meetings, because his presence would change the room. Seeing Jobs through Catmull’s eyes is one of the book’s quiet pleasures.
If you would rather hear him tell it, Catmull walks through these ideas in his 2014 talk at Stanford, and Glasp has the video with a summary and transcript.
The Glasp Take: Share the Ugly Draft
The Braintrust is, at heart, a system for sharing notes. Someone shows unfinished work. Others respond honestly. The owner keeps control and decides what to use. Everyone gets better, including the people giving the notes, because they see how others solve problems.
Most knowledge work does not have a Braintrust. People read, think and draft alone, and share only the finished version. The ugly baby never leaves the room, so nobody can help it grow.
Glasp is a small step toward the other way of working. Glasp stands for Greatest Legacy Accumulated as Shared Proof. When you share your highlights and notes while you are still reading, you are showing work in progress: what caught your eye, what confused you, what you are not sure about yet. Others can see how you are thinking, and you can see how they are. Nobody has authority over your notes. You decide what to take from theirs. It is the same reason we keep saying you should learn in public, before you feel ready.
One more idea from the book applies here, in our reading: a process is there to help people think, not to think for them. Sharing notes does not replace reading. It just means you do not have to read alone.
The safe room in the AI era
Many of us now get our first notes from an AI. We paste in a draft and ask what is wrong with it. That can help, and Catmull’s principles suggest how to use it well. What follows is our reading of them, not something he said.
Keep the authority with the author. The Braintrust works because it cannot order changes. Treat AI feedback the same way: notes to weigh, not a verdict to accept.
Protect the ugly baby. A model learns from what already worked, so its edits tend to pull a strange early idea back toward the familiar. If a draft feels new, do not let the first round of machine edits sand it down.
Keep people in the room. An AI can tell you a sentence is unclear. It is less likely to tell you that you are avoiding the hard part, the way a colleague who knows you can. Use the machine for speed, and keep a small room of humans for candor.
A question for you
What are you working on right now that you have not shown anyone yet, because it is not ready? Who would you show it to if you knew they would be honest and had no power over it? Tell us in the comments.
Forwarded this email? Subscribe to Glasp Newsletter
Start highlighting the web with Glasp
Would you like to take Glasp on the go?
With the Glasp mobile app, you can highlight and organize your favorite content anytime, anywhere. Stay productive on the move and never miss an insightful quote.
Partner With Glasp
We currently offer newsletter sponsorships. If you have a product, event, or service you’d like to share with our community of learning enthusiasts, sponsor an edition of our newsletter to reach engaged readers.
Catmull's Braintrust works because the notes come with no authority: the director decides what to keep.
Where do you get honest notes on unfinished work today, and who gets to see your "ugly baby" before it's ready? I'd love to hear in the replies.
If a line in this essay stuck with you, highlight it here and it will be saved to your Glasp library:
https://glasp.co/save?u=https://read.glasp.co/p/candor-needs-a-safe-room
I love how you tied Catmull’s ideas to both AI and Glasp in a way that feels grounded in real creative struggle rather than abstract theory. The contrast between the Beast and the Ugly Baby really stuck with me, especially your point that most knowledge work never lets the ugly baby leave the room. The framing of AI as part of a Braintrust that must never hold authority over the author feels like a genuinely useful mental model for how to work with these tools without losing what makes risky ideas special.