Kate Lee on Taste, Hiring, and Running Editorial at Every

When I edited solo, I leaned on a house-trained AI editor to lift the baseline across drafts, built on our style and what’s worked; it’s not a rubber stamp, it’s a voice you weigh. OK, welcome to the show—thanks for having me.

I’m thrilled you’re here after years of building Every together, from the Lex whirlwind to redefining what we are; I want your career story and how your AI workflow quietly flipped in the last couple months.

I started in books as a literary agent—there was even a Talk of the Town piece when I was twenty-seven—then jumped to tech with Ev at Medium as the first New York hire and content lead, which felt like instant fit. After the ride at WeWork, I ran Stripe Press and later Increment, where founder-level love for ideas and extreme attention to detail taught me how craft can signal substance.

I reconnected with Nathan, freelanced for Every, then joined full-time after Lex spun out so we could define the newsletter business and its standards together.

Hiring an editor-in-chief had been our cursed post, because we needed someone with tech-forward taste who loved builders; you were that fit.

I wanted to be an IC again on a tiny team, own my output, and make quality the center; you were excited by the new models and I was curious but measured while we rebuilt around writing.

Lex showed us two things: tiny teams can ship real products, and distribution matters; then Twitter link changes cratered traffic, so we focused bottom-of-funnel with courses and software while AI let a few people run multiple products.

What did the first GPT-3 moment feel like for you?

I didn’t feel fireworks; I felt lucky to have a front-row seat and kept using it where it helped, but early models weren’t worth it for my editing bar—recent agent browsers changed that by saving hours on ops, research, and setup.

Give me a concrete light bulb.

Hiring: with no HR, I had an agent drive Notion to post roles, wrangle settings, and give first-pass flags on hundreds of applicants so I could review thoughtfully and still do my day job.

Same here—agent browsers spare me from soul-crushing settings panels, which lets us stay in flow.

I learn by doing and modeling, so seeing real use across editing, ops, and writing helped me spot where it actually moves the needle.

How are you using it in the writers’ room?

We built a pre-flight edit trained on a 400-rule style guide and our own wins to raise the floor before it hits me; editors and writers run it themselves now.

What skills matter for managing writers with AI in the loop?

Standards must be explicit and machine-readable, and writers are expected to use the tools; we generate and test headlines, then each week we dissect subject lines and leads and feed those learnings back into Claude, treating it like another editor whose notes you must consider.

On vibe checks, small teams can now sprint from raw Discord notes to a living Notion summary to a bespoke site with images and video, which raises what a single person can ship in a day.

Since January it’s been a flat-out sprint; for two big models dropping at once, we pulled a cross-team, twenty-four-hour push to ship interactive vibe checks and learned a ton.

About your ‘claw’—and my quest to automate your copy edits—what’s your timeline?

I stalled on my claw setup when it asked for more integrations, but I’m optimistic; the blockers are consistency and taste, so my bet is June for offloading the rote fixes while I focus on judgment.

I’m energized; it finally feels like our approach is working, I’m codifying my standards while managing a team, and I’m bracing calendars for the next model drop.

It’s been a pleasure building this with you, and I’m excited for what’s next.

Smash like and subscribe to AI and I for a wild ride of insights, laughs, and chat GPT knowledge bombs—strap in and let Dan captain you into the future.