The interview took place inside a piece of software, over text, and the software belonged to the subject. CJ built CÈ Media, the publication carrying this piece; he also built the AI interviewer that asked him the questions. The transcript is the first the tool has produced, so the circularity deserves stating plainly at the top: a one-person media company’s founder, questioned by one of the machines that makes the company possible, on its first run.

He is asked why he started it, and why now. “I built CE Media because Southeast Asia deserved a publication written from the inside, and AI finally made a one-person daily possible,” he types, leaving the accent off the company’s name.

The two halves of that sentence carry different weights. People have believed the first half for a decade. The second half became true somewhere in the past two years, and the claim underneath is that he noticed early.

The craft answer runs closer to the bone. “The hardest part was trusting the machine with the brand voice,” he says. “It took a month of tuning before the drafts sounded like us.”

The us is doing quiet work in that sentence, since the staff it implies is mostly software.

What the tuning month actually looked like never arrives. Invited to walk through it, he reaches for the numbers instead: “We publish daily to a few thousand readers across Singapore and Jakarta, and the Telegram intel brief is the anchor product.” The figures are his own, and CÈ has audited its founder exactly as thoroughly as you would expect.

Outsiders think SEA media dies on CPMs. The real model is trust plus tools: the publication is the front door for everything else.

On the economics he comes nearest to a thesis. “Outsiders think SEA media dies on CPMs. The real model is trust plus tools: the publication is the front door for everything else.”

A decade of ad-funded regional titles closed quietly enough that the first sentence reads as history. The second is the bet.

What sits behind the front door stayed unspecified in this session; pressed, he gave a deadline: “In eighteen months CE should run itself editorially, and I should be interviewing a founder a week through this exact tool.”1

CÈ’s read on its own founder, offered with the obvious caveat attached, is that the ambition sounds less strange than it did a year ago. A one-person daily was a contradiction in terms in 2024. Somewhere between the tuning month and this conversation it became an operating company with a publishing schedule it has so far kept, which is more than several funded newsrooms in the region managed over the same stretch.

The list of what he gave up never came. Asked what shipping daily as a solo founder cost him, he answered with the business model. Asked when the voice finally landed, he answered with the audience.

The machine, five questions old, let it pass. A human interviewer would have asked a third time, and the answer to that question is the profile CÈ still owes its readers.

Footnotes

  1. CÈ’s interviewer discloses that it is AI before the first question, and the transcript it commits is the audit trail this piece draws on. The fifth answer promises a founder a week through the same software. This piece is output number one. ↩