Andrew Warner opens this conversation with Lex founder Nathan Baschez with a blunt theory: a pioneering AI writing product was overtaken as ChatGPT and Claude absorbed similar capabilities. Nathan agrees that competition from general-purpose tools was a large part of the story. But he does not describe a simple case of one company copying Lex and ending it. The customer, the job the product did, and the economics of that job mattered too.
Nathan says Lex is winding down by the end of 2026. It helped writers think through and improve their own work; it was not mainly a machine for generating high volumes of copy. That distinction made it a loved product for some writers, but a harder standalone business once broad AI apps could also converse with a draft. He also names a deeper challenge: helping someone make good writing better is often a less urgent, less measurable purchase than helping them do work they could not do before. This is his interpretation in an interview, not a full financial postmortem.
Watch the Interview
Credit: Andrew Warner's interview on The Next New Thing and the supplied transcript are the source for Nathan's reflections. The episode is presented by Zapier. Our founder questions and reader precautions below are editorial additions.
Lex Was Built for the Writer, Not Instead of the Writer
Nathan says he started with a frustration familiar to anyone who writes in Google Docs: the editor did not feel designed for the way he wanted to think. Before ChatGPT launched, he used that blank page as a place to experiment with AI. The goal was not the formulaic production of ad copy. He wanted to keep the writer in the center of the page and put a research and feedback partner beside the draft. At 03:18, he describes being able to ask a question without disappearing into a search detour.
That original purpose is documented in Nathan's 2023 origin and funding essay and is still visible in Lex's feature guide: feedback, comments, revision history and AI checks around a human draft. A 2022 nights-and-weekends project inside Every became an independent company in August 2023, when Lex announced a $2.75 million seed round led by True Ventures. Nathan rounds that figure to $2.8 million in the interview and says the company also generated revenue. He does not disclose enough financial data here to calculate burn, retention or the full economics behind the decision to wind down.
Why the Market Became Harder
General AI became a good-enough alternative
What Nathan says
ChatGPT made it harder to explain why a writer needed a separate AI writing workspace. Andrew asks whether newer chat and document features displaced Lex; Nathan says that is a substantial part of it.
What it does not prove
The interview does not show feature-by-feature copying, a user-loss curve, or a single event that closed the business. Nathan contrasts writing with coding, where specialized products such as Cursor and Claude Code still grew beside general assistants.
Better craft is harder to price than new capability
What Nathan says
Lex often helped someone spend more care on a piece of writing. Coding tools, by comparison, can save obvious time or enable projects a customer could not otherwise attempt.
Founder question
Can a buyer name a costly problem, budget and repeat usage for this improvement? Some writers valued Lex deeply; that does not automatically make the segment large enough for a venture-backed standalone company.
The audience had mixed feelings about AI
What Nathan says
Many writers see writing as thinking and worry that AI will produce disposable prose or undermine their identity. Nathan shares those concerns while finding AI valuable as a sparring partner and idea generator.
Founder question
Does the product serve people already eager to use the technology, or does every sale require resolving a philosophical objection? The latter can be a real market, but it needs a clearer message and stronger evidence.
What Nathan Would Tell a Founder
Andrew tries to distill two conventional lessons: sell into a valuable market and target customers enthusiastic about AI. Nathan accepts both, then resists turning his experience into a formula. At 08:51, his advice is to follow the work you are genuinely curious about without assuming it will become a large business. He does not say founders should ignore the market. He says obsession makes people build and learn, while only real use and payment can validate a company.
He also does not present the seed round as an obvious mistake in hindsight. Given what the team knew at the time, Nathan says he does not think the journey would have unfolded very differently. He remembers building with the team, shipping ambitious work and hearing from people the tool helped. The interview ends with a sober distinction: Lex was a real product with users and revenue, yet continuing as an independent business no longer made sense.
A practical reading for builders: pursue the interesting idea, then test its business case early. Ask who would be upset if the product vanished, what they currently pay or spend time on, whether the work is frequent, and why a broad assistant plus a document editor is not enough. These are our validation questions, not metrics Nathan reported for Lex.
For People Still Writing in Lex
As of this article's date, Lex's site remains accessible, and its pricing page is labeled for existing users only. Nathan says Lex will wind down by year-end but gives no exact cutoff, export format or transition schedule in this interview. Keep independent copies of important drafts, note which shared documents collaborators rely on, and follow direct communications from Lex for the actual migration steps. Do not assume that a public page or an old subscription FAQ settles those details.
Video Chapters
| Time | Conversation |
|---|---|
| 00:00 | Why Lex is winding down |
| 01:12 | Nathan on the future of Lex |
| 02:42 | The original idea behind Lex |
| 03:18 | AI as a thought partner for writers |
| 04:03 | Saving time versus improving a draft |
| 05:33 | How ChatGPT changed the market |
| 06:18 | Choosing the right market |
| 07:48 | Lessons for AI founders |
| 08:15 | Value and customer enthusiasm |
| 08:51 | Follow curiosity, not a formula |
| 10:21 | Should Lex have stayed a side project? |
| 11:33 | Build what you love without assuming an outcome |
| 12:36 | Why experimentation builds judgment |
| 13:57 | The best moments of building Lex |
| 14:36 | The seed round and revenue |
| 14:51 | Why Lex could not continue independently |
Sources and Limits
- Primary interview: Andrew Warner with Nathan Baschez, with the transcript provided for this post. The winding-down plan, competitive interpretation and retrospective lessons are Nathan's account; they are not an audited financial analysis.
- Original company history: Nathan's Lex origin and $2.75 million seed announcement, plus Lex's product guide.
- Current product status: Lex homepage and pricing page, checked 25 October 2026. Neither page provides the detailed end-of-service migration plan described as forthcoming in Nathan's announcement.