Story of a startup that closed after ChatGPT

Summary

Lex was an AI-assisted writing product built around a better writing experience. Users could write in a document while asking an AI assistant questions, getting feedback, and generating ideas. Founder Nathan Baschez said the product was motivated by dissatisfaction with Google Docs and a desire to experiment with AI before ChatGPT existed.

Lex is winding down as an independent business at the end of the year. Baschez is proud of the product and team, but concluded that continuing independently no longer made sense. A key challenge was that general-purpose AI tools such as ChatGPT and Claude added features that overlapped with Lex’s core experience.

Business Lessons and Tactics

• Prioritize substantial value over incremental improvement. Baschez’s central product lesson was that helping people do something they could not do before—or do it dramatically faster—can be a stronger proposition than making an existing activity somewhat better. He saw AI coding tools as an example of the former; Lex mostly supported and improved an existing writing process.
• Choose a market with clear willingness to pay and value. The interview contrasts writing essays or books with use cases such as marketing copy and coding, where the financial value of the output may be more direct. The interviewer suggested that ad-writing customers might be more likely to pay for dedicated software than essay writers.
• Consider customer attitudes toward AI. Lex targeted writers, some of whom view AI as a threat to creativity or as a source of low-quality writing. By comparison, the speakers described programmers as generally more eager to use AI to gain an advantage. Selling AI to a skeptical audience can make adoption harder.
• Be genuinely differentiated from general-purpose tools. Lex was specific to writing, but Baschez felt it was not sufficiently vertical or distinct to withstand competition from broad AI products as they added document and conversational capabilities.
• Build from genuine interest, then learn from the market. Baschez advised founders to pursue problems they are personally excited about rather than choosing ideas only because they appear to satisfy conventional business criteria. He recommended not assuming an idea is a business until evidence supports that, while continuing to adjust as the team learns.
• Treat experimentation as a source of judgment. Baschez argued that founders develop useful instincts by acting on them, learning from both successes and failures, and refining their judgment over time. Avoiding every idea because of its risks can prevent that learning.
• Value the work and relationships, not only the exit outcome. Baschez highlighted building with the team, making decisions, setting ambitious goals, shipping work the team was proud of, and hearing how the product helped users. He also noted that Lex’s work drew the attention of people at Notion, illustrating that a startup’s value can extend beyond its standalone business outcome.

Product, Operations, and Financial Details

• Lex positioned AI as a thought partner: writers could ask questions or seek feedback while continuing to write, rather than having AI generate an entire article.
• The product was aimed at people writing essay-like posts, newsletters, or books—not primarily at mass-producing marketing content.
• ChatGPT’s arrival and subsequent product additions were described as a major competitive pressure. The speakers did not frame this as proof that general-purpose AI eliminates every AI startup, but as a particularly difficult challenge for Lex’s product category.
• Lex raised a $2.8 million seed round and also generated revenue. Baschez said the funding lasted longer because the company had revenue, but no revenue, runway, or growth figures were provided.
• No specific performance targets, timelines beyond the planned year-end closure, or standard KPIs were stated.

Presenters and Sources

• Andrew Warner — interviewer
• Nathan Baschez — founder of Lex
• Zapier — sponsor mentioned in the interview