OpenAI is narrowing its focus on things that make money
TL;DR
Over the past year OpenAI experimented broadly: video platform, shopping portal, even AI erotica. Now the company is pivoting hard toward revenue.
Key Points
- CEO Sam Altman announced the erotica feature last October after reports of declining time-on-site for ChatGPT.
- Testing reportedly failed to reliably filter out references to bestiality and incest, according to the Financial Times.
- OpenAI is retreating from risky consumer features and doubling down on business tools — just as competition from Anthropic intensifies.
- Enterprise customers want text generation and agent workflows, not erotic chatbot experiences.
Nauti's Take
The fact that OpenAI even announced an erotica feature — only to have it fail moderation tests for bestiality and incest references — is not exactly a confidence-inspiring story. The underlying motivation was clearly panic: declining engagement ahead of an IPO is a serious problem, but launching edgy features is not the solution.
Refocusing on enterprise tools and agentic workflows is the right call strategically, though it arguably should have been the priority all along. Anyone watching the next wave of flashy ChatGPT feature announcements should pay close attention to which ones actually ship.
Context
OpenAI faces a classic tension: consumer reach drives attention, but enterprise customers drive predictable revenue. With an IPO on the horizon, the pressure to demonstrate scalable business models is intense. Pulling back from controversial features signals a preference for safe, recurring revenue over viral but risky experiments.
Meanwhile Anthropic is gaining ground in the enterprise segment — OpenAI cannot afford to cede that territory.
Sources