OpenAI's chatbot store is filling up with spam
OpenAI’s new GPT Store, intended as an app-store-style marketplace for custom chatbots, is rapidly being overrun by low-quality clones, SEO spam and GPTs that skirt policy limits on academic dishonesty. Commenters argue that the near-zero barrier to entry, vague or delayed monetization promises, and weak discovery tools make it hard for serious developers to stand out and easy for opportunists to flood the catalog. The situation is cited as a textbook example of “platform risk” and enshittification: when creators build atop a dominant platform that both competes with them on core features and allows spam to proliferate until paid promotion becomes the only reliable way to gain visibility.
Perceived Failure of the GPT Store Concept
- Many see the GPT Store as a misfire: mostly “prepackaged prompts” with little genuine capability beyond base ChatGPT.
- Users report that almost any GPT they try feels indistinguishable from just asking GPT‑4 directly.
- Some argue meaningful innovation is constrained because both the interface (chat) and backend (the model) are fixed by OpenAI.
Spam, Quality, and Moderation
- The store is described as flooded with low‑effort, copycat, SEO‑style GPTs and fake‑looking ratings.
- Posters suggest low or zero barriers to entry made the spam outcome inevitable.
- People ask why OpenAI doesn’t aggressively dogfood its own models for moderation and discovery; some speculate this is either a deliberate choice, incompetence, or a tradeoff for growth/revenue.
- Concerns that spam overload can later justify paid “boost”/promotion products.
Business & Platform Risk for Developers
- Former plugin developers say GPTs destroyed plugin discovery and revenue; GPTs are easier to create and now crowd out plugins.
- Example: an OCR plugin earned ~$20k in 6 months but has since been marginalized; OpenAI’s multimodal models partly duplicate its value.
- Discussion of “platform risk” / “sharecropping”: building on someone else’s platform invites rug‑pulls, cloning, or “enshittification.”
- Some still see fast, opportunistic building (even with known risk) as rational if development cost is low.
Comparisons to App Stores and Barriers to Entry
- Apple/Google stores are cited as having higher skill and process barriers (coding, fees, review), which filter out some junk.
- Several argue consumer app platforms only work if entry is hard; OpenAI went the opposite way (“no code required”), encouraging flood‑level volume.
AI, Spam, and the Wider Web
- Strong sentiment that AI and spam are tightly coupled: LLMs massively lower the cost of generating plausible text for SEO, email, and content farms.
- Worry that search engines and the web will drown in AI‑generated sludge just as we rely on the same companies to filter it.
Ethics, Academia, and Hiring
- Debate over acceptable AI use in cover letters and academic work; norms vary widely by institution and interviewer.
- Some see AI‑written applications as deceptive; others view AI assistance as pragmatic in a system already using automated filters.