Shows which parts of an AI-native IDE users say matter most and fail most: context indexing, model routing and agent workflows. Useful for choosing or building coding tooling.
Key takeaways · AI-distilled
The authors found that 71% of discussion in Cursor's public forum concerned workflow-critical mechanisms such as context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → indexing, model routingSending each request to a model chosen by the difficulty of the task, rather than using one model for everything.Full definition → and AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → workflows.
Negative experiences clustered around usability, pricing and dependability, including installation and setup, usage limits and consumption, crashes, and latency.
Over the two-plus years studied, complaints shifted from early access and setup barriers toward pricing, robustness and keeping workflows reliable over time.
Terms in this piece · Glossary
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
model routing — Sending each request to a model chosen by the difficulty of the task, rather than using one model for everything.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.