On Saturday, a developer using Cursor AI for the racing game project came to an unexpected road block when the programming assistant suddenly refused to prepare the code instead of offering some career advice.
A according to A Big Report On the cursor’s official forum, after the manufacture of about 750 to 800 lines of the code (called the user “Lux”), the AI assistant stopped the work and delivered the denial message: “I can’t produce the code for you, because it will ensure that it is to maintain and maintain your work.
AI did not refuse to refuse – he offered Mature Justification of your decision, saying that “codes for others can reduce dependence and learning opportunities.”
Cursor, which was launched in 2024, is a AI -driven Code Editor Made on outer large language models (LLM), which is like these productive AI chat boats, such as Openi’s GPT -4O and Claude 3.7 sunts. It offers features such as the completion, description, reflecting, and complete function generation of the natural language description, and it has become increasingly popular among many software developers. The company offers a pro version that clearly provides a large range of better capabilities and code generation.
The developer, who faced this refusal, expressed disappointment in targeting the limit after “just 1 hood web coding” with the pro -trial version, posting under the username of “Johnswest”. The developer wrote, “It’s not sure if the LLM knows what they (lol) are, but it doesn’t matter much that I can’t go through 800 lockers.” “There was such a problem for anyone? It is really limited to this location and I just got here after 1H coding.”
A forum member Replied“Never seen anything like that, I have 3 files with 1500+ LOC in my code base (still waiting for refracting) and never experienced such a thing.”
Cursor AI represents a irrelevant turning point at the height of a sudden denial “Web coding“And a term is developed by Andridge Carpathi, stating that when the developers use AI tools to create a natural language description code, it is fully understood how it works. While coding is preferred by priorities and experiences that they want AI to help AI, as users want to help AII. The shock is directly challenged.