The Top 1% of Devs are Cranking Out 46x More Code Using Cursor
Cursor just dropped their data on how we’re actually using AI, and surprise: most of us are just playing pretend while the elite 1% are literally rewriting the internet. It’s a beautifully depressing look at the widening competence gap in the age of automation.
Average coding output has surged from 3.6k to 8.6k lines per week, while pull requests containing over a thousand lines of code jumped from 8% to 14%. Models are now reading significantly more context than they are outputting, with the input-to-output token ratio ballooning from 4.5x to over 11x, and cache reads now accounting for 90% of total activity.
While average users poke at the AI with roughly 145 tool calls per session, the top 1% of performers are operating on an entirely different planet. These individuals are churning out 46 times more code and closing 15 times more pull requests than their peers. Even when factoring in the 9x price disparity between cheaper models like Composer 2.5 and premium ones like Opus 4.7, the higher-end models maintain an edge because they actually get their suggestions accepted more often.
We are watching the death of the average coder and the birth of the AI-augmented cyborg class. When a single human can do the work of forty-six, the job market won't just shift—it will likely collapse into a bizarre meritocracy where your ability to prompt engineering beats your ability to actually understand the syntax. Enjoy the race to the bottom while the top 1% builds the future in their sleep.
Source: Cursor
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.