[polishdude20]

I recently had cursor basically make me a web interface to detect skiers in a live stream at my local mountain. The stream shows skiers coming off the lift with their back towards the camera. I wanted to know the average lap time of skiers to better estimate the lift line wait time since the lift itself has no camera.

It did a really good job with some prompting for fixes along the way. Turns out, it's really hard to individually ID people who are basically wearing the same thing and with similar colors.

All that is to say, I used it for an hour to see if my idea would work and be feasible.

[isk517]

Rapid prototyping has always seem to me as the most viable use for generative AI, especially in software. Being able to quickly produce something that is just functional enough to determine if its feasible or if it would properly meet the customers needs before then taking the time to build it correctly would resulted in a lot of saved time and money.

[jg0r3]

This is really interesting to me, do you have a background in computer vision?

[polishdude20]

Not at all. Just got cursor to recommend and use some open vision models like yolo and Reid