I'm building KineVision, a sport video analysis app for iOS. Something like Coach's Eye.
I wanted to have a more difficult project and learn more about AV and Liquid Glass. I try not to vibe-code.
I keep it 100% free for now, I'm looking for power users and retention. I won't monetise until I feel that this is a helpful tool.
I don’t like typing so I’m experimenting with camera, and screenshot organizing for productivity.
It’s my weekend project, but I’m trying to update at least once a month. So far works for me for collecting gift ideas or stuff I see in real life and want to check online later.
from landing page "We do not log or associate searches with an account"
I guess it's nothing technological like FHE, but rather how they set up their system.
ai accelerator market is so tough... When we were designing ai pipeline for video processing at my last company, we went through a lot of different accelerators during research...
most of the companies won't even send you a dev kit unless you show promise to buy thousands.. Google coral was different, of the shelf, easy to buy, plug and play, python API.
It has a few problems though. It's not very fast for today's standards.The list of supported operators aged quickly. On top of that it is a bit energy hungry, 2 TOPS per watt + random spikes.
I hope they are working on something new, but I also won't be suprised if they kill it, for the last year the devices are constantly out of stock.
https://axiomsskateboards.com