Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I know this is your baby and building it all out on a raspberry pi is pretty cool, but I feel like that could be a bottleneck long-term.

First of all customers will need a physical device so you have to figure out shipping. Second, the RPi puts a limit on your computation. If you ever want a bigger and better model, you’re out of luck.

Why not just make it into an iPhone/Android app? Either do the processing locally or use the app as a thin client for a SaaS ($$$). Hell, people are running stable diffusion on an iPhone these days.

If it’s an app, people can easily try it out, and if it works for them you have a new customer overnight. Plus you don’t have to deal with pain in the ass RPi supply chain issues. Maybe it already exists as an app in which case ignore everything I said.

So personally what I would do is spend $10k porting to a mobile ecosystem and go back to YC saying you need cash for a couple mobile devs and maybe a temporary designer. If you’re comfortable doing the ML yourself you can save on that too.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: