On: Show HN: I Built an AI Recruiting Toolkit Without ...

[mnahkies]

I'm curious how your COGS (cost of goods sold) are going to line up with the one time payment/lifelong subscription model.

ML Inference is pretty expensive in my experience - it's not clear to me whether you're running models yourself or paying someone like openai to do so, but either way GPUs aren't cheap

[kranzelbart → mnahkies]

We ran a token model for some time for precisely this reason, but came to the realization that the costs were comparatively low. We use OpenAI in the background.

[mnahkies → kranzelbart]

Fair, for what it's worth I think the one off model is better for the consumer so if you can make the numbers work that's great!