suvamsh

I genuinely believe talking to machines is the future and I significantly reduced the amount of typing I do. So I took a few days to just build an open source one, Rust + whisper.cpp, Metal GPU acceleration, runs entirely on-device. The AI models are free, so why are we still paying for wrapper apps?

> The answer was always yes. The question was the speed and cost.

That's the right assessment, as a result of cost the final decision would be we cannot. The cost to at least prototype and experiment is dramatically lower which lets you try things you'd never have thought was possible.

> it cannot build the infrastructure to scale that app

Why do you say this? I haven't seen proof of this, inversely teams with already scaled infrastructure are able to launch greenfield at scale faster with AI.

We just launched our beta last week and are looking for feedback. It's currently free to use!

I'm curious to know why a chatbot sucks more than a web form in your opinion? In my experience web forms are static, tedious, not interactive and as a candidate can feel quite disrespectful that the company does not want to "talk" to me.

I agree that a candidate can be screened with a web form but this method isn't very popular as web forms can be quite boring and unpleasant to use. A chatbot can strive to replicate the human aspect that a recruiter provides. Also talking to a bot is probably more fun than filling out a form.

Resume parsing is it's own problem that has no universal solution and we aren't tackling this at the moment. Some candidates may prefer to apply with LinkedIn, GitHub, etc. rather than resume and hence I believe there are multiple places to obtain candidate information. As this is an MVP we haven't built this feature yet, but we are actively working on the next iteration.