On-Device AI Revolution: Custom Hardware & Agent Management

The engineer in me is just begging to ask you some questions about your product and and also I want you to answer too like when at what point are these going to be commercially available to people because I don't think they are yet. But if I were to try to build something like what you're doing on a laptop, I would put a like a a whisper front end of, you know, voice to to text and then I would put like a small, you know, low latency LLM in place with like a four-bit quant, so it's fast. And then I would put like a really low latency text to speech engine, right? Maybe like a Kakuro or something like that that is really fast. Um, but if I were to do that, that's going to cost, you know, I'm going to have to have a pretty decent GPU or I'm going to have to have like an AMD with unified high bandwidth RAM or something. How are you going to be able to to combine these elements and I I know I just described a simplified version. Uh, but how are you going to combine these ...