Not everybody wants their client talking to Google (or OpenAI etc) - it needs to at least have an off switch
I have something even more fascinating, it's still underdeveloped but it definitely has a potential. nostr:nevent1qqsv0006qurr3wg84zgvqt3h50t0ehlyxm8vk0fe59j7efkphpwpkygpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qygxx7urh795e65x0j25k22l7hlavqh7xss4eacu3pzwetxu26h2gl5psgqqqqqqsts3h4e
In-device models can take care of private events. But otherwise, I am not sure they are that interesting for public events. Constantly updating the local model to match the newest events coming in online is challenging.
I was attempting that, and it's quite challenging. So, I will be working on it with Gemini. This will focus on private tasks like TTS, PDF summarizer, Basic Q/A etc., using small, task-specific language models with WebGPU.
this tbh