Oddbean new post about login | logout If you have a PC, you can run all the cutting-edge language models locally with just one click in your browser. Tiny models also work on Android Chrome, and support for iOS will be added, at least for new iPhones. Temporary demo: https://iefan-ws.web.app/ https://image.nostr.build/ca976421700bb3591eda64d92667fbd108888b875122f23fb333fa81b6897ecf.jpg#m=image%2Fjpeg&alt=Verifiable+file+url&x=443ce0688c16f2c26bd753ba6c16a4e4e2de207427bd16d9ca746206ed1ab00f&size=71841&dim=980x1920&blurhash=%5D55%7Dv%7DsqNFV%5BM%7B%5E%2CtjMyX5M%7B_Nr_S0V%5BRjxbo%7BVubYWBtQepkUaMkBD%25X4sEbZofInVutQaMt7DiX4t8bYofInV%5BtRWBof&ox=ca976421700bb3591eda64d92667fbd108888b875122f23fb333fa81b6897ecf
Wow it’s working great!
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.
