
OpenAI ライブストリーム
OpenAIがライブストリーム配信イベントを開催します。放送中に具体的な発表、新製品発表、またはデモンストレーションが明かされる予定です。
OpenAIが前回予告なしのライブストリームをやった時、GPT-4 Turboをドロップして、一晩で価格設定を完全に変えました

I was using Gemma’s models in the plane offline while reading a book, to aid my studying, and I thought about this. Think about it this way. Most phones can run Gemma’s 2B model (many the 4B too), and open source models will get cheaper and better, and CPUs more optimized for AI. Gemma 4B is almost on par with GPT 4o which was the #1 model at its time. I think in the future, it will work this way: Client: Request -> Local LLM using phone as server -> Server: Response -> Phone view (client)