Vishrut Jha
@vishrutkmr7
iOS, full stack dev. Building something 🆒 in 🐦🔥 heat! Visca el Barça 💙❤️ | 🇺🇸
ID: 67285883
https://www.vishrut.co/ 20-08-2009 10:59:11
25,25K Tweet
629 Takipçi
486 Takip Edilen
Aakash Gupta This post banks quite significantly on a very weak premise that LLM inference is expensive and at scale it's not gonna work out. The reality is it's come down atleast 100x in the past two years and continues to drop. Coming even this close to Google in 3 years is a significant