In the latest round of #MLPerf industry benchmarks, the first submission of NVIDIA Blackwell delivered up to 4X more performance on #Llama 2 70B, while the NVIDIA H200 Tensor Core GPU continued to deliver outstanding results across every data center test.
TECHNOLOGY
-
Engineer Builds Mechanical Hand Prosthesis Without Power Requirements
By
–
Este ingeniero está construyendo una prótesis para su mano que no necesita energía. Un sistema 100% mecánico que puede mejorar su calidad de vida.
— Juan Merodio (@juanmerodio) 28 août 2024
Sin duda un trabajo digno de admiración 👏 pic.twitter.com/5tgFQPN5HdEste ingeniero está construyendo una prótesis para su mano que no necesita energía. Un sistema 100% mecánico que puede mejorar su calidad de vida. Sin duda un trabajo digno de admiración
-
Cerebras Debuts World’s Fastest Inference Technology
By
–
It was an interesting couple of days hearing from Cerebras about their move into inference. I think the company made a compelling pitch for different technology needed to enable different kinds of Gen AI queries. AI startup Cerebras debuts ‘world’s fastest inference,’ with a
-
Gemini could receive model upgrade
By
–
This is one of the possibilities that Gemini will get a model upgrade
-
ChatGPT MacOS Update: Selection of Custom GPTs
By
–
ChatGPT app for MacOS got updated and it seems that now you can select custom GPTs by typing in the Chat bar widget However, they are not working if you select to open chats in Companion mode (opening custom GPTs in a main chat works fine)
-
FastHTML Framework Size Comparison with FastAPI
By
–
I believe that FastHTML is far smaller than FastAPI. It’s definitely not a big framework.
-
SpaceX Starlink offers free emergency mobile services worldwide
By
–
After thinking it through, SpaceX Starlink will provide emergency services access for mobile phones for people in distress for free. This applies worldwide, subject to approval by country governments. Can’t have a situation where someone dies because they forgot or were unable
-
SambaNova Achieves 114 Tokens Per Second Record with Llama
By
–
I've been playing with @SambaNovaAI
's API serving fast Llama 3.1 405B tokens. Really cool to see leading model running at speed. Congrats to Samba Nova for hitting a 114 tokens/sec speed record (and also thanks @KunleOlukotun for getting me an API key!) -
Atomic Human Book Release in US and Canada
By
–
Atomic Human available in the US and Canada from 3rd September.
-
Tesla SpaceX Collaboration: Space Ride Hailing Innovation
By
–
Potential Tesla/SpaceX collab: ride hailing works even if you’re in space!