C’est un tournant majeur. À ce stade, est-ce qu’on ne serait pas déjà en train de franchir le seuil de l’Explosion d’Intelligence ? Voici l'analyse complète : https://
youtube.com/watch?v=K0v7oc
0RWn4
…
AI HARDWARE
-
GPU Performance for Large AI Models: Loading vs. Inference Speeds
By
–
-
Consumer GPU limitations for running large AI models
By
–
I'm not trolling. The original statement by @jeremyphoward was that it couldn't run on "consumer GPUs" (plural), and I genuinely didn't understand why. If the original statement was "it's too large to run efficiently on a single consumer GPU", then that would make more sense.
-
Silicon Molding Advances With New Mold Technologies
By
–
Molds of all sorts to do more silicon molding
-
10M Token Processing and Infinite Context Length Breakthroughs
By
–
Wonder what compute times we will see with 10M input tokens on a rather beefy system. The NiH tests are also insane, It’s hard to believe we are moving towards infinite token context lengths with near perfect retrieval.
-
Robots becoming more uncanny as technology advances rapidly
By
–
👀😩🤷🏽♀️ Robotics! Tech is an enabler. These robots are getting more uncanny, feels creepy. pic.twitter.com/m1dDW7ywsq#Robots #robotics #Robotech #technews #RoboticGang #tech #IoT #Engineering #Gadgets #ML #programming
— Catherine Adenle (@CatherineAdenle) 6 avril 2025Robotics! Tech is an enabler. These robots are getting more uncanny, feels creepy. #Robots #robotics #Robotech #technews #RoboticGang #tech #IoT #Engineering #Gadgets #ML #programming
-
Gemma 3 Open Source Models Run Single GPU TPU
By
–
Fwiw, this exact reason is why we made the Gemma 3 open source models something that developers could easily run on a single GPU or TPU.
-
STL Model Development: Inevitability of Commercialization
By
–
I think at least one person is working on an STL model but suspect it will be a while before anyone gets to commercializable results. (Thought of doing it myself; seems like not the thing to spend years of life on.) In principle it’s ~inevitable it happens.
-
Running Large Models on Consumer GPUs: Parallelism and Quantization
By
–
I don't really use GPUs, as most of our work uses TPUs. I was just confused by @jeremyphoward 's statement that it wouldn't run on consumer GPUs, when the normal techniques of model parallelism, quantization, right compression, etc. seem like they should work just fine to run it