Trying to develop the next huge LLM is fun, but how about a side project with a more reasonable scope like the 1 LLM + 1 GPU + 1 Day NeurIPS efficiency challenge: https://
llm-efficiency-challenge.github.io Below a list of the approved models (PS: Lit-GPT was chosen as the official starter kit )
AI HARDWARE
-
NeurIPS LLM Efficiency Challenge: Train Model in One Day
By
–
-
Cerebras-GPT: First Open-Source Compute-Optimal Models
By
–
Cerebras-GPT – Deploying the first open-source compute-optimal models – https://
hubs.li/Q01Z9tgQ0 -
Jensen Huang to Keynote SIGGRAPH 2026 on August 8
By
–
Save the date for our founder and CEO Jensen Huang's special keynote at #SIGGRAPH on August 8. https://
nvda.ws/3CRNLhO -
Portfolio Companies’ Massive GPU Cluster Long-Term Spending Overlooked
By
–
forgot the millions they spent on locking in large GPU clusters for 3 years for portfolio companies
-
AWS Launches NVIDIA H100 Instances to General Availability
By
–
in case you missed it @awscloud just shipped their nvidia H100 instances as GA yesterday only the 8xH100 flavor is available for now, so it's a bit hefty https://
aws.amazon.com/ec2/instance-t
ypes/p5/
… -
Hands-on Testing of New AI-Powered Phones Begins
By
–
We've got our hands on the phones. So rest assured, more hands-on tests / thoughts to follow!
-
Samsung Foldable Phones: Innovation in Mobile Hardware Experience
By
–
This is Samsung's opportunity to show off its best high-spec devices, and it's making a big splash with nifty features – like folding the phone halfway to watch stuff on the go – in an effort to get people to upgrade to the "folding phone" experience.
-
Major Players in AI, Hardware, and Aerospace: SpaceX, OpenAI, Nvidia, TSMC, Lockheed
By
–
spacex openai
nvidia tsmc
lockheed
… -
Neuromorphic Computing and the Potential for Hyper-Realistic AI
By
–
Neuromorphic Computing And The Potential For Hyper-Realistic AI
#AI #AIio #BigData #ML #NLU #Futureofwork @TopCyberNews @SpirosMargaris @MarshaCollier @MHiesboeck
@MHcommunicate @Fisher85M @MikeQuindazzi @NealSchaffer
http://
ow.ly/nJFf30swnqF -
Replicate Models Now Available on NVIDIA A40 GPUs
By
–
You can now run Replicate models on @nvidia
's A40 GPUs. For many models, A40s can be 80-90% as fast as A100s but much cheaper.