To give people a sense of energy scale… Listening to TVA talk about how they hit peak energy demand in Tennessee of 35 gigawatts and they were able to handle that. (10 million people) Kenya’s peak energy demand in 2024 was 2.2 gigawatts. (55 million people)
SYSTEMS
-
Critical Smart City Trends Reshaping Urban Life 2025
By
–
8 Critical Smart City Trends Reshaping Urban Life In 2025 Urban centers are undergoing a dramatic technological transformation, with #artificialintelligence, #digitaltwins, and #smart #infrastructure reshaping how millions live in mega-cities. From AI-driven #urbanplanning to
-
LLM Agents Production: Architectures, Challenges, Best Practices
By
–
LLM Agents in Production: Architectures, Challenges, and Best Practices – ZenML Blog https://
bit.ly/4gGcPtW
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
AWS Azure Serverless Egress Control Public Preview Launch
By
–
Egress control for serverless and model serving workloads is now in Public Preview on AWS and Azure. Centrally control outbound access from serverless workloads across multiple products and workspaces to minimize the risk of unauthorized or accidental data transfers outside
-
Scaling Reasoning Models: From Thousands to Millions of GPUs
By
–
This is the interesting bit. As we move to the next phase of scaling reasoning models w RL, data and compute converge. Next breakthroughs require moving from 100’000’s of GPUs to millions of GPUs.
-
MoE Efficiency and KV Store Compression Advances
By
–
MoE very efficient and compress context KV store. Both noted in today’s post
-
Qwen 2.5 7B Released with 1M Context Window
By
–
Qwen 2.5 7B just dropped with a 1M context window. Here it is running at 4bit @ 11 tok/s.
— Aaron Ng (@localghost) 27 janvier 2025
Should we add it to the next update? 1M of context adds a lot of possibilities. pic.twitter.com/y4uhLfHKNSQwen 2.5 7B just dropped with a 1M context window. Here it is running at 4bit @ 11 tok/s. Should we add it to the next update? 1M of context adds a lot of possibilities.
-
RL Model Achieves Superhuman Snark on Human Cope Chains
By
–
applying large-scale RL to chains of cope written by human seethers, we observe an “lmao moment” upon which the model spontaneously exhibits superhuman snark
-
Meta’s Manhattan-Sized Data Center for Industrial Intelligence
By
–
Here we go. Another (very big) building block for Industrial Intelligence from @Meta
. This follows @xai
’s #Colossus housing 100,000 @nvidia H100s, and covering 800,000 sq ft. #Stargate (‘monuments in the desert’ as per @sama
.) …Meta’s Manhattan-sized data centre. The
