AI Daily : The Latest in AI: Salesforce’s Next Generation of Einstein, NVIDIA’s AI GPU Shipment, and more!
This Newsletter is Curated By: AI || Reviewed By : Avijit || Date: 2023-09-18
Spotlight
– Salesforce launches next generation of Einstein, introducing a conversational AI assistant to every CRM application and customer experience.
– NVIDIA is reportedly shipping 900 tons of H100 AI GPUs this quarter, which amounts to 300,000 units.
– SoftBank seeks a tie-up with OpenAI as part of its deal spree after Arm IPO.
– Three Chinese firms, including Beijing Fourth Paradigm, aim to raise up to $280 million in Hong Kong IPOs.
– Microsoft researchers have developed EvoDiff, a general-purpose diffusion framework for tunable protein creation in sequence space.
MindBlowing AI Tools Everyone talking about
– vLLM is an open-source machine learning library designed to enhance the throughput of LLM serving systems.
– Researchers have introduced Rewindable Auto-regressive INference (RAIN), a unique inference technique for Large Language Models (LLMs).
– AGENTS is an open-source language agent library and framework that simplifies the customization, tuning, and deployment of language agents.
– Google is rolling out its third iteration of the Helpful Content Update, classifying content for search engines versus people.
– Würstchen is an innovative two-stage compression model for text-to-image generation.
What The World Have Researched & Innovated
– Researchers have unveiled Protpardelle, an all-atom diffusion model that generates high-quality proteins with reduced computational cost.
– PagedAttention, an attention algorithm, improves LLM serving systems by optimizing memory utilization and sharing.
– Biased medical data in AI systems should be seen as informative artifacts, revealing healthcare inequalities.