Section
Infrastructure
The latest Infrastructure news, analysed daily by the i10X newsroom.
FilterPerplexity
-
pplx-embed-v2-late: Perplexity Late-Interaction Multimodal Embeddings
Discover pplx-embed-v2-late, Perplexity’s open late-interaction models that deliver OCR-free retrieval for PDFs and visuals. Use 9B for indexing and 0.6B on edge devices to optimize RAG performance. Learn more.
By
-
Perplexity Portable Computer on AMD Ryzen AI Max PCs
Perplexity and AMD enable local 27B AI agent workflows on Ryzen AI Max hardware. Run private, recurring tasks…
-
Perplexity Personal Computer: AI Agent on Mac Mini
Perplexity launches Personal Computer, turning Mac mini into a dedicated 24/7 AI agent for enterprises. Explore integrations with…
-
Perplexity AI’s CoreWeave Partnership: NVIDIA GB200 Inference
Discover how Perplexity AI’s multiyear deal with CoreWeave leverages NVIDIA’s GB200 Grace Blackwell superchips for superior AI inference…
-
On-Device AI: Aravind Srinivas’s Decentralized Vision
Discover Aravind Srinivas’s contrarian AI strategy at Perplexity, emphasizing on-device AI for privacy, speed, and human curiosity. Challenge…
More stories
-
Perplexity’s Playbook: Run Trillion-Parameter LLMs on Cloud GPUs
Discover Perplexity AI’s open-source optimizations for running trillion-parameter LLMs efficiently on commodity cloud GPUs. Overcome network bottlenecks with EFA and NCCL to scale AI without costly…