1Hours of Humanoid Teleop, Recorded in Real Homes (opens in new tab)(topicqueue.substack.com)3DISCURSIVE17h ago0Save
2Knowledge curation (not search) is the AI big data problem (opens in new tab)(daft.ai)3DISCURSIVE5mo ago0Save
3Cutting LLM Batch Inference Time by Half with Dynamic Prefix Bucketing (opens in new tab)(daft.ai)2DISCURSIVE7mo ago0Save
4Agentic systems are just query engines for unstructured data (opens in new tab)(daft.ai)3DISCURSIVE7mo ago0Save
5Benchmark: Spark vs. Ray Data vs. Daft on Multimodal Workloads (opens in new tab)(daft.ai)1DISCURSIVE8mo ago0Save
6Show HN: Flotilla, the Distributed Engine for Multimodal Pipelines (opens in new tab)(daft.ai)2DISCURSIVE8mo ago0Save
7Profiling multimodal workloads: lessons from Daft (opens in new tab)(daft.ai)1DISCURSIVE9mo ago0Save
8Processing 24T tokens for LLM training with 0 crashes (what made it possible) (opens in new tab)(daft.ai)1DISCURSIVE10mo ago0Save
9We Hit 100% GPU Utilization–and Then Made It 3× Faster by Not Using It (opens in new tab)(daft.ai)17DISCURSIVE10mo ago11Save
10Elasticsearch Was Great, but Vector Databases Are the Future (opens in new tab)(thenewstack.io)16DISCURSIVE1y ago7Save
11Have you faced any challenges of using Elasticsearch for vector search?We heard from many that it'll get expensive quickly, and the performance is suboptimal. Any thoughts?show more1DISCURSIVE1y ago0Save
13Enhance information retrieval with learned sparse embeddings (opens in new tab)(zilliz.com)1DISCURSIVE2y ago0Save
