DISCURSIVE

7 ms·

1.
The Pi0-Fast Libero-10 Baseline Is 25 Points Low(topicqueue.substack.com)
2 points by DISCURSIVE 9h ago|0 comments
2.
Nvidia's Openness Promise Leaves Out Hosting, Gating and Licensing(topicqueue.substack.com)
1 points by DISCURSIVE 7d ago|1 comments
3.
Perceptron Ships 36B Robot Weights Under Apache-2.0(topicqueue.substack.com)
2 points by DISCURSIVE 14d ago|0 comments
4.
Model Success Rate May Count the Wrong Runs(topicqueue.substack.com)
1 points by DISCURSIVE 21d ago|0 comments
5.
RoboDojo Scores 32 of 34 Robot Policies Under 10% Success(topicqueue.substack.com)
2 points by DISCURSIVE 28d ago|0 comments
6.
Dyna-2: A Million Hours of Human Video, Zero Robot Frames(topicqueue.substack.com)
1 points by DISCURSIVE 1mo ago|0 comments
7.
Gemini Robotics Released One Policy from Feet to Fingertips(topicqueue.substack.com)
2 points by DISCURSIVE 1mo ago|0 comments
8.
5k Hours of Tactile Data, and Four Ways to Get It(topicqueue.substack.com)
1 points by DISCURSIVE 2mo ago|0 comments
9.
Nvidia Puts a 4B World Model on the Robot [Weekly Physical AI Roundup](topicqueue.substack.com)
2 points by DISCURSIVE 2mo ago|0 comments
10.
Turning robot video into training-ready data(eventual.ai)
2 points by DISCURSIVE 2mo ago|0 comments
11.
Xiaomi Opens a 38B World Model Built to Generate Robot Data(topicqueue.substack.com)
4 points by DISCURSIVE 2mo ago|0 comments
12.
Querying Physical AI Data with Daft(eventual.ai)
1 points by DISCURSIVE 2mo ago|0 comments
13.
LeRobot v0.6.0 Adds World Models, Reward Models, and an Open GR00T(topicqueue.substack.com)
1 points by DISCURSIVE 2mo ago|0 comments
14.
A Push on Tactile Data, and a Warning That the Benchmark Was Leaking(topicqueue.substack.com)
2 points by DISCURSIVE 3mo ago|0 comments
15.
Hours of Humanoid Teleop, Recorded in Real Homes(topicqueue.substack.com)
3 points by DISCURSIVE 3mo ago|0 comments
16.
Knowledge curation (not search) is the AI big data problem(daft.ai)
3 points by DISCURSIVE 8mo ago|0 comments
17.
Cutting LLM Batch Inference Time by Half with Dynamic Prefix Bucketing(daft.ai)
2 points by DISCURSIVE 10mo ago|0 comments
18.
Agentic systems are just query engines for unstructured data(daft.ai)
3 points by DISCURSIVE 10mo ago|0 comments
19.
Benchmark: Spark vs. Ray Data vs. Daft on Multimodal Workloads(daft.ai)
1 points by DISCURSIVE 1y ago|0 comments
20.
Show HN: Flotilla, the Distributed Engine for Multimodal Pipelines(daft.ai)
2 points by DISCURSIVE 1y ago|0 comments
21.
Profiling multimodal workloads: lessons from Daft(daft.ai)
1 points by DISCURSIVE 1y ago|0 comments
22.
Processing 24T tokens for LLM training with 0 crashes (what made it possible)(daft.ai)
1 points by DISCURSIVE 1y ago|0 comments
23.
We Hit 100% GPU Utilization–and Then Made It 3× Faster by Not Using It(daft.ai)
17 points by DISCURSIVE 1y ago|11 comments
24.
Elasticsearch Was Great, but Vector Databases Are the Future(thenewstack.io)
16 points by DISCURSIVE 2y ago|7 comments
25.
Have you faced any challenges of using Elasticsearch for vector search?
1 points by DISCURSIVE 2y ago|0 comments
26.
BGE-M3 vs. Splade(zilliz.com)
1 points by DISCURSIVE 2y ago|0 comments
27.
Enhance information retrieval with learned sparse embeddings(zilliz.com)
1 points by DISCURSIVE 2y ago|0 comments
28.
by DISCURSIVE 2y ago
Yeah, we usually just say vector embeddings are the numerical representation of that piece of unstructured data. This glossary page put it together quite nicely. https://zilliz.com/glossary/vector-embeddings
29.
The emerging trend of vector databases(zilliz.com)
2 points by DISCURSIVE 2y ago|0 comments
30.
The Future of Vector Database(zilliz.com)
1 points by DISCURSIVE 2y ago|0 comments

More ›