emschwartz's Likes

RSS doesn't necessarily means firehose
andregarzia.com·24w·
📰RSS Reading Practices
Reading RSS content is a skilled activity
doliver.org·24w·
📰RSS Reading Practices
Two tower models instead of 'hybrid search'
softwaredoug.com·23w·
Discuss: Hacker News
🔍Vector Search Algorithms
Two tower embeddings instead of 'hybrid search'
softwaredoug.com·23w
🎯Vector Search
Why performance optimization is hard work
purplesyringa.moe·23w·
🧮Compute Optimization
Qwen 3 offers a case study in how to effectively release a model
simonwillison.net·23w·
Discuss: Hacker News
🤖AI
Separate Indexes or Compound Indexes in SQL
use-the-index-luke.com·23w·
Discuss: Hacker News
📇Indexing Strategies
🎲 The Models Want to Reason
benjamincongdon.me·24w
🧠LLM Inference
Cross-Encoder Rediscovers a Semantic Variant of BM25
shaped.ai·24w
🔍Information Retrieval
Semantically rank everything using LLMs (companies, candidates, ideas etc.)
eamag.me·24w·
Discuss: Hacker News
🏆LLM Benchmarking
🚨This week's top AI/ML research papers:
threadreaderapp.com·24w
🏗️LLM Infrastructure
One Embedding to Rule Them All
shaped.ai·24w·
Discuss: Hacker News
🎯Vector Search
Reproducing Hacker News writing style fingerprinting
antirez.com·25w·
🔎Tantivy
Cursor 0.49: Rules generation, improved agent terminal and MCP images
cursor.com·25w·
Discuss: Hacker News
📟Terminals
Benchmarking LLM-based Relevance Judgment Methods
arxiv.org·25w
🏆LLM Benchmarking
Two Years of Rust
borretti.me·25w·
🦀Rust Compiler Internals
An intuitive introduction to text embeddings
stackoverflow.blog·26w·
Discuss: Hacker News
📊Embeddings
Foundation Model for Personalized Recommendation
netflixtechblog.medium.com·29w·
Discuss: Hacker News
🎛️Feed Filtering
The Third-Worlding of America
paulkrugman.substack.com·26w·
Discuss: Substack
🏛️Politics