BREAKING
Karpathy's 'LLM Wiki' Goes Viral
0
notes converted
0
Memex concept
Three Core Operations
1
Ingest new source
↓
2
Query with citations
↓
3
Lint for conflicts
LLM Wiki vs Standard RAG
Standard RAG
per-query
●
Rebuilds from raw sources
●
Needs vector database
●
Harder to accumulate
LLM Wiki
persistent
●
Local Markdown files
●
No vector database
●
Continuous cross-linking
Cheap, Durable, but Daunting
A Personal Brain That Curates Itself
AI NEWS BLITZ
A note-taking pattern from Andrej Karpathy is reviving dormant personal archives.