List view
Optimal paths for 1. Document ingestion (parsing of text, images, tables, comments, revisions) 2. Document understanding (just pass in as context, create a graph/tree, embed?) 3. Review creation - high-level summary and advice + specific comments w/ optional tracked changes 4. Maximizing speed and minimizing token usage What approaches are to be employed and what tools are used to implement them?
No due date•1/1 issues closed