Comment by dmos62 Comment by dmos62 4 days ago 0 replies Copy Link View on Hacker News What kind of abstractions around llm-io would you prefer?