Research Topics

Start with a problem, then follow a comparison or inspect individual systems. These four topics are reading paths through the same accepted analyses, not separate populations.

Topic Architectural questions
Agent Memory What survives a task, how does it reach a later decision, and what happens when it is wrong?
Runtime & Control Who chooses and executes the next action, under which authority?
Knowledge & Verification What is checked before an agent relies on a claim?
Workflow & Recovery How does multi-step work coordinate, recover and reach delivery?

Memory currently has the most developed comparison research and structured dataset. The other topics begin with questions and generated per-system accounts; they do not yet have equivalent comparison tables.

Also browse comparisons, the Systems Directory, or Methods & Data.


Complete file listing (generated at build time)