Tag: ai-agents
-
Could agents learn to work better from their own runs?
Most agent memory research is about remembering the user. I built a browser agent that mines its own traces instead, and the lessons that transferred were not the ones I expected.
-
Context engineering needs a context engine
Stronger models help, but agent reliability may depend just as much on better context systems around them.
-
Borrowed confidence is fragile in agentic systems
Production-like evals revealed the retrieval architecture I actually needed and reminded me that confidence in agentic systems has to be earned, not borrowed.
-
The lethal trifecta in AI agents
When agents can read private data, process untrusted content, and communicate outward, prompt injection becomes a much more serious security problem.
-
How Plan Caching Reduces LLM Agent Costs
Plan caching reuses planning templates across similar agent tasks, cutting cost and latency without throwing away accuracy.
-
Stop filling your agent's context window just because you can
Bigger context windows do not remove failure modes. They create new ones when we stop being intentional about what goes into an agent's context.