A request to learn what breaks in real agent use

A computer science student recently focused on agent and . The work included building and a , then stopping it after the evidence did not support it strongly enough.

The current goal is learning what actually fails when people run agents in real settings, not selling a product or building one right now. The request asks for 2 or 3 short answers about the most annoying recent agent failure, with a 10-minute text conversation as an option.

Key points

  • The recent focus was agent and .
  • A was built and benchmarked, then dropped because the evidence was not strong enough.
  • The main question is what breaks for people running agents in real life.
  • There is no product pitch or active build described.
  • The requested feedback is short: 2 or 3 questions over about 10 minutes by text.
Read original