Real-world agent test: flight and hotel rebooking burned tokens, manual app was still easier
curious_vii · x · 2026-11-03
User curiousvii reports a frustrating, token-intensive attempt to rebook a flight and hotel with an AI assistant. Ultimately tapping through the American Airlines app was easier, and the email the assistant drafted to hotel staff was verbose. Verdict: agent performance on long-tail real-world tasks remains hit-or-miss.
More from coding & agent
- Workers IO launches deterministic simulation infra to verify AI coding agents' code — julianweisser · 2026-11-03
- Open-source AI pit crew for iRacing runs 5 agents on Kimi, checks every number against telemetry — JeremyCMorgan · 2026-11-03
- Agent control plane benchmark: Smalltalk-80 cascades hit 100% compliance vs 2.5% for Gherkin — RandalSchwartz · 2026-11-03
- Karpathy: one prompt to Claude Code deploys a YouTube thumbnail-swap browser extension — amaarora · 2026-11-03
- Cherry Pi proposes layered harness architecture: agent loops compose with governed workspaces — sull · 2026-11-03
- $200 Claude Max runs overnight multi-subagent Opus 5.5 job, burns only 6% of weekly quota — lxfater · 2026-11-03