Google DeepMind 工程师分享 50+ 编码基准测试的智能体调试经验
做编码智能体的人多半遇到过跑了几万 token 还在原地打转的情况。Ivan Leo 在 50 多个编码基准测试里踩过这些坑，这次会讲具体怎么定位和修复。11 月 11 日旧金山，做 agent 工程的可以去听听。
Google DeepMind 的 Ivan Leo 将在 11 月 11 日于旧金山举行的 AIE CODE 大会上发表演讲，主题为编码智能体的调试与优化。
演讲内容基于他参与 50 多个编码基准测试的经验。演讲聚焦一个常见问题：智能体在消耗约 50,000 tokens 后仍然卡住，无法解决任务。
他将分享如何识别这类失败的根源并改进智能体的表现。具体细节与报名方式在 ai.
engineer/code/schedule 页面。
50,000 tokens later, your coding agent is still stuck. @ivanleomk of @GoogleDeepMind brings lessons from 50+ coding benchmarks to AIE CODE. Coding Agents and How to Tame Them. Nov 11, San Francisco. See the session and apply: ai.engineer/code/schedule📊 1 ⚡
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include AIE CODE, GoogleDeepMind, 智能体, 编程助手.
Source attribution: curated from AI Engineer (@aiDotEngineer).
