编码代理对前端开发加速效果最显著，其次是后端开发，基础设施和科研工作加速幅度依次递减。前端开发因编码代理熟练掌握TypeScript、JavaScript及React、Angular等框架而大幅提速；后端开发虽也有提升，但仍需人类工程师处理边界情况和安全漏洞；基础设施涉及复杂权衡和深度调试，AI帮助有限；科研工作除编码外还有大量非编码环节，编码代理的加速作用最为有限。理解这些差异有助于合理设置团队预期并优化组织方式。
Longer training horizons have not slowed AI progress because modern architectures (like transformers) and techniques such as reward shaping distribute gradients effectively, making credit assignment over long episodes tractable without exponential difficulty increases.
AI progress hasn't slowed despite longer training horizons, possibly due to huge FLOP efficiency gains from fixing bugs, unreliable human intuitions about near-human intelligence, and capabilities depending on traits beyond intelligence like persistence.