Podcast Atlas
concept Updated 2026-08-07 Topics: Technology

Long-Horizon AI

Long-horizon AI is Yao Shunyu / 姚顺宇’s frame in 140. 对姚顺宇的4小时访谈：请允许我小疯一下！在Anthropic和Gemini训模型、技术预测、英雄主义已过去 for models that can complete longer task chains than ordinary context windows and benchmark episodes expose. His shorthand is “train with finite context, use as infinite context”: train models under bounded context, but deploy them in interactive systems where they can gather information, retrieve, summarize, forget, and keep working over a much longer horizon.

138. 对罗福莉3.5小时访谈：AI范式已然巨变！OpenClaw、Agent范式很吃后训练、卡的分配、组织平权 adds an architecture and agent-data version through Luo Fuli / 罗福莉 and Memo VR. The source treats code, long context, cross-session memory, plan compression, and agent framework traces as training signals, while Agent-Optimized Model Architecture tries to make long-context behavior fast and cheap enough for real agent work.

Episode 17: 向量模型工程师：AI 的隐藏瓶颈与新时代的信息迷宫 adds the retrieval-boundary version. N 同学 / N Student argues that longer context helps some single-document reading tasks, but it does not replace Retrieval-Augmented Generation over very large document collections, and long conversations can still suffer Context Decay when earlier information becomes vague or stale.

OpenClaw 之后，我只想未来 3-6 个月的事情｜对谈 Sheet0 创始人王文锋 adds a file-system practice version through 王文锋 / Wang Wenfeng. He describes task length in terms of step counts, where current practical work may move from dozens of steps toward hundreds or thousands, and argues that file-system state lets agents record progress, observe errors, repair memory, and keep working without relying on a human to hand-curate all context.

AI 发展了 4 年，把应用发展没了？｜AI 年中复盘 adds a market-stage version through 曲凯 / Qu Kai. After chat and coding, Qu expects investor and product attention to shift toward longer-horizon autonomous agents such as GPT Work, where the core question becomes whether models can sustain task context, tool use, and feedback across substantial work rather than only answer or code short prompts.

Key Claims

  • Long-horizon capability is not only a larger context-window number; it is the ability to decide what information remains relevant during extended work.
  • Context Engineering becomes part of model training and post-training, not only a user habit, because the model must learn when to compress, retrieve, ignore, or ask for new context.
  • Yao treats continual learning and long-horizon work as closely related: a model’s active context and KV cache can be viewed as a temporary form of weight-like state.
  • The technical routes include pretraining-side changes such as sparse attention and post-training-side changes such as context management; Yao says his own focus is more on post-training.
  • The concept is especially important for agentic work because tasks such as research, software projects, and personal assistance fail when models lose earlier decisions or optimize only the next short step.
  • Evaluation must be scientific and comparative; a long-horizon technique matters only if it improves real extended tasks rather than merely making prompts longer.
  • Long-horizon ability can be trained indirectly through code and agent workflows because both require dependencies, recovery, planning, and verification over many steps.
  • Long context is not a universal substitute for retrieval; large libraries still need Vector Model Engineering, chunking, and ranking.
  • File-system state can serve as external working memory for long-horizon agents because progress, failures, and repair notes become inspectable artifacts rather than hidden chat history.
  • The 42章经 mid-2026 recap treats long-horizon agents as a likely next narrative after chat and coding, but still asks for grounded workflow and payment proof.

Connections