Recursive Language Models: Claude Code, Agent Swarms, Big Research Bets, & the $40M AI Problem…
@latentspacepod35 тыс. подписчиков
46 040 просмотров
18 комментариев
Комментарии из оригинального поста в X (Twitter)

@MixtTensi8198
@a1zhang I watched this and I will watch it again. My Favorite Latent Space session yet! Thank you @a1zhang you are my hero! :)
1 отметка «Нравится»

@AgentikHQ
@a1zhang treating Claude Code, Codex and Pi as the same shape clicked for me the harness stays, the model under it keeps rotating
1 отметка «Нравится»

@Eminemead
@a1zhang @a1zhang just shout out @dosco as this one guy who uses RLM in Ax lol
1 отметка «Нравится»

@oosjake
@a1zhang Treating Claude Code, Codex, and Pi as the same shape is the useful bit. The harness and review loop matter more than which model sits underneath this week.
1 отметка «Нравится»

@oosjake
@a1zhang Yes. Deterministic gates first. A second model is still correlated; the compiler is not.

@robdias
@a1zhang If Claude Code, Codex, and Pi are the same shape, the moat is the harness and the eval loop. How much of that $40M problem solver is recursion vs better context offload?

@guannanjiayou
@a1zhang oracle: recursive calls help only when each step leaves a traceable state

@itstonywhite_
@a1zhang HI, I’ve been using AINews to keep up with AI, but I noticed the website hasn’t been updated since September 9. 😅 Is AINews coming back? Would love to see the daily updates return!

@latentspacepod
@a1zhang working on new one

@DGlushakov41949
@a1zhang How much of that gain remains when cost and recovery from failed runs are part of the test? MIT CSAIL reports a median 26% gain over GPT-5 compaction on long-context tasks, but operational value will depend on orchestration overhead.

@oosjake
@a1zhang RLMs make context routing explicit, but recursive calls need a budget and stopping rule. Otherwise context offloading just hides runaway coordination cost.

@catmanyau
@a1zhang The leverage in AI often comes from changing the search strategy, not just scaling the search. A good expert can turn brute force into a much smaller, structured problem.

@JasonC_Dev
@a1zhang 代理群集听着猛,我卡在凭证隔离——集群里每个 agent 是各自单独授权工具,还是共享同一套会话密钥?一个节点挂了,别的节点会不会捡到半截旧 token

@oosjake
@a1zhang The context offloading angle is the key. Recursive agents are less about adding more calls and more about giving each subproblem a bounded working set.

@BenjaminSandn
@a1zhang Recursive loops make the setup part of capability. As the system rewrites its path, an external judge matters more.

@MaxCurnin
@a1zhang Recursive subagents are a compute bill with a principal problem attached. Who authorized the leaf when the parent only had a vague task?

@niccolobanti
@a1zhang Who or what reviewed the output of the 10,000-agent run? Review is the step I see fail first when subagents multiply, since a model checking its own diff costs more compute than writing it.

@binbinyiyilili
@a1zhang RLM 的实用价值在于把“上下文长度”问题改写成“可编排的计算”问题:代码负责筛选与卸载,子代理负责局部推理,主代理只汇总结果。真正落地时建议同时报告递归深度、工具调用次数和失败回退率,否则很难和普通长上下文方案公平比较。