Recursive Decision Models

or, Operating System One

Raymond Weitekamp

recursivedecisionmodels.comoSo.computer

What if “reasoning” was actually many small decisions?

RLM × Jev ⇒ RDM

Omar Khattab on X, Oct 15, 2025: Alex is a second-month PhD student who did this work in 4 weeks, and he suspects Alex has secret recursive Alexes doing his work. It quotes Alex Zhang announcing Recursive Language Models. Raymond Weitekamp replies: smells like claude code to me ;) Raymond Weitekamp on X, Sep 16: everyone has been building divergence (generate), i'm very excited for jev as convergence (decide), yin and yang. It quotes his post showing the TypeSafe AI 'You're in!' account email.

Last time

Recursive Coding Agents

  • The bottleneck is not intelligence. It's reliability. It's trust.
  • RLMs are the new reasoning models.
  • Trick question: RLMs are Recursive Coding Agents.
Raymond Weitekamp on X, Apr 20, 2026: the article RLMs are the new reasoning models

RLM / LLM

Root RLM (depth=0)
├── Sub-RLM A (depth=1)
│   ├── LLM A1 (depth=2)
│   └── LLM A2 (depth=2)
└── Sub-RLM B (depth=1)
    ├── LLM B1 (depth=2)
    └── LLM B2 (depth=2)

RDM / Decision

Root RDM (depth=0)
├── Sub-RDM A (depth=1)
│   ├── Decision A1 (depth=2)
│   └── Decision A2 (depth=2)
└── Sub-RDM B (depth=1)
    ├── Decision B1 (depth=2)
    └── Decision B2 (depth=2)

RLM / LLM

Root RLM (depth=0)
├── Sub-RLM A (depth=1)
│   ├── LLM A1 (depth=2)
│   └── LLM A2 (depth=2)
└── Sub-RLM B (depth=1)
    ├── LLM B1 (depth=2)
    └── LLM B2 (depth=2)

RDM / Decision

Root RDM (depth=0)
├── Sub-RDM A (depth=1)
│   ├── Decision A1 (depth=2)
│   └── Decision A2 (depth=2)
└── Sub-RDM B (depth=1)
    ├── Decision B1 (depth=2)
    └── Decision B2 (depth=2)

RLM / Decision as a tool

Root RLM (depth=0)
├── Decision A (depth=1)
├── Sub-RLM B (depth=1)
│   ├── Decision B1 (depth=2)
│   └── LLM B2 (depth=2)
└── LLM C (depth=1)

RDM / LLM as a decision

Root RDM (depth=0)
├── Sub-RDM A (depth=1)
│   ├── Decision A1 (depth=2)
│   └── LLM A2 (depth=2)
└── Sub-RDM B (depth=1)
    ├── Decision B1 (depth=2)
    └── Decision B2 (depth=2)

What if “reasoning” was actually many small decisions?

Recursive Decision Models

A decision that can decide to make another decision.

Three rules

  1. 1The context stays outside the model, and code cuts it into slices.
  2. 2A decision can choose to call another decision, up to a maximum depth.
  3. 3It works with no generative model at all, and it composes with LLMs and agents.

Decisions all the way down

What if the pins could decide?

Same number. Any color.

Hierarchical search, without the beam

Hierarchical search

Twelve thousand categories, and no request shows more than one level of them. Each answer decides which categories to ask about next: every one at least half as likely as the likeliest. The same walk as TypeSafe's hierarchical classification cookbook, without the beam.

recursivedecisionmodels.com/hierarchy ↗
A product listing

Laptop backpack with a padded sleeve, 20 litres.

The search Laya's answers followed rated, not followed

A replay, not a live run. The laptop backpack, board game and hiking backpack are Laya's recorded answers (a 450 MB decision model, in the browser). The cat perch and headphones are an oracle replay: the right path, with odds set by hand.

Nathan LeClaire @dotpem · Oct 18, 2026 𝕏

Here's an example of stacking jev on jev inspired by Lily. Jev divides and conquers an image to find a bunch of bounding boxes. For straightforward detection it's worse/slower than SAM3 but for something intelligent - like finding all sheep who might "yee haw" - it does well.

View on X ↗
lillyguisnet @lillyguisnet · Oct 16, 2026 𝕏

Replying to @lotusfracture

First image, increase number of image cuttings allowed up to 15 times. Second, same but also included feedback. Each produced box is checked and asked if correctly on a beetle. If no, repeat cycle, if yes keep the box.

Greyed squares, Jev said no beetle. Other squares, Jev said 1 beetle then asked to produce the box.

Green dot, ground truth. Pink box, Jev box. Yellow dot, Jev box centroid. Blue dotted box, independent cuts for another test (can ignore here).

Overall, the feedback loop helps throw away bad boxes but doesn't help finding more. 1 correct beetle was rejected too.

Nice trick! :)

A case of pinned butterflies and beetles cut into nested squares, with Jev's boxes on the beetles, cuts allowed up to 15 timesThe same case with a feedback check on each box
View on X ↗

What if “reasoning” was actually many small decisions?

Decisions can only choose from what code has already laid out.

Pure decisions need every option filled in ahead of time. For games, browsers and computer use, that might be enough.

It must compose with LLMs

LLM: diverges decision model: converges decisions 0 / 10

An LLM fans out options, a decision model keeps one or two, and each one kept opens a smaller fan.

Generation x Decision Demonstrate Emergent Graphs

Call graphs of two decision-only runs: regular trees laid out by codeCall graphs of four runs with an LLM: no two the same shape; the largest is an 80-request cycle that only a limit stops
Blue: a request to the decision model. Orange: something the LLM wrote. Left, decisions only; right, with an LLM. After docs.typesafe.ai/demos/smart-home.

A decision model can only point, so something has to say what it is pointing at.

A map links decision models to LLMs: names, each with a description for the model and a thing for the code.

What worked goes back on the map

Fast and cheap means we can speculate

Run many decisions at once, and let the system decide which to trust.

Try it

rawwerks/dspy

dspy.experimental.RDM, in a fork of DSPy (branch feat/rdm-v0).

Python 17 1 Updated Oct 4, 2026
rawwerks/rdm

Recursive Decision Models

Rust Updated Nov 7, 2026

A woman in a gym sits on a leg machine looking at her phone instead of exercising A man leans into a laptop so far that his face comes out through the screen Seen from above at night, a person at a desk with two monitors, two keyboards, three phones and a mug of coffee
Four glowing silhouettes: typing on a laptop, a laptop with an AI sparkle, talking to an AI by voice, and sitting cross-legged with orbits around them

Radical HCI

Raymond Weitekamp @raw_works · Oct 11, 2026 𝕏

spent some time this weekend getting back to my roots in real-time machine learning

this is a recursive decision model, deciding when it can decide that a fast decision is decisive enough

as the rdm learns from repetition, it autonomously begins to trust a very small model

by the end of the video the full computer use response time is ~45 ms

decision agents are here

View on X ↗
Matt Mastracci @mmastrac · Oct 29, 2026 𝕏

We put up a version of hunch that you can play Pictionary against in your browser (no server!). It's really good, even if your drawings are not.

No Jev or OpenRouter calls here, nothing leaves your system! All powered by @GoogleGemma's EmbeddingGemma 2.

https://tryhunch.dev/pictionary

View on X ↗

How can we decouple our bodies from our devices?

What would it mean to work on a computer without touching it?

What would it mean to work on a computer without being able to see it?

Operating System One

Join the waitlist at oSo.computer

oSo is…

  • …a computer use agent built on RDM
  • …a recursively improving system that learns on-the-fly
  • …a tool to control your computer without touching it
  • …an experiment in radical human computer interfacing
  • …real-time machine learning

Operating System One

Join the waitlist at oSo.computer