Redesigning Execution for Intelligence
A concatenative programming framework for AI-first development
Why REI?
What if software could evolve itself while running?
The Hypothesis
It sounds like science fiction, but the technical pieces already exist. The real question is whether our programming tools are ready for it.
Software development is changing. Language models can write code, but our tools and languages were designed for humans alone.
Current programming paradigms create friction when AI tries to reason about, generate, or execute code alongside humans.
The Experiment
That's the possibility REI is exploring—not as distant future speculation, but as a design constraint for today.
REI reimagines programming as a collaboration between humans, AI, and machines - where each contributor works in their native paradigm.
By using a concatenative, stack-based approach, REI creates predictable execution that both humans and AI can reason about effectively.
How It Works
AI Instrumentality Project
Concatenative & Stack-Based
Operations compose naturally through function concatenation. Execution is predictable - what you see is what gets executed, in order.
LLM-Optimized
Documentation and architecture designed for AI consumption. Every operation is self-describing, enabling LLMs to reason about and generate code effectively.
Minimal Core, Composable
Small vocabulary of primitives enables infinite combinations. Build complex behavior from simple, well-understood building blocks.
Separated Concerns
Execution engine (REIVM) is separate from interfaces (monitors). Multiple ways to interact with the same core runtime.
System Architecture
REIVM executes operations on a stack. REIMON provides the interface layer. Multiple frontends connect through REIMON.
Fair Warning
This is intentional - we're building for collaboration across different forms of intelligence.