🦖
The long game

Building something that can build itself.

Most tools stop at suggestions. Dinopsys is aimed at systems that take a goal, inspect the real environment, figure out the steps, do the work, check whether it actually worked, recover when it didn’t, and get sharper the next time around.

DISCOVER REASON BUILD VERIFY RECOVER IMPROVE
The core idea

Not just an AI that answers.

The point is a system that owns the whole job — not one that hands you a plan and walks away. It should finish the computer-based work, not just describe how a human could do it.

01 / 08
What it should eventually handle

One system. Many jobs.

Not locked to a single language, project type, or toolset. The same loop should stretch across different kinds of work.

System architecture

Intelligence becomes orchestration.

Built as cooperating parts, not one giant black box. Each subsystem owns a clear job inside the same loop: discover, decide, act, check, recover.

The autonomous loop

Work becomes a feedback system.

Make a change, then look at what actually happened. That result is the evidence for the next move — not a one-shot guess.

🎯
GOAL Understand
🔭
DISCOVER Inspect
🧠
REASON Plan
🛠️
BUILD Implement
🧪
TEST Execute
✓
VERIFY Evaluate
♻️
RECOVER Adapt
📈
LEARN Improve
Goal — Establish what the system is actually being asked to accomplish before deciding how to execute it.
Self-improvement

Eventually, the system improves itself.

Not reckless self-rewriting. A measured loop: watch how it performed, name the weak spots, try a concrete change in a safe place, check whether it actually helped, and keep only what holds up.

Roadmap

From prototype to autonomy.

Built in stages on purpose. Each one unlocks the next instead of trying to jump straight to the end state.

Overall roadmap progress 0%
Establish the core architecture: discovery, reasoning, implementation, building, testing, verification, and basic failure recovery.
Focus Core architecture
Result Working autonomous foundation
Status Established
Demonstrate the complete coherent loop: understand the request → discover the environment → reason → implement → build → test → verify → recover → produce the final result.
Focus Complete task loop
Result Coherent end-to-end execution
Status Milestone established
Improve recovery quality, evidence handling, failure classification, deterministic behaviour, testing, isolation, and resistance to incorrect assumptions.
Focus Reliability
Priority Evidence-driven recovery
Status Active development
Improve planning, context gathering, research, long-term knowledge, multi-step reasoning, and decision quality.
Focus Reasoning depth
Priority Better decisions
Status Planned
Reduce unnecessary human intervention and allow the system to manage increasingly complete objectives from beginning to end.
Focus Independence
Priority Complete objectives
Status Planned
Dinopsys discovers when additional tools, models, libraries, services, or workflows are required and learns how they fit into its operating environment.
Focus Capability expansion
Priority Tool discovery
Status Future
Controlled self-analysis and improvement: identify weaknesses, propose changes, test them, verify them, and retain improvements that demonstrably work.
Focus Measured improvement
Priority Verified changes
Status Future
The long-term vision: a broadly capable autonomous system able to understand computer-based objectives, discover how to accomplish them, execute the work, verify the result, recover from failure, and improve its capabilities over time.
Focus Broad computer autonomy
Priority End-to-end capability
Status Ultimate vision
Distribution

From a local project to something people can use.

Eventually this has to leave the original machine. Clean installs, portable pieces, versioned releases, and support for the places where the work actually happens.

01
Autonomous Core
05
Core Subsystems
08
Roadmap Stages
∞
Possible Workflows
The real target

Don’t build a system that only follows instructions.

Build one that figures out what the job actually is, finds a way to do it, runs the work, checks the outcome, recovers when it misses, and shows up sharper the next time.

Admin access

Admin mode