Ambient->Trex->Hyperfleet Feedback loop

ambient-trex-hyperfleet-feedback-loop.md

Dogfooding Feedback Loop: Ambient → TRex → Hyperfleet

The Self-Reinforcing System

Ambient (AI tooling) develops TRex (template) which generates Hyperfleet (real services) which provides feedback back to both TRex and Ambient.

┌─────────────────────────────────────────────────────────────────┐
│                    DOGFOODING FEEDBACK LOOP                     │
└─────────────────────────────────────────────────────────────────┘

    ┌──────────────┐
    │   Ambient    │ ◄─────────────────────┐
    │ (AI Tooling) │                       │
    └──────┬───────┘                       │
           │                               │
           │ AI-assisted development       │
           ▼                               │
    ┌──────────────┐                       │
    │    TRex      │ ◄────────┐            │
    │  (Template)  │          │            │
    └──────┬───────┘          │            │
           │                  │            │
           │ Generates        │            │
           │ services         │            │
           ▼                  │            │
    ┌──────────────┐          │            │
    │  Hyperfleet  │          │            │
    │   + Pull     │──────────┘            │
    │   Secret     │                       │
    │  (Real World │ Real bugs, patterns,  │
    │   Services)  │ requirements flow     │
    └──────────────┘ back to both          │
           │                               │
           └───────────────────────────────┘

Project Glossary

Ambient - AI spec-driven development

Ambient is a cloud-native platform for spec-driven development and running agentic workflows at scale.

https://github.com/ambient-code/

TRex - Trusted REST Example

TRex is a fully functional API project meant for cloning new projects. It bootstraps a new API and generates all the boilerplate code.

https://github.com/openshift-online/rh-trex

Hyperfleet - Next gen fleet services for Managed OpenShift

Hyperfleet contains new TRex clones as part of the project. One (Hyperfleet API) is a straight API server representing a testable upgrade path. The other (Pull Secret Service) contains a single feature (interfacing with a cloud API and managing image registry tokens).

https://github.com/openshift-hyperfleet/

How Dogfooding Works

1. Ambient → TRex (AI Develops Templates)

• Use Ambient to build and refactor TRex itself
• AI agents handle TRex generator improvements
• Spec-driven development of template features
• AI coordinates TRex codebase changes

2. TRex → Hyperfleet (Templates Generate Services)

• Generate Hyperfleet services from TRex
• Apply TRex updates to existing Hyperfleet services
• Propagate template improvements downstream
• Automated service upgrades via generator

3. Hyperfleet → TRex (Reality Improves Templates)

• Production bugs expose template gaps
• Real-world requirements drive new features
• Successful patterns backported to template
• Edge cases become generator improvements

4. Hyperfleet → Ambient (Reality Trains AI)

• Real services validate AI coordination
• Production constraints refine AI workflows
• Complex scenarios expose orchestration gaps
• Actual codebases improve AI decision-making

Concrete Feedback Examples

Example 1: Nested Resources Discovery

Hyperfleet need: Customer service requires /customers/{id}/subscriptions Feedback to TRex: Add nested resource generation to generator Development with Ambient: AI agents implement nested resource feature in TRex Result: Pull Secret service gets nested resources automatically

Example 2: Event-Driven Pattern

Hyperfleet need: Multiple services need async event handling Feedback to TRex: Event controllers become base template Development with Ambient: AI refactors TRex to include event patterns Result: All new services event-driven by default

Example 3: Production Bug

Hyperfleet bug: Race condition in database advisory locks Feedback to TRex: Fix lock pattern in template Development with Ambient: AI applies fix across TRex codebase Result: Bug prevented in all future services

The Acceleration Effect

Cycle 1:
  Ambient develops TRex feature
  → TRex generates Hyperfleet service
  → 2 weeks manual refinement
  → Patterns flow back to TRex
  → Ambient learns from gaps

Cycle 2:
  Ambient improves TRex with feedback
  → TRex generates better service
  → 1 week manual refinement
  → New patterns flow back
  → Ambient automates previous manual work

Cycle 3:
  Ambient enhances TRex further
  → TRex generates production-ready service
  → 2 days refinement
  → Edge cases flow back
  → Ambient handles most coordination

Cycle 4:
  Ambient optimizes TRex
  → TRex generates deployment-ready service
  → Minimal manual work
  → Advanced patterns flow back
  → Ambient fully automated

Real-World Projects as Feedback Sources

Hyperfleet Services

• Customer Management
• Subscription Service
• Notification System
• Billing Service

Pull Secret Service

• High-security requirements
• Complex authorization patterns
• Performance constraints
• Compliance needs

Both provide real production feedback that Ambient and TRex don't get from toy examples.

Why This Loop Works

1. Ambient needs real projects: Open-source AI tooling benefits from real Red Hat collaboration
2. TRex needs validation: Templates must work in production, not just demos
3. Hyperfleet needs velocity: Real services need fast, reliable generation
4. Continuous improvement: Every cycle makes the whole stack better

Success Metrics

• AI Development Speed: Time for Ambient to implement TRex features
• Template Coverage: % of production code generated vs. manual
• Service Deployment: Days from spec to production
• Feedback Velocity: Time from bug discovery to template fix
• Cross-Service Updates: Time to propagate TRex improvements

The Strategic Advantage

For Ambient: Real Red Hat projects provide authentic collaboration feedback For TRex: Production services validate and improve templates For Hyperfleet: Generated services reach production faster with fewer bugs

---

Core Principle: Ambient develops TRex using AI, TRex generates Hyperfleet services, Hyperfleet reality feeds back to both TRex and Ambient. Each project makes the others better.