LURNR

The description states:

• LURNR is a synthetic-only AI-assisted algebra intervention system.
• It follows one learner through an evidence sequence involving incorrect submissions, hints, reasoning under support, and transfer checks.
• The system removes instructional support and changes the problem representation to test if reasoning survives — if not, it routes the case to a teacher.
• It uses GPT-5.6 as a bounded proposal engine, responding via strict JSON schema.
• A Haskell engine evaluates whether AI proposals are legal under current learner snapshots and course policies.
• The system tracks protected obligations, evaluates structured transfer evidence, and authorizes claims or repairs before committing state.

Inference LURNR is not a general-purpose tutoring tool but a specialized system for tracking and evaluating learning in algebra with a focus on preventing premature mastery claims.

The description states:

• The project was inspired by the question: “What did the learner actually learn?”
• It aims to avoid building another “tutor chat with a nicer haircut.”
• LURNR focuses on narrower, harder questions:
  • What help is allowed?
  • What thinking still belongs to the learner?
  • Does a teacher need to step in?

Inference LURNR positions itself as a system that prioritizes educational correctness over task completion, distinguishing between assisted performance and demonstrated understanding.

The description states:

• LURNR is built for algebra learners.
• It includes a teacher dashboard to route cases requiring attention.
• The system is designed for intervention in algebra learning, not general education or assessment.

Inference The primary users are likely teachers and instructional designers working with algebra students, particularly those seeking systems that avoid false mastery claims.

Not evidenced.

The description does not mention any pricing, monetization strategy, or business model. It only describes the technical architecture and use case for a demo.

The description states:

• Built with Haskell (core logic), TypeScript/Node.js (network layer), Astro/LiteShip (UI).
• Uses PostgreSQL, Railway, OpenAI Responses API, and JSON Schema.
• The system includes a trusted core in Haskell, which evaluates proposals before they can be committed.
• A deterministic proposal path is used for canonical cases, while live GPT calls are isolated and not part of the main learner record.

Inference LURNR shows strong technical discipline with clear separation between AI generation and system authorization. It uses a hybrid deterministic/live approach to manage AI integration safely.

Not evidenced.

There is no mention of actual users, customers, revenue, or adoption beyond the hackathon demo. The system runs publicly but with synthetic data only.

Not evidenced.

The description does not compare LURNR to existing tools or platforms in the AI education space.

• No real-world testing: The system is demonstrated only with synthetic data and has no evidence of being tested with actual learners.
• Single-founder project: The team size is listed as one, which may limit scalability or operational capacity.
• Limited scope: It focuses only on algebra and does not appear to support broader domains or educational levels.
• High technical complexity: The use of Haskell and strict authorization layers may hinder adoption or integration into existing systems.
• Unclear path to production: While the demo is public, there’s no indication of how it would scale beyond a hackathon-level prototype.

1. What are the key assumptions about how teachers will interact with the system and interpret its routing decisions?
2. How does LURNR plan to validate that its transfer checks are meaningful in real-world settings?
3. Has there been any pilot testing or feedback from educators using synthetic or real data?
4. What is the roadmap for expanding beyond algebra into other subjects or grade levels?
5. How does the system handle edge cases like learners who do not respond to hints or fail consistently?
6. Are there plans to integrate with existing LMS platforms or educational infrastructures?