HarborGrid

• The description states HarborGrid is a "planning demonstration for marine EV charging in the Salish Sea and Juan de Fuca Strait."
• It uses a deterministic optimizer to score combinations of 6 synthetic candidate harbour sites against 8 weighted demand nodes.
• A budget slider (1–5 chargers) drives an interactive Leaflet map and dashboard showing coverage gains and gap reductions.
• The system includes a structured facts layer that feeds verified data into GPT-5.6 for explanation, with strict instructions to never select, rank, recommend or invent.
• The author states the app degrades gracefully if no API key is configured, using a deterministic template with the same numbers.
• It is described as a "docs-first project" with PROJECT.md, DECISIONS.md and STATUS.md files.

Not evidenced No actual product, revenue, customers, or commercial use cases are described beyond this demonstration.

• The description states: “The optimizer picks the sites — AI only explains them.”
• It positions itself as a demonstration of an "unfashionable idea" that AI belongs in the explanation layer, not the decision loop, for infrastructure planning.
• The author claims to have built this as a solo project with no team, and submitted it to a hackathon.
• The tagline: “Deterministic marine EV charging planning for the Salish Sea” is self-reported.

Not evidenced No evidence of positioning evolution or market traction. Claims are based on self-description only.

• The description states the project focuses on "marine EV charging in the Salish Sea and Juan de Fuca Strait."
• It is described as a demonstration for "the industry" to help get sponsorship for EV travel.
• The author mentions “a hand-mapped 584 NM Vancouver Island circumnavigation network” as a real dataset, suggesting future target customers may be marine infrastructure planners or EV charging network developers.

Not evidenced No identified customer segments, personas, or ICP beyond the author’s personal interest in EV travel and hackathon submission.

• The description states no pricing or business model is evident.
• It is described as a demonstration project with no revenue streams.
• The author notes that the app degrades gracefully if no API key is configured, suggesting possible future monetization via API access or premium features.

Not evidenced No pricing, monetization strategy, or business model details are provided beyond the demo.

• Built solo in two days (July 14–15) using Next.js, React, TypeScript, CSS Modules, Leaflet + React Leaflet over OpenStreetMap tiles, and OpenAI JavaScript SDK.
• The core design rule is: deterministic code makes every siting decision.
• The optimizer exhaustively scores site combinations using straight-line Haversine nautical miles.
• A budget × range scenario matrix (budgets 1–5 × 10/15/20/25 nm) powers the slider with zero AI calls and zero client-side solver runs.
• The explanation service uses GPT-5.6 with strict instructions to never select, rank, recommend or invent.
• A structured facts layer feeds deterministic data into the model.
• The system includes rate limiting, an in-memory cache keyed by SHA-256 hash of facts, and a 503 fallback path.

Not evidenced No evidence of production deployment, scalability, or long-term technical strategy beyond this demo.

• The description states the project was built solo in two days.
• It is described as a hackathon submission to the OpenAI 2026 hackathon on Devpost.
• The author mentions that the real dataset exists (584 NM Vancouver Island circumnavigation network with 36 real stations) but no evidence of adoption or usage beyond this demonstration.
• No evidence of revenue, customers, or product-market fit.

Not evidenced No traction, adoption, or maturity beyond a solo-built demo.

• The description does not mention any competitors or direct market comparison.
• It is described as a solo project focused on marine EV charging planning.
• The author notes that the idea of placing AI in the explanation layer rather than decision loop is "unfashionable" but potentially valuable for infrastructure planning.

Not evidenced No competitive landscape, market positioning, or competitor analysis provided.

• Single-person development: The project was built solo with no team — a high-risk signal for scalability and long-term viability.
• No commercial traction: The description states it is a demonstration with no revenue, customers or adoption.
• Unproven market demand: No evidence of customer validation or real-world use cases beyond the author’s personal interest in EV travel.
• Limited scope: The project is focused on a specific geographic area (Salish Sea) and uses synthetic data — not scalable to broader markets without significant expansion.
• AI explanation layer only: While novel, this approach may not be sufficient for commercial adoption without integration into actual decision-making systems.

Not evidenced No evidence of risk mitigation or strategic planning beyond the demo.

1. What is the validation process for the synthetic demand data used in the demonstration?
2. How does the author plan to transition from this solo-built demo to a scalable product or service?
3. Are there any real-world partners or stakeholders interested in using or investing in HarborGrid beyond the hackathon submission?
4. What are the technical and commercial risks of scaling the deterministic optimizer for larger candidate sets?
5. Is there a roadmap for integrating real-world data beyond the 36 stations mentioned?