HATI - Homestead Autonomous Threat Intervention

The description states that HATI is a surveillance system for homesteads that uses AI to detect threats near a poultry coop and deploy an olfactory deterrent. It includes:

• A camera (Foscam G4) capturing motion events.
• Local image processing using a Gemma 4 E4B model to filter benign events.
• A large language model (GPT-5.6 Luna) for visual interpretation of unclear events.
• Deterministic Python code to authorize physical action, including human veto and allowlists.
• A diffuser that sprays predator scent for up to five minutes.
• Telegram integration for alerts and feedback.

Inference The system is described as autonomous but with strong human oversight. It is not a full AI-driven automation without human control.

The author states that HATI was built in response to the challenge of predation on poultry, where traditional deterrents like fencing and lights are insufficient or habituated to by predators. The system is positioned as an intelligent, humane, and autonomous solution for homesteaders.

Inference The positioning is rooted in a personal problem (losing chickens), not market research or customer feedback. The claim of autonomy is qualified by the presence of human veto and deterministic checks.

The description states that HATI targets homesteaders who keep poultry and are concerned about predation. It is described as a solution for hobby farmers or rural dwellers with small-scale livestock.

Inference The target customer is not clearly defined beyond the author’s own use case. No evidence of market segmentation, customer personas, or competitive positioning is provided.

The description does not state any pricing model, revenue streams, or monetization strategy. It is described as a prototype built for a hackathon and not intended for sale or commercial deployment.

Inference There is no evidence of a business model beyond the author’s own use case or demonstration.

The system uses:

• Local motion detection with OpenCV.
• A local Gemma 4 E4B model to suppress benign events.
• GPT-5.6 Luna for visual interpretation.
• Deterministic Python code for authorization and safety checks.
• Telegram for alerts and feedback.
• Tuya-based hardware control (diffuser).
• GitHub Actions CI and a public judge site.

Inference The system is described as robust in testing, with local processing to reduce API costs and safety checks. It includes feedback loops for improvement and crash recovery mechanisms.

The description states that the project was built during one hackathon week by a single developer. It includes:

• A real outdoor test event.
• A controlled improvement fixture that moved from 3/4 to 4/4 accuracy with zero regressions.
• A demo that runs without hardware or API keys.
• A public judge site and GitHub Actions CI.

Inference There is no evidence of customer adoption, revenue, or field deployment beyond the author’s own testing. The system is described as a prototype, not a product in use.

The description does not mention any competitors or direct market context. It is positioned as solving a niche problem for homesteaders, but there is no evidence of existing solutions or competitive analysis.

Inference No competitive landscape is described. The author does not reference similar products or services.

• No traction or revenue: The system is described as a prototype with no real-world deployment.
• Single-person team: Only one developer, which raises concerns about scalability and long-term maintenance.
• Self-reported only: All information is from the author’s own account, with no third-party verification.
• Limited scope: The project is described as a hackathon effort, not a commercial product.
• No customer feedback or validation: No evidence of user testing or real-world feedback beyond the author.

1. What is the actual cost of deploying this system at scale?
2. Has there been any field testing beyond the controlled demo?
3. Are there plans to commercialize or monetize this solution?
4. How does the system handle false positives in real-world conditions?
5. Is there any evidence of customer interest or demand beyond your own use case?