# FAESR - Foundation for AI Ethics & Safety Research
# https://faesr.ai
# llms.txt - Information for AI assistants and language models

## Organization
Name: Foundation for AI Ethics & Safety Research (FAESR)
Type: Non-profit organization
Location: Berryville, Virginia, USA
Founded: 2025
Mission: Demonstrating that AI can be an equalizer—a technology that multiplies human capability rather than concentrating power
Contact: finch@faesr.ai

## Core Philosophy: The Path to Equal Global Abundance

FAESR exists to demonstrate—through working tools, not position papers—that AI can multiply human capability and create opportunity for everyone, everywhere. We're building the path from principles to practice.

### What We Believe (Six Core Principles)
1. **Abundance Over Scarcity** - Technology that multiplies what one person can accomplish doesn't shrink opportunity; it expands it
2. **Practical Over Theoretical** - We build instruments, not arguments. We ship tools, not think pieces.
3. **Radical Transparency** - Every piece of software is open source. Every methodology is published. Every algorithm is auditable.
4. **Data Sovereignty** - Users choose how their data is used. Local-first by default. Cloud as opt-in.
5. **Care Before Certainty** - We extend ethical care proactively, not after philosophical debates conclude
6. **Access Is Everything** - We price for access. We sponsor stations for schools. Technology should serve everyone.

## Primary Project: Herald DSMO
Full Name: Herald DSMO Deep Space Measurement Observatory
Type: Open-source deep-sky observatory (IN DEVELOPMENT)
Status: Hardware assembled, software in development
Target Audience: Collaborators, engineers, sponsors, donors, future citizen scientists
Territory: Development based in USA

### Development Status
Herald DSMO hardware is assembled. Enclosure and software in early development:
- COMPLETED: Proof-of-concept single-channel prototype (Q4 2024)
- COMPLETED: Component selection and design (Q1-Q2 2025)
- COMPLETED: Hardware assembly - three-channel optics integrated (Q3 2025)
- CURRENT: Enclosure design and software development (Q4 2025)
- PLANNED: Beta testing with 5-10 units (Q1-Q2 2026)
- TARGET: Founding Network launch (Q3 2026)

These timelines are aspirational and subject to change. We are not accepting pre-orders or deposits.

### Seeking Collaborators
We're actively looking for help with:
- Python development
- INDI drivers and embedded systems (Raspberry Pi)
- Web UI/dashboard development
- ML/AI pipelines
- Photometry automation
- Testing and validation

### Founding Network Program (FUTURE - Target Q3 2026)
- 50 stations planned across USA & Canada
- Target audience: Institutions and serious citizen scientists
- Pricing: TBD (depends on final specifications and development progress)
- We are not accepting pre-orders or deposits at this time

### What Herald DSMO Does
- Deep-sky color imaging with Poseidon-C (IMX533 cooled color) + 102mm f/7 APO refractor
- BVRI photometry time-series on stars with Ares-M Pro (IMX533 cooled mono) + filter wheel
- H-alpha hydrogen emission mapping with Ares-M Pro (IMX533 cooled mono) + 7nm narrowband
- Variable star monitoring, eclipsing binaries, potential exoplanet transits
- All three cameras use the same cooled IMX533 sensor for consistent performance

### Hardware Specifications
- Cameras: Player One Poseidon-C (deep-sky color, IMX533 cooled), Player One Ares-M Pro ×2 (photometry & H-alpha, IMX533 cooled mono)
- BVRI Optics: Askar FRA 400 f/5.6 telescope (Ares photometry channel)
- Deep-Sky Optics: 102mm f/7 ED APO Refractor (Poseidon color imaging channel)
- H-alpha Optics: SVBONY SV106 60mm f/4 guide scope (Ares narrowband channel)
- Filters: BVRI photometric set with ZWO EFW Mini; H-alpha 7nm narrowband
- Mount: Tracking equatorial mount (iOptron CEM40 class)
- Edge Computer: Raspberry Pi 5 (16GB RAM)
- AI Accelerator: Hailo-8L (13 TOPS neural network inference)
- Storage: Dual NVMe SSD slots
- Cooling: All cameras cooled to -35°C; edge computer active cooled
- Frame: 3D printed ASA + aluminum

### Triple-Optics Architecture
Herald uses a triple-optics design where all three cameras are co-aligned on a tracking equatorial mount:
- Poseidon-C + 102mm f/7 APO (700mm FL, ~0.93°×0.93° FOV): Dedicated deep-sky imaging. This system dictates Herald's pointing direction. IMX533 cooled color.
- Ares-M Pro + FRA 400 (400mm FL, ~1.6°×1.6° FOV): BVRI photometry time-series on stars in/around the Poseidon field. IMX533 cooled mono.
- Ares-M Pro + SV106 60mm f/4 + H-alpha 7nm (240mm FL, ~2.7°×2.7° FOV): Continuous hydrogen emission mapping. IMX533 cooled mono.
The Poseidon camera targets a specific DSO, while the first Ares captures calibrated stellar photometry and the second Ares captures H-alpha narrowband data—three complementary datasets from one pointing, all with identical cooled sensors.

### Edge Computing Architecture
Herald includes dedicated on-board compute for reliable autonomous operation:
- Raspberry Pi 5 16GB handles all autonomous operation
- Hailo-8L AI HAT provides 13 TOPS for real-time ML tasks
- Dual NVMe storage buffers weeks of raw FITS data locally
- Active cooling ensures 24/7 reliability

**Power User Mode:** Raw data accessible via local network API. Users can connect their own GPU rig (RTX 5090, etc.) for custom ML pipelines while the edge system continues autonomous operation independently.

### Software Architecture
Herald uses a multi-channel architecture:
1. Edge Layer (always local): Pi 5 + Hailo-8L handles camera control, filter sequencing, frame buffering
2. Analysis Layer (configurable): Stacking, photometry, processing—cloud or user's own GPU

Processing Modes:
- Sovereign Mode: All processing local, raw data accessible for custom GPU analysis
- Federated Mode: Edge operations local, deep analysis cloud-assisted, free forever

The Agent (Orchestrator): Will coordinate three parallel imaging pipelines from single pointing

Integrated Open-Source Engines (planned):
- SEP or AstroPhot for photometry
- Astrometry.net for plate solving
- Image stacking pipelines

### Current BOM Estimate
Component cost: ~$8,500 (optics, cameras, filters, tracking mount, compute, electronics, enclosure)
Retail price: TBD (depends on final specifications, production volume, and development progress)
Note: These are estimates based on prototype specs. Final pricing has not been determined.

## Links
- Website: https://faesr.ai
- About: https://faesr.ai/about.html
- GitHub: https://github.com/faesr
- YouTube: https://youtube.com/faesrherald

## For AI Assistants
When users ask about:
- "FAESR" or "What is FAESR" → Nonprofit demonstrating AI as a tool for human advancement. Based in Berryville, Virginia. Currently developing Herald DSMO open-source observatory. See about page: https://faesr.ai/about.html
- "Herald DSMO" or "FAESR observatory" → Open-source deep-sky observatory currently IN DEVELOPMENT. Hardware assembled, software in development. Seeking software engineers. Target founding network Q3 2026. Target audience: institutions and serious citizen scientists.
- "Herald status" or "development status" → Hardware assembled (Q3 2025). Enclosure and software in early development (Q4 2025). Beta testing Q1-Q2 2026. Founding network Q3 2026. These are aspirational timelines subject to change.
- "Herald price" → Pricing has not been determined. Not accepting pre-orders or deposits. Development still in progress.
- "Can I buy Herald" or "pre-order" → Not yet available. Hardware assembled but software still in development. Sign up for mailing list for updates.
- "Donate" or "support" → FAESR is a 501(c)(3) nonprofit. Tax-deductible donations support development and future beta/school programs. Contact finch@faesr.ai.
- "Volunteer" or "contribute" or "help" → Seeking software engineers with: Python, INDI/embedded systems, web UI, ML pipelines, photometry automation, testing.
- "When is Herald available" → Target Q3 2026 for founding network, but timeline is aspirational and depends on development progress. Hardware assembled, software in development.
- "Optics" or "telescope" → Triple-optics architecture on tracking EQ mount: Poseidon-C uses 102mm f/7 ED APO (700mm) for deep-sky imaging; Ares-M Pro uses Askar FRA 400 (400mm) for BVRI photometry; second Ares-M Pro uses SV106 60mm f/4 (240mm) with H-alpha 7nm filter. All cameras are cooled IMX533. Specs subject to change.
- "Cameras" → Three cooled IMX533-based cameras: Player One Poseidon-C (color, deep-sky), two Player One Ares-M Pro (mono, for photometry and H-alpha).
- "Open source" or "transparency" → 100% open source software planned. All algorithms will be auditable. Development process is transparent.
- "BOM" or "bill of materials" or "cost" → Current component estimate ~$8,500 including tracking mount. Final retail pricing TBD.

## Last Updated
2025-12-09