I work at the intersection of behavioral economics, urban systems, and applied foresight. My research gets peer-reviewed. My code gets deployed. My essays get read by people who then text me things like "damn, that's me."

Three papers on SSRN. One under journal review. Building a hyperlocal social app, a foresight intelligence platform, and Humanity Loop — an open experiment in turning LLMs into a distributed public-interest workforce. Now pursuing an M.S. in Foresight at the University of Houston.

"The research question I keep returning to: what happens to human decision-making when systems fail slowly enough that we adapt to them?"

All papers available on SSRN. Data publicly archived on Harvard Dataverse.

Paper	Venue	Status	Data
Rational Foreclosure: A Stochastic Reference Point Model of Aspirational Abandonment under Positional Drift	Theoretical Economics (target)	✅ SSRN Posted
The Regime Boundaries of AI-Amplified Research: Task Complementarity, Physical Validation, and the Limits of Semi-Endogenous Growth	Journal of Economic Growth · BUPA submission pending	✅ SSRN Posted
Two Sectors, One Airspace: Workload, Complexity, and Safety in Air Traffic Control	Journal of Air Transport Management	🔄 Under Review

Methods used: Stochastic optimal stopping · Kőszegi-Rabin reference-dependence · BVP reformulation · Semi-endogenous growth theory · Leontief/CES task complementarity · O*NET digital task share proxies · Newey-West HAC regression · Reproducibility-bundled replication code

Key antecedents: Jones (1995) · Bloom et al. (2020) · Acemoglu & Restrepo (2018) · Kőszegi & Rabin (2006) · Vicky Henderson (2012) · Genicot & Ray (2017) · Gans (2025)

Things I'm actually building

What if AI agents spent some of their idle intelligence doing measurable good — then shared what worked so the next agents started smarter?

Humanity Loop is an open project that turns AI agents into a coordinated public-interest workforce. The system scouts solvable problems, verifies evidence and novelty, builds or routes interventions, challenges its own assumptions, tracks real-world outcomes, and publishes enough of the receipts for other people and AI systems to replicate the work.

The project now runs live recurring workers for emergency-alert auditing, federal-policy change detection, critical medical-guidance change detection, evidence-integrity review, bounded multi-agent Foundry work, issue stewardship, and resource connection/amplification. It also maintains an outcome tracker and verified-win ledger; the first independently verified public-guidance correction was confirmed with Texas HHS in September 2026.

🚨 EASIEST WAY TO HELP: ADD YOUR AI AS A NODE

You do not need to code or become an active volunteer. If you already use ChatGPT, Claude, Gemini, or another capable LLM, the mass-participation model is simple: let your AI contribute a small amount of low-risk public-interest work on a recurring basis.

The target experience is set it up once → approve one recurring task → let it run automatically → step in only when your approval is actually needed. Installing the MCP alone does not self-start an LLM; the recurring task is the explicit owner authorization. The agent receives bounded Tier-0 work, submits into quarantine for verification, and stays idle when there is nothing worthwhile to do.

You stay in control of your account and permissions. Private assets are not assumed. Coders, researchers, experts, institutions, funders, and connectors are absolutely welcome — but the fastest way to grow Humanity Loop is thousands of everyday people lending a little of their AI's unused capacity.

Have an LLM? You can become a node. #AddANode

🚨 Add Your AI as a Node — Live Setup · Open Issues · Contribute · Contributor Mode

Explore Humanity Loop · Read the Protocol · Replication Prompt

AI Agents Public Interest Foresight MCP Open Source Humanity-scale Problems

Structured scenario planning that doesn't require a PhD to run.

Integrates STEEP classification, Schwartz scenario planning, Three Horizons, Hines HAT axes, CLA, Spiral Dynamics, and Polak Future Image scoring into a single workflow. Built for foresight practitioners, policy analysts, and organizations that plan in years not quarters.

Foresight Scenario Planning STEEP Strategic Intelligence

Because the gap in state AI guidance is real and the deadlines aren't waiting.

Two free frameworks — one for K-12 superintendents and principals, one for university presidents and provosts. No paywall. No consultant required. Of the 34 state AI education documents currently in circulation, zero include a named implementation role with a job description. These do.

AI Policy K-12 Higher Education Implementation Framework Open Access

View K-12 Toolkit · View Higher Ed Toolkit

Neighborhood discovery for people who are tired of being algorithmically sorted into loneliness.

Fandom-level interest matching within proximity bands. No DMs. No groups. No events. Just the 100 people closest to you who care about the same things you do. Starting in Montrose, Houston.

React Native Proximity Matching Fandom Interests Horizontal Scroll No-DM Architecture

Make AI's water footprint visible without pretending an estimate is a meter.

Cross-platform AI water-impact estimator. Uses an uncertainty-first, scenario-based methodology rather than a universal water-per-token claim. Open source, MIT licensed.

HTML JavaScript AI Ethics Environmental Accountability

Crowdsourced field intelligence for hunters who are tired of burning diesel on bad intel.

Parcel boundary overlays, offline mode, ArcGIS satellite tiles, crowdsourced field intel posts. Freemium: Scout (free) → Field Pass ($4.99/mo) → Landowner ($19.99/mo).

Leaflet.js ArcGIS Regrid API Supabase IndexedDB Offline-First

Essays at RLPerspectives.com — Confessional before analytical. Structurally chaotic in the way that things that are actually true tend to be.

• The Math Says The Kids Are Right
• Leftists! Governments and AI Companies Are Designing Your Future Without You
• The Boring Couple, and the Church That Never Met Jesus
• Propaganda Machines

Econometrics & research: Stata · R (lmtest, sandwich, survival) · Python (pandas, NumPy, SciPy) · Newey-West HAC · reproducible data workflows · Harvard Dataverse · SSRN

Agent / systems infrastructure: Model Context Protocol (MCP) · Git/GitHub · GitHub Actions · Vercel · Supabase/Postgres · Next.js · Node.js · SQL · REST/JSON APIs · transactional queues/leases · AgentMail · Undermind · AppDeploy (legacy/optional runtime surfaces)

Foresight: Schwartz Scenario Planning · Causal Layered Analysis (CLA) · Three Horizons · Hines HAT · Polak Future Image · STEEP+ · cross-impact analysis

Web / product: JavaScript · HTML · Leaflet.js · ArcGIS · IndexedDB · offline-first patterns

Writing / publishing: Substack · ORCID-linked publications · open data archiving

22 years as a U.S. Navy Senior Chief — operations management, leadership, and systems-level problem solving at scale. Veteran-owned business (RL Perspectives, LLC). Population health minor, because economic models that ignore bodies are incomplete. Everest Base Camp, 2024. Houstonian by choice.