Hi, I'm Lester Roberts

I'm an economist with a background in data analytics and AI. I'm passionate about making the world a better place through public policy. I'm also into fencing, piano, and pickleball. Check out some of my art and projects!

Portrait of Lester Alaric Roberts

What you'll find here:

  • Projects — HTML experiments and small web apps.
  • Art — A collection of my artwork.
  • Issues — My thoughts on policy issues.
  • Book Recs — Some of my favorite books.
  • Hire Me — Why you might want to employ me.

Featured

Excellence in Economic Analysis Award
Federal Communications Commission

Excellence in Economic Analysis

Awarded the FCC's Excellence in Economic Analysis Award for my work making the Emergency Alert System more resilient, enhancing public safety.

View Official Document →
UC Fuzzy Logic Competition
University of Cincinnati

1st Place — Fuzzy Logic Competition

Won 1st Place, most innovative, most computationally efficient in the UC Fuzzy Logic Challenge.

View Announcement →

Selected Works

Digital Paintings, Sketches, and 3D Renders

I've loved sketching for a very long time, and have since branched out into digital illustration and 3D modeling. I draw a wide range of things - whatever strikes my interest. Often that's ships, lizards, sci-fi or fantasy soldiers, or portraits.

Projects

HTML web apps. You can find these in the /projects folder on the website's GitHub.

Game Map

X Defense

Defend Earth from an Alien Invasion. LLM-driven enemy strategy, roleplay, and battle resolution. Uses RAG to feed the LLM prompt world state information to maintain consistency.

Formatted for Desktop

Requires an LLM API key
Game Logo

Chinese Language Learning

Learn Chinese through a customizable LLM-driven chat interface. Combines an LLM chat interface with an open source dictionary. Extremely versatile: customize scenarios, NPCs, player information, and a vocab list for the LLM to emphasize.

Requires an LLM API key

On the Issues

When someone burns fossil fuels, the rest of society pays a small cost in the form of environmental harm, a harm not priced into the decision to burn those fossil fuels.[1] A Carbon Tax, Cap and Trade System, or hybrid of the two is one of the most effective ways of dealing with this harm and protecting the environment.[1] It makes those who consume fossil fuels account for those costs, and the money doesn't disappear: it can be repaid out as a dividend to everyone, compensating them for the environmental harm and offsetting the typical regressivity of energy taxes in high-income countries, or invested in infrastructure or renewables.

With some energy intensive technologies, like AI or Cryptocurrency, a common critique is that they need to be regulated, restricted, or banned as they use too much energy. I prefer, instead, a solution that treats this source of harm equally across all uses. A carbon price would account for some of the potential environmental harms of these technologies, encourage them to use renewable sources, while being even handed: not favoring one type of economic activity over the other.

Many jurisdictions already have some sort of Carbon Pricing scheme.[3] They are not universally popular: polluters don't like having to pay the price and voters tend to prefer less visible costs like pollution and regulatory drag over very visible costs like taxes. Australia and Canada have both rolled back Carbon Pricing schemes.[4] However, Carbon Pricing remains one of the best tools available to tackle the biggest environmental problems of our time.

  1. Carbon pricing is widely considered by economists the most economically efficient way to reduce emissions: see Carbon price (Wikipedia); the 1997 Economists' Statement on Climate Change signed by over 2,500 economists including nine Nobel Laureates; and a 2024 Nature Communications meta-analysis of 21 carbon pricing schemes finding 17 caused "immediate and substantial emissions reductions". ↩
  2. As of 2025, 28% of global GHG emissions are covered by carbon pricing initiatives across dozens of national and subnational jurisdictions. See World Bank Carbon Pricing Dashboard; Carbon price — Scope and coverage (Wikipedia). ↩
  3. In Australia, the Clean Energy Act 2011 carbon pricing scheme was repealed on 17 July 2014 by the Abbott government. See Carbon pricing in Australia (Wikipedia); In Canada, the federal consumer carbon tax was set to 0% effective April 1, 2025 by Prime Minister Mark Carney; the industrial carbon tax (Output Based Pricing System) remains in place. Carbon pricing in Canada (Wikipedia). ↩

The best taxes are market-correcting, like Carbon and Tobacco taxes, raising revenue by penalizing activity that harms others. The worst taxes are market-distorting, penalizing behavior that benefits others. Property tax schemes typically tax the value of the land + improvements, so turning an empty lot into something economically productive raises the cost of holding that land, penalizing productive activity. Land Value Taxes are better than Property Taxes as they only tax the value of the land.[1] Given that, for most jurisdictions, land creation is a nonstarter, this minimizes the economic distortion of the tax.[2]

I would encourage local municipalities to, generally speaking, favor Land Value Taxes, and switch from a Property Tax to the revenue-equivalent Land Value Tax where feasible.

  1. The supply of land is essentially fixed (perfectly inelastic), meaning the burden of an LVT falls entirely on the landowner and cannot be passed to tenants — a key reason economists consider it nondistortionary. See Land value tax — Efficiency (Wikipedia) ↩

Like many people, I am concerned about the potential for AI to consolidate more economic power into the hands of fewer people. In previous economic revolutions, automation of human tasks has largely been to the benefit of humanity. More capital-intensive forms of production have allowed for more people to work in more productive, higher-paid jobs, and led to greater economic prosperity. The rise of Generative Intelligence threatens to be different. As frontier AI models get more capable each year, they become more capable of replacing more and more human labor. Mass job loss as a result of AI is a very real possibility, which would strip people of their incomes and centralize power around datacenters.

AI is only currently possible because of the public commons.[1] It is trained on code, art, paintings, poetry, wikipedia articles, and so much intelligent human activity that was made free and easily accessible for the benefit of humanity.[1] As such, it's important to ensure that the public, humanity, that has made this possible, benefits.

There are various ways to structure an AI windfall tax, some suggest taxing AI company revenues or profits at a higher rate, while others propose a robot tax on automation-driven productivity gains.

  1. Major AI models are trained on large-scale web crawls of publicly available content. The Common Crawl dataset, a nonprofit web archive capturing billions of webpages, has been used to train LLMs from OpenAI, Google DeepMind, and Anthropic. As of 2024, Common Crawl and The Pile were the two main training datasets for AI models. Wikipedia articles, open-source code repositories, and public domain art are also commonly included in training corpora. ↩

The Founding Fathers set the length of Copyright to 14 years plus an option to renew for another 14 years.[1] Current US law, Life + 70 years[2], is absolutely ridiculous.

Copyright is a state-mandated monopoly, limiting competition over a product. State-mandated monopolies can be good. With copyright, authors have a period to reap the rewards of the labor of inventing valuable intellectual property, either by continuing to make products or selling the rights. There is value in this. However, culture and storytelling thrive on the ability to for a whole society to adapt and innovate off of ideas. An incredibly long copyright period suppresses this while providing only a heavily time-discounted benefit to the actual creator. Reduce the copyright window to 14 + 14 years, more than enough for a creator to reap the rewards of their work, and help culture flourish instead of being increasingly centralized under a handful of entertainment giants.

  1. The Copyright Act of 1790, signed by George Washington, provided an initial term of 14 years with one optional 14-year renewal — a total of 28 years maximum. ↩
  2. The Sonny Bono Copyright Term Extension Act of 1998 (also known as the "Mickey Mouse Protection Act") extended U.S. copyright to the life of the author plus 70 years, or 95 years from publication for works made for hire. See Copyright law of the United States — Duration (Wikipedia). ↩

Book Recommendations

A selection of my all-time favorites

Stories, real or fictional, that have left an impact on me through their storytelling, philosophy, interesting concepts, or more. Mostly sci-fi and historical nonfiction.

Roadside Picnic, Cover
Arkady and Boris Strugatsky

Roadside Picnic

A deeply-philosophical science fiction classic that inspired the (also) excellent film 'Stalker'. Brooding, mysterious, and deeply human.

Project Hail Mary, Cover
Andy Wier

Project Hail Mary

Andy Wier's simple prose is a vessel for an incredibly deep technical exploration of some very interesting hard science fiction concepts. He effortlessly blends high-stakes adventure with analytical problem solving in a way almost anyone can enjoy.

Blind Mans Bluff, Cover
Sherry Sontag, Christopher Drew, with Annette Lawrence Drew

Blind Man's Bluff

My all-time favorite book. A series of vignettes that provides a look into the undersea theatre of the Cold War.

1776, Cover
David McCullough

1776

Not only a compelling work of military history but a humanizing character study of George Washington. A look at how he braved, by the skin of his teeth, the opening year (and a half) of the American Revolution.

I, Robot, Cover
Issac Asimov

I, Robot

A speculative look at robots and AI told through a series of little 'detective stories'.

This is How You Lose the Time War, Cover
Amal El-Mohtar and Max Gladstone

This is How You Lose the Time War

A very surreal time travel love story. Memorable in how incredibly abstract and strange it is.

Hire Me!

I'm an award-winning economist looking for interesting work.

What I do:

  • Develop Algorithms to Solve Problems
  • Analyze Data and Deliver Insights

Experience:

Federal Communications Commission

I was priveledged to work along the amazing people of the Office of Economics and Analytics, being able to contribute to critical public safety and public interest work. For my exceptional analytical work, I won the Agency's Excellence in Economic Analysis Award and a Quality Step Increase.

I contributed to the public interest by:

  • Developing a network analysis algorithm to map, visualize, and test for single points of failure in Emergency Alert System distribution hierarchies. I was able to automate processes to detect single points of failure, reducing the likelihood of catestrophic failures in the EAS Network.
  • Performing research, data collection, and analysis for multibillion dollar telecommunications mergers, contributing to the Agency's 'public interest' mandate.
  • Created data pipelines to format, clean, and analyze industry data.
  • Designed market simulations to perform cost effectiveness analysis to
  • Build geospatial visualizations to inform leadership decision-making on topics such as satellite ground station regulations and educational subsidy allocation.

Excellence in Economic Analysis

Me with FCC Commissioner Jessica Rosenworcel

Image Credit: Federal Communications Commission

With FCC Commissioner Jessica Rosenworcel. Receiving recognition for contributions to public safety through analysis of the Emergency Alert System.

Education:

MS in Applied Economics

  • STEM-accreddited coursework. Panel Data, Time Series, Geospatial Analytics

BA in Economics, Minor in Political Science

Skills:

  • Programming: R, Python (matplotlib, scikit, Conda), MATLAB, STATA, Visual Basic, Git, OpenGL
  • Artificial Intelligence (AI): Machine learning, genetic algorithms, fuzzy inference systems, behavior trees
  • Econometrics: Regression modeling, time series, panel data, geospatial analytics
  • Geospatial Analytics (GIS): Geospatial visualization and analysis in ArcGIS, R, GeoDa
  • Business Analytics: Statistical analysis in Excel, use of Microsoft Office Suite (Word, PowerPoint)

Get in touch:

Email me at [email protected]

Contact Me
Loading…