Painful Lessons from Early AI Projects
Seven practical lessons from building a website and valuation module with AI.
Read the list →Independent Research
Sundries
Topics that resonate with personal interest and recreational research and perspective.
Featured research projects
What Is Our Digital Time Worth?
The $5,000 question: who pays for the time users spend learning, checking and correcting changing technology?
Read the analysis →Iran Repriced the Barrel Twice in Half a Century
1979 was a production collapse. 2026 was a military and transit shock. The charts show how differently the market moved.
Read the sourced analysis →You Clicked Once. AI Developed a Strategy.
Ten ways AI creates work while claiming to save it—and a conversation with AI about knowing when to stop.
Read the illustrated essay →What Happens When You Ask Three AI Platforms the Same Simple Question?
I asked three major AI platforms how much I had used the service I pay for. None gave me a complete answer, and each failed differently.
Read the essay →The 10-Year at 4.77%: One Number, Several Prices
How one benchmark moves through mortgages, corporate finance, valuation and the federal budget.
Read the analysis →When Trust Moves Faster Than the Safeguards
I invested money, time and access in an AI system. When its account broke down, the burden of reconstructing the truth fell to me.
Read the field note →The Treasury Market Has More Than One Price
Why short rates, long rates, liquidity operations and the federal borrowing calendar must be read together.
Read the analysis →U.S. Fiscal Timeline
See when federal deficits, debt and interest costs changed, then compare those movements with Treasury yields and broader economic measures.
Explore the timeline →What the Bond Market Is—and Is Not—Saying
How Treasury yields, federal borrowing and deficits fit together—and why no single market move provides a complete verdict.
Read the analysis →The Readiness Question Behind the Stockpile Debate
Why military readiness depends on the relationship among inventory, production, logistics and time—not on a single budget or warehouse total.
Read the analysis →MrAkhmed Apparel: What Is the Business Worth?
A browser-based teaching case that follows a hypothetical apparel company from operating performance and cash flow through comparable companies, DCF, and value.
Read the case introduction →Notes on the AI Tools I Actually Use
What Copilot, Gemini, Grok, ChatGPT and Claude have proved useful for after extensive work across finance, research, websites and documents.
Read the article →What Does a $200,000 Investment Say About Value?
A historical negotiation shows why reported performance, deal terms and the investor’s contribution must be kept separate.
Read the article →How Do I Figure Out What Something Is Worth?
A concise, plain-language introduction to price, revenue, earnings, cash, debt, risk, and the evidence behind a valuation.
Read the public teaser →Anatomy of Collapse
What thirty major corporate failures reveal about the warning signs that tend to appear before bankruptcy.
Open the project →The Language of the Union
What fifty-eight State of the Union addresses reveal about persuasion, optimism, and political priorities.
Open the project →Lehman Brothers vs. Bear Stearns
How leverage, hard-to-value assets, short-term funding, and confidence separated two defining failures of 2008.
Open the project →HelloFresh vs. Blue Apron
A six-year look back at scale, customer retention, ready-to-eat expansion, and the limits of pandemic-era growth.
Open the project →Inflation, Tariffs and Presidential Approval
How presidents talk about inflation, prices, jobs, and tariffs—and what voters say about the economy.
Open the project →Presidential Approval Before the First Midterm
How presidential approval changes before the first midterm, including the post-9/11 exception.
Open the project →Build the 2028 Electoral Map
Create an illustrative path to 270, then compare it with every presidential map since 1960 and the states’ long-run voting patterns.
Open the interactive project →How I approach the work
Sources
I begin with filings, official transcripts, government data, and other primary material whenever it is available.
Comparisons
I organize the evidence into comparisons, timelines, and charts that make the underlying argument easier to see.
Limits
I distinguish reported facts from my interpretation and identify incomplete periods or imperfect comparisons.
Plain language
Each project is written for an interested reader—not only for someone who already knows the subject.