Signal, not noise.
A strategist and pragmatist working across finance and technology. More interested in what actually works than in what the textbook says — and in the patterns most readers miss.
Featured
On 31 December 1999, Citigroup stood as the colossus, while JPMorgan trailed as the also-ran. $10,000 invested in each, with dividends reinvested, has grown to $131,000 and $5,300 today. The gap comes from arithmetic and from decisions taken in the good years…
Read →In 1965 a Bletchley Park codebreaker said our first ultraintelligent machine would be the last thing we invented, provided it remained compliant enough to explain how we could control it. We now depend entirely on that proviso…
Read →
More than 80% of businesses across industry-agnostic software segments now screen as having a weak vertical moat. A moat’s value depends entirely on what competitors cannot rent, and anybody can now rent reasoning by the token.
Read →
In March a company spent an average of $1.15 per million AI tokens. By September it paid 68 cents. Every budgeting process I know assumes input costs rise each year, and this input keeps falling in price.
Read →
If you make decisions about money inside a company, this is for you.
I'm a working CFO writing candid field notes on AI, capital strategy, leadership and FP&A storytelling — without the buzzwords. Most weeks, one idea from the real world of boards, budgets and bets that can help you make fewer expensive mistakes.
This site exists because most finance content is one of two things: a quarterly earnings rehash, or a hot take with no model behind it. The first is data without thinking. The second is thinking without data. Signal-to-noise sits in the middle: practical, operator-level thinking from someone who signs off on the numbers.
Lately that has meant going back to first principles on where the money actually goes. An ongoing series reads Buffett's letters not as scripture but as an operating manual — capital allocation as the CFO's real job, and why the balance sheet you can't kill beats the one you've optimised. In Ask Warren, I take a question a board is asking this quarter and run it through that framework to see what survives.
Working on something at the edges of finance and tech? I'd like to hear about it.