What I Am Actually Trying to Learn

The through-line in all of this is one stubborn question: where does value sit when the accounts cannot see it?

It shows up in a supplement aisle that cleared at six times revenue and at one, eight weeks apart. It shows up in a finance workflow whose entire saving is invisible on a balance sheet. It is the same question wearing different clothes, and I have not got to the bottom of it yet, which is precisely why I keep going.

The deal work

Five consumer wellness businesses changed hands in six weeks this summer at prices seven times apart. I plotted them, worked out what each buyer was really paying for, and wrote a note on each.

P&G / Thorne5.8x2026E revenueCelsius / Alani Nu<3x2024A revenuePepsiCo / poppi$1.65bnnet of tax benefitsBain / Vitabiotics~$1.2bnno public revenueYellow Wood / Nestlé0.8x2025 sales

Read the full argument, including the chart and where I think I am wrong.

Inside a public company

BODiNinety minutes to three. $784K sized and presented to the MD of Accounting. The assets AI builds that nobody recordsThe essay that came out of it. The best output of a deployment never lands on the balance sheet.

Wall Street’s Trojan Horse

A podcast, in progress

A show about how quietly AI has already walked into investment banking. Not the version in the headlines, with the press releases and the transformation decks. The version where a first-year analyst stops doing four hours of work on a Tuesday and mentions it to nobody.

I am collecting perspectives from two groups: first-year analysts who are living inside the change, and the people building the tools that caused it. They describe the same event in almost completely different language, and the distance between those two accounts is the actual show.

Artwork and first episodes to come. If you are a first-year analyst, or you build these tools, I would like to record with you.