My Growth-Focused SIPP Has Four Risk Sleeves
My self-invested personal pension is built to maximise growth and convexity, but not by treating every risk as the same bet. Each sleeve has its own job, payoff profile, and risk control.
How I Size Lottery-Ticket Stocks
This note uses three recent small-cap resource signals to explain my lottery-ticket approach. EST and HREE are speculative positions with a small fixed bet size. THX is an operating business, so it gets different risk treatment.
THX: Cheap Cash Flow, Real Business, Real Concentration Risk
Thor is not a lottery ticket in the explorer sense. It is a West African gold producer on low valuation multiples, with the key constraint that one producing asset still does most of the work.
HREE: Big Rare-Earth Project, Bigger Financing Gap
Harena is a Madagascar rare-earth development equity with project numbers large enough to justify a tiny speculative position. It also has the kind of funding gap that can leave those numbers disconnected from the equity for a very long time.
EST: Partner-Backed Copper Optionality, But Still a Lottery Ticket
East Star is not a normal operating business. It is a Kazakhstan exploration ticket where the case rests on geology, partner support, and surviving without getting diluted into irrelevance before any of that pays off.
The Return of the Deferred Builder
For experienced technologists with ideas, taste, judgement, and half-finished projects...
Designing a ‘Get Home’ Wallet
A small exercise in payment resilience.
The Shape of Unicode Corruption Is a Clue
If you are debugging broken accented characters, do not stop at “Unicode is broken.” The shape of the corruption often tells you which part of the pipeline to inspect next.
Samsung, Hiya, and a Very Short Transparency Horizon
Samsung’s `Caller ID and spam protection` prompt leads into a real data-sharing arrangement. I wanted to know who gets my call data. Beyond Samsung and Hiya, the documentation gave me categories instead of names.
How to write security findings for systems that are still being implemented
Security reports on work in progress often fail in one of two directions: caveat-laden mush or false certainty. The job is to make implementation state, evidence, intent, timing, and residual uncertainty legible to both implementors and management.