Disclosures

What I will not write

• No opinion without evidence. If I cannot show the basis for a claim, I do not make it.
• No client or confidential specifics. Operator examples are anonymized; I do not name systems that are not mine to discuss.
• No predictions, futurism, hot takes, or engagement bait. That is trend-following dressed as insight, and it is the thing I write against.
• Nothing about my family or my personal life.

How I use evidence

Every piece carries at least one concrete artifact: a chart from an evaluation, a named clause of the regulation, a primary-source document, or a metric from a deployment. Where the work is not public, the evidence is generalised and anonymised. Where it is public, such as peer-reviewed research, it is cited so you can check it. If a piece has no evidence behind it, it does not go up.

Conflicts of interest

I build and ship AI as well as write about it. My affiliations, stated so you can weigh what I say: CTO at Levana (regulated health), Tech Lead at Foreveryone (social sector), and founder of Legalithm (legal and compliance software). I do not write to promote them, I do not use a founder or executive title as the credential for a claim, and every argument here is meant to stand even if any of those ventures closed tomorrow. Where a topic touches one of them directly, I say so in the piece.

Use of AI

I use AI tools in my work and sometimes in drafting. The judgement, the evidence, the argument, and the conclusions are mine, and I stand behind them as my own.

Contact

Corrections and questions: hello@pedrammadani.com. If something here is wrong, tell me and I will fix it.