PoliticStack
A new publication. Does not contain: mainstream echo.
I have to be honest. When it comes to politics, my domain knowledge is pretty basic compared to finance and economics.
No matter how much weekend reading time I put in after a week on the trading floor, I got so frustrated with the context void from the news cycle that I hired a geopolitics tutor. Rob was great. I learnt more from Rob than from any of the institutional experts I had access to because they all talked past my ability to understand them.
And I could ask Rob really stupid questions that revealed just how basic my knowledge was. I didn’t need to impress him.
So when I thought about creating PoliticStack — a separate, MarketStack-adjacent publication — to cover all the politics beats on Substack, I needed it to be something that non-domain experts — following geopolitics for markets or simply to stay informed — could understand.
But I didn’t want it to be an echo of mainstream media. Nobody needs another summary of politics news, even if it’s written by Rising Substack authors. None of that stuff moves markets. Let’s face it, we come to Substack for originality and depth, stuff like:
Built on a confidential memo not publicly available, with direct quotes.
Concrete and newsworthy shifts not found in mainstream coverage.
Mainstream wires have not assembled this forensic chain.
And we’re all starved for context and credibility: author context, claim credibility and conflicts of interest, stuff like:
She writes from a perspective critical of US military intervention and Iranian regime-change advocacy.
The characterisation of intent is the article’s interpretation, not an established legal finding.
The leaked memo’s authenticity cannot be independently verified.
(The author) runs GreenMedInfo, a website promoting alternative health claims, and is personally suing CCDH as part of a group of vaccine-sceptic figures targeted by the organisation — a direct and deep conflict of interest.
All of those lines from The First Edition of PoliticStack.
The aim throughout is the same one that drives the whole project: that someone with no prior domain knowledge should be able to read any entry and understand not just what a piece says, but whether to trust it.
So I built it like this:
How PoliticStack works (to start with)
If you didn’t already see, PoliticStack is a separate publication.
Every week it scans Substack’s leaderboards across every politics beat: U.S. Politics, World Politics (Geopolitics), International, Health Politics, and Climate & Environment — pulling that day’s top 100 Bestsellers and top 10 Rising publications from each.
For each publication, it looks at every article published in the past seven days.
Each article is screened for originality: does it contain original documents, primary sources, or investigative findings that Reuters, the BBC, or the New York Times would not cover? Articles that simply recap known events, offer expert commentary on obvious situations, or are ideological arguments without new evidence are removed.
Each article that passes is checked vs mainstream news coverage to filter out the stories that have already spread.
Survivors are distilled by AI into a four-part note: what the piece argues, why it was chosen, what to watch out for (context), and who the author is. Here’s an example from the first edition containing 9 articles, published moments ago:
All entries are then ranked by editorial importance and ordered by global and market relevance.
Every entry requires no prior knowledge — people, institutions, and context are explained inline.
Claude can make mistakes:
The originality filter is imperfect. It will occasionally pass articles that turn out to have been covered elsewhere, and filter out pieces that deserved inclusion. The web search verification step catches the most obvious cases, but not every one.
The web search step has limits. It catches stories where mainstream outlets used similar search terms. It can miss cases where the same story was covered under a different headline, or where Substack broke a story that wire services then followed under their own framing.
Summaries reflect what’s publicly available. For paywalled articles, Claude only reads the free preview. For articles where the key finding is buried deep in the piece, the summary may miss it.
Author context can be outdated. Claude’s knowledge has a training cutoff. Stated backgrounds may no longer be accurate, and conflicts of interest may be missing.
The importance ranking is a judgment call. Ranking by originality, significance, and evidence quality is an automated editorial opinion. Two equally valid editors might order these differently.
Non-English articles are a weak point. Harder to assess accurately at the screening stage, meaning more false positives and less reliable summaries.
You can subscribe to PoliticStack here.
Visit MarketStack The Edit · MarketStack Terminal
MarketStack is free today. But if you value my work, you can pledge for a future subscription. MarketStack is an independent, anonymous publication summarising publicly available commentary and views from across financial media. Nothing here constitutes financial advice or a recommendation to buy, sell or hold any security. All views are a synthesis of public information. Past performance is not a guide to future results. This publication is not authorised or regulated by the Financial Conduct Authority. The author writes anonymously in a personal capacity.