Anirudh Vemprala

Eventually you gotta wake up and decide the world is yours for the taking.

Ok. Here we go Ari, you know better than this. You’ve been inside the room. You understand how alliances actually function, not just how they’re talked about on cable hits. NATO was never meaningfully consulted here. Not brought in as partners. Not treated as allies whose buy-in mattered. Instead, for years they’ve been publicly dressed down, threatened, and told outright that they’re on their own. When the President of the United States repeatedly questions the value of the alliance, floats walking away from Article 5, and even talks about things like taking Greenland, you don’t get trust—you get hedging. So now there’s a major war raging on their own continent, and those countries are being asked to stretch even thinner for an operation they had no role in shaping, led by a president who has made clear he views alliances as transactional at best and disposable at worst. Of course they’re cautious. Of course they’re calculating risk. And yes—of course they’re worried they’ll be left holding the bag when Trump inevitably changes course or loses interest. That’s not freeloading. That’s rational behavior in response to uncertainty we created. You’re right that some European countries have underinvested in defense. That’s been true for years, and many have started correcting it—especially since Russia’s invasion of Ukraine. But let’s not pretend this moment exists in a vacuum. Trust is cumulative. And it’s been burned down repeatedly. And the idea that this is about “refusing to help the U.S. rid the world of Iran” ignores the bigger strategic picture. European nations are dealing with an active land war, energy insecurity, domestic political strain, and the very real possibility that U.S. commitments to NATO could evaporate overnight. You don’t expand commitments under those conditions—you consolidate. You know this, Ari. And I think you know why this argument doesn’t hold up. But somewhere along the way, you traded that understanding for applause lines. You’ve sold yourself at the altar of popularity instead of leveling with people about the complexity here. Alliances aren’t maintained by ultimatums and public humiliation. They’re maintained by trust, consultation, and consistency. We’ve offered too little of that lately—and now we’re seeing the result.

Trump: "The US can't take of daycare. That has to be up to a state. We're fighting wars. Medicaid, Medicare -- they can do it on a state basis. We have to take care of one thing: military protection. But all these little scams that have taken place, you have to let states take care of them."

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

Not the most important point, but this is a well written corporate memo

Former ICE attorney Ryan Schwank on why he left the agency: "I swore an oath to uphold the Constitution when I joined ICE…I followed it when I resigned…The legally required training program at the ICE academy is deficient, defective, and broken."