Andrew Morral

Gun Policy in America; Director @GunResearch. Greenwald Chair in Gun Policy; @RANDCorporation.

[Andrew Morral @AndrewMorral]

RAND's Gun Policy in America Project, with support from @Arnold_Ventures, has publicly released several sets of data, code, and other resources that gun policy researchers may find useful. This thread highlights what we have made available to date...

[Andrew Morral @AndrewMorral]

Reminder that the nomination period for the $5,000 Greenwald Family Award for Firearm Violence and Injury Prevention Research Excellence closes in one month. Peer-reviewed papers published within the past two years are eligible for this prize.

https://t.co/sZGysWmfAA

[John Arnold @johnarnold]

After many conversations over past year with friends, business associates & policymakers about the future of AI job disruption, I’ve tried to get my thoughts in order. With the caveat that I have no specific AI expertise, here they are. Comments and corrections encouraged.🧵
1/n

[Andrew Morral @AndrewMorral]

@DavidJo30329121 What data are you modeling here as ownership?

[Andrew Morral @AndrewMorral]

The 2026 Greenwald Family Award is now accepting nominations of research papers that have most advanced understanding the nature, causes or prevention of firearm violence or other harms.

Nomination instructions for this annual $5000 prize are here:

https://t.co/xrcGrxklv2

[Andrej Karpathy @karpathy]

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.

[Andrew Morral @AndrewMorral]

Good stuff

[David Hogg 🟧 @davidhogg111]

Whoever is the 2028 nominee needs to have a plan to fix this and a many of the other cuts Trump made to science and research. Massively cutting NIH research is incredibly myopic. It hurts Americans health, our security and economy to not be a leader in research.

[kitze 🛠️ tinkerer.club @thekitze]

me and claude code all day every day

[Andrew Morral @AndrewMorral]

https://t.co/XPcXVfq1XJ

[Andrew Morral @AndrewMorral]

This is very damaging to US scientific infrastructure and progress. I feel terrible for young researchers trying to get started in this environment. This will have long lasting effects.

[Andrew Morral @AndrewMorral]

Many thanks to @EpiProfCharlie for his supportive comments in this piece!

[Andrew Morral @AndrewMorral]

These projects have been immensely important to me, and I'm sorry to see them come to an end. But so long as gun violence remains a leading cause of death and grief in this country RAND will work to bring evidence and policy solutions to address it.
https://t.co/rbV2V7rBdW

[Andrew Morral @AndrewMorral]

@gregmidgette

[Andrew Morral @AndrewMorral]

Huge state differences in suicide/homicide rates are often explained away by, for instance, economic conditions, demographics, firearm ownership rates, political lean, etc. How plausible are those claims, and what portion of state mortality rates cannot be so explained? [1/2]

[Andrew Morral @AndrewMorral]

I think our methods here are useful for thinking through what covariates may be important to control for in gun policy analysis. We wrote them up here (and thanks to my coauthors, Greg Midgette and Terry Schell!).

https://t.co/JVY6RULUg3

[Andrew Morral @AndrewMorral]

An excellent review of our Science of Gun Policy report released last week by Gina Napoletano @WeAreTheMighty.

https://t.co/OJox4EEo29

[Andrew Morral @AndrewMorral]

We are very grateful to @Arnold_Ventures for supporting this initiative for the past 8 years!

[Andrew Morral @AndrewMorral]

We just released our last scheduled update to Science of Gun Policy, our report and web pages on the known effects of gun laws. 🧵 https://t.co/uN7C6qsR6S

[Andrew Morral @AndrewMorral]

This work has been incredibly important to me. When we started, there was remarkably little evidence on gun policy—despite enormous stakes. Here’s a look at what a decade of focused research has achieved: https://t.co/19Uw0Z8ytF