Blog

Shelf with hundreds of casette tapes on it

Transcribing the Kohler Tapes

25 February 2026

I set up a custom AI pipeline to transcribe over 1000 oral history interviews.

The AI Moral Panic

20 February 2026

Current public sentiment around AI bears all the hallmarks of a moral panic.

Topic Toggle

26 January 2026

I built an app for my grandma.

Short Form Video

12 January 2026

I watch a lot of reels. I decided to try and make some of my own.

State of AI - January 2026

1 January 2026

What do today's modern geneative AI systems have to offer?

Five Nights at Freddy's characters in Anglo-Saxon style, generated by Nano Banana

Join Us For a Bite

13 December 2026

My nerdiest project yet? I translate 'Join Us for a Bite' into Old English.

Little robot running through a text world.

Creating an AI Agent to Solve Text Games

8 December 2025

For a class final project I trained an agent to solve text-based puzzles.

Alex Lyman Studios: A Year-End Restrospective

4 December 2025

Where does my Lego account stand at the end of the year?

Bible Semantic Search

19 November 2025

I built a new Bible search engine, powered by semantic search.

MTGPT-2 — Training GPT-2 to Generate Magic Cards

6 October 2025

I finally did something I've wanted to do for 3 years, and trained a Language Model to generate MtG cards.

Getting Into 3-D Game Development

10 September 2025

I got back into game development, and I'm really excited about the progress I've made so far.

Using LLMs as a Tool for Inference in the Political and Social Sciences

8 August 2025

How should political and social scientists use LLMs? In a new paper, my team provides some guidelines.

Automated Detection of Spanish Borrowings at IberLEF 2025

22 June 2025

I entered another academic competition with the goal of demonstrating how well commercial LLMs perform NLP tasks.

Don Asay as a missionary in Manila circa 1965

Preserving Grandpa Don's Mission Journal

18 June 2025

I made an interactive viewer for my grandpa, Don Asay's, mission journal.

Balancing Large Language Model Alignment and Algorithmic Fidelity in Social Science Research

23 May 2025

Aligning LLMs to behave like helpful, friendly assistants is great, but can that hamper their abilities? In new research, we explore the tradeoff between alignment and capability.

Several original Lego creations from my account.

Growing Alex Lyman Studios

7 May 2025

Becoming the World's Worst DJ

I try to gain a new skill: becoming the worst DJ known to man.