What research taught me

I worked on the AI Risk problem with Derek Anderson @Mizzou (Columbia, MO) with the Center for Geospatial Intelligence, and we worked closely with the Naval Research Laboratory. Shoutout to Derek, Jim, Fred, Chris, Grant, Curt, Charlie, Bryce, Jeff, Matt, and many more who helped me along the way.

“I do not love the bright sword for its sharpness, nor the arrow for its swiftness, nor the warrior for his glory. I only love that which they defend.” -J.R.R. Tolkien

I learned six years ago that AI has the potential to destroy humanity, so I buckled down and spent the next half-decade doing everything I could to fight this threat, until I burned out.

During this Process, I learned a lot about AI, but I learned even more about the Process of science and what it takes to solve a massive problem. My half-decade of research and five gray hairs culminated in a 73 page thesis that will inevitably fade into obscurity.

Science is awesome.

I learned how to take a huge problem and break it down into the smallest actionable steps possible. You can’t solve something as big as AI in one step. The most influential science papers in the world are just a slightly bigger step forward than most other papers. That’s all that is needed, a bunch of people making miniscule amounts of shareable, verifiable progress at a time. Then all of a sudden, boom, we’ve got Large Language Models living in our cellphones that are smarter than anyone I’ve ever met.

I learned how to get my cartoons published. It’s pretty easy when you pair them with months of extensive experiments testing new math, excruciating attention to detail and sources, and effective copy that conveys the correct ideas – mainly, the idea that cartoons are cool! I’ll leave it up to you to find them in my Thesis. ;)

I learned how to convey technical math concepts using no equations, just pretty colors and cool diagrams. Hey, even us enginerds love to be entertained during presentations, not bored out of our minds.

I learned how to write. I’ve always found science papers extremely boring, but it’s actually quite the skill to make them that boring – you have to be very accurate, thorough, attentive, and logical in your storytelling, to create the effect that everyone knows what is being said and the result is the most obvious thing in the world!

Oh yeah, I guess I learned about computer science too. But really, coding is the easy part of AI research. Hell, I barely learned Python, and had tons of trouble just getting experiments running. But now I know how to use a computer from the bottom-up (okay, from the Terminal and C up, I don’t “assemble” anything).

I’m a Master of science, muahaha! I wouldn’t call myself a computer scientist, nor an engineer. I’m more just a nature lover who pays really close attention to details, and is unemotional about my personal connection to the fickle ideas I currently hold in my head. I’m a learner, I’m probably wrong about everything I currently know, and I’m 100% okay with that because I know the Process of science, and thus, can learn more about anything in this universe.

M.S. Thesis

Journal Articles

Conference Articles

Poster Presentations