Hello! 👋

My name's Ryan and I'm a software engineer based in New Jersey. I use he/they pronouns.

I work for the Seung Lab at the Princeton Neuroscience Institute on FlyWire, and other petabyte-scale connectomic projects. We volumetrically image nervous tissue at nanoscopic resolution at the largest spatial scales possible, as fast as possible, to observe, reconstruct, and understand the structure and connectivity of nervous systems. See some of my thoughts about this and other stuff below.

Previously, I worked at Zocdoc and Lingraphica, and studied computer science and engineering at Rutgers University.

You can find me on:

• Twitter
• LinkedIn
• GitHub
• Chess.com

Have a good one!

Thoughts

Updated 2024-10-20

Generally

I think the human brain is the most interesting object known to humans, and studying it is perhaps the most difficult and most important scientific challenge of our time. I also believe the development and application of powerful computing and artificial intelligence is perhaps the most important and difficult engineering challenge of our time, and that general, powerful, and safe AI will be the critical factor in solving the great challenges we face, including: the climate crisis and abundant energy; global human and animal health, well-being, and freedom; and understanding the fundamental nature of this universe and its contents, including ourselves.

On Natural Intelligence

I work to accelerate our understanding of nervous systems at the physical level of cells, synapses, and circuits, but I'm deeply interested in the work understanding brains and the nature of intelligence within each level of analysis, and am especially interested in ideas which may bring together our understanding across those levels.

On Artifical Intelligence

I think deep learning is working really well, and is currently being scaled to create extremely powerful and general AI systems. Even current systems provide capabilities I thought were decades away a few years ago. The frontier of AI is accelerating rapidly, and it's possible systems will exist in a number of years that exceed human capability in many or all domains. However, I also see major lapses in the abilities of the most sophisticated current systems to generalize out of distribution to truly novel tasks and acquire new skills, and it's very possible very novel ideas and approaches are needed to create systems that are truly "generally" intelligent in the way that animals are. Scale might be all we need, but it might not.