I co-founded durable.ai, a platform for synthesizing and maintaining long-running software from requirements drawn up by non-technical enterprise users.
Before Durable, I co-founded Canvas Technology, a Boulder, Colorado startup developing vision-based mobile autonomy for manufacturing and logistics. Canvas was acquired by Amazon in 2019.
Post-acquisition, I was a venture partner at Xplorer Capital, helping the team source and assess early-stage deals in deeptech, AI, and sustainability.
I hold a PhD from the University of Colorado Boulder, where my research focused on visual perception and control for agile autonomous vehicles.
You can read my writing, browse my talks and research, or get in touch.
I found the ACX post on the Midjourney Scanner helpful. My takeaway was that there are limitations to resolution and coverage of ultrasound tomography (for example, ultrasound can’t penetrate bone and struggles with cranial imaging). With large scale and their expertise, Midjourney may be able to squeeze more out of the data. I agree that it’s AlphaFold-like: the data is observational and any model needs to learn or approximate the mechanisms behind how it’s generated.