A portfolio of experiments at the intersection of human endeavor and AI
Identifying and addressing unexpected and undesired outcomes caused by AI.
Discovering and building AI-enabled workflows that were previously impossible or uneconomical.
As the lab's touchstone, the Crossfactors Framework is an comprehensive repository of psychological engineering, cognitive effects, technological principles and human factors that often have unexpected and undesired effects.
A summary can be found below.
User-generated input from a variety of sources is parsed to organize information that would otherwise be difficult to find. LLMs are used to both parse and structure the information as well as derive consensus in the presence of conflicting information.
PDplease.com was built as an illustrative project - a directory of PD-friendly optometrists.
Generate well-defined data structures and composite types from unstructured data, in real-time given prior loosely structured or incomplete examples.
Plumpager.com was built as an illustrative project - a platform that creates secondhand for sale listings in seconds from narrated video, including illustrations, descriptions and prices.
Authenticates the provenance of text as human-written based on typing metrics and generates shareable, verifiable assets.
Typecha.com hosts the platform, a data analysis server and a browser based text editor. C2PA integration currently in the works.
Breathing new life into early internet content trapped in archaic platforms and orphaned on the modern web. The focus is on evergreen technical content from domain experts.
More details TBA...
I am happy to respond to relevant inquiries.
(Link opens the contact form on my personal webpage)