Adventures in Building

I build stuff. Cabinets. Vintage Trucks. Hybrid AI. Bigbee Labs is my personal experimentation space where I work on coding and technology. This is my blog, where I write about all that stuff.

My big project right now is Ren – A Hybrid AI Partner in Care. I’m planning on using this blog as sort of “field notes”. Backgrounders on what I’m doing. Architecture and design concepts. Experiences and learning.

Here’s the Big Picture

Informal Caregivers: Healthcare’s Big Bet explains why I’m focusing on caregiving in the home. TL;DR – Caregiving in the home is a big deal. Caregivers don’t get the support they need. We need to figure this out before the healthcare system implodes.

Ren: Supporting Caregivers at Scale describes the six kinds of support Ren provides – Educate and Translate; Coordinate and Navigate; Advocate and Escalate; Guide, Coach, and Train; Validate and Reassure; and Protect and Sustain.

Ren: A Hybrid AI provides an overview of what I mean when I say “Hybrid AI” and discusses the nine AI technologies and traditions that I am synthesizing to create Ren.

Deeper Dives on AI

Hybrid & Neuro-Symbolic AI: A Different Take discusses what the AI community means by Hybrid AI and Neuro-Symbolic AI (spoiler: the community thinks they’re the same thing). I have a different take (spoiler: Hybrid is much broader than Neuro-Symbolic.

Knowledge Representation and the Epistemic Substrate Ren operates in a messy world. Sometimes information is complete and accurate. Other times, it’s incomplete or ambiguous. Knowledge Representation and the Epistemic Substrate is the key to navigating this world.

Targeted Knowledge Acquisition In simple terms, this is about helping an AI know what it knows, recognize what it doesn’t, and then act on that ignorance in an intelligent, inspectable way.

Epistemic Construction In the past, knowledge engineering has been a major stumbling block for symbolic reasoning. Here’s how Ren uses LLMs to tackle this challenge.

Rants & Ramblings

This is Not a Hammer: Why “Agentic” Systems are not Agentic The term “agentic” is being widely misused these days. One reason these systems are not agentic is that they lack intrinsic motivation.

Human Slop about AI is a lament but also a prompt for assessing slop in the wild.

Friends Don’t Let Friends Say, “AI is going to transform healthcare” is a reminder of why AI is in fact NOT going to transform healthcare.