Seeds seeds
Evals for a RAG bot: a hands-on loop for conversation designers
A field-tested workflow for catching hallucinations, citation drift, and unsafe outputs on a Langfuse-traced bot, written for designers who have never run an eval before.
Seeds
Ideas waiting to germinate.
seeds
Bootstrapping in the strict, Engelbart sense
Not the idiom. A recursive engineering discipline: A-activity does the work, B-activity improves how A is done, C-activity improves how B is done. The leverage is at C, and almost nobody runs it.
seeds
Collective IQ
Engelbart's central concept: the capacity of a group, organization, or society to address complex urgent problems. The thing he kept naming, and the thing the industry kept losing.
seeds
H-LAM/T: the augmentation system is more than the tool
Engelbart's 1962 framework. Augmentation isn't the tool alone, it is Human + Language + Artifacts + Methodology + Training, co-evolving. UX has spent forty years polishing the Artifact and ignoring the other three.
seeds
Scaling laws for augmentation
Engelbart said scaling is a science. Dimensional scaling applies to organisations, to human-plus-agent collectives, and to the problems we are trying to solve. Each regime needs the form redesigned.
seeds
The proof layer LLMs are missing
LLMs solved the language-translation half of knowledge interop. The verification half is largely unbuilt: provenance, claim attribution, signed assertions, grounding evals, shared verification standards.
seeds
User-user-machine interfaces
Engelbart designed for groups, not individuals. H-LAM/T and the Networked Improvement Community suggest the unit of design isn't user plus machine but user plus user plus machine. A reminder for the Guildford talk.
seeds
Ideas, brainwaves and jottings for Guildford talk
Running scratchpad for the Guildford conference talk. Ideas as they arrive.
seeds
From conversation to application: building on method documents
The next step beyond conversation-driven interaction: using Claude Code to turn method documents into working applications and dashboards.
seeds
Reading notes: bacteria to AI
Reading notes on cognitive assemblages, anthropocentrism, and the case for extending cognition beyond human consciousness, from bacteria to AI.
seeds
A reading log for Bacteria to AI
Some kind of reading log for Bacteria to AI, and perhaps even a reminder to actually read.
seeds
Conversational patterns for human-machine interaction
Schegloff mapped conversational patterns for human-human interaction. We need new ones for human-machine interaction, and conflating the two harms both.
seeds
My changing role in an ultrasmall development team
Being PO, scrum master, prompt designer, tester, eval writer and UX designer in a team of 3 means rethinking roles and processes on a daily, almost hourly basis.
seeds
The return of the button
Claude Code is reintroducing structured, constrained input: buttons, numbered options, limited choices. Which is actually IVR. And that might be exactly right.
seeds
Thematic analysis as interaction model and research method
What if we modeled human-AI interaction on thematic analysis rather than conversation? And what if we used that same method to discover themes in a knowledge garden?
seeds
Chatbots without AI
Can a chatbot-style UI work without generative AI, using decision trees, hyperlinking, and serendipity instead?
seeds
Dutch-aware quantization
Standard quantization damages non-English languages disproportionately. A Dutch-first approach to model compression shows there's a better way.
seeds
Embeddings for knowledge gardens: a research gap
Academic research on embeddings hasn't caught up with how PKM tools actually use them. There might be something worth writing here.
seeds
Knowledge gardens and serendipity
Should a digital garden be searchable and useful, or is there beauty in letting people explore?
seeds
Style guide for AI
Language rules for how I talk about AI in my work.
seeds
The disappearance of authentic voice online
When did we stop caring about who actually wrote the words?
seeds
Writing as a conversation with yourself
A mechanism that catches your train of thought. Something that responds in another space.
seeds
AI as writing partner
What's the right metaphor for AI's role in my writing process?
seeds
AI transparency in content
How should I indicate when content was made with AI, and share the prompts?
seeds
How to build a digital garden
What if I wrote a comprehensive guide on how I built this site?
seeds
Principles behind this garden
What design principles shaped this site, and why do they matter?
seeds