A collection of notes and articles at various stages of development.
Watching Maggie Appleton's GitHub Universe talk on agent collaboration. Running notes on alignment as the new bottleneck, the ACE environment, conversations under the hood, and a comparison with Engelbart's NIC.
Watching Doug Engelbart's 2004 Accelerating Change keynote, 36 years after the demo. Running notes with Claude on tool systems vs human systems, collective capability, dimensional scaling, and what a post-conversation augmentation field could look like.
Watching Engelbart's 1968 'Mother of all demos' live, in 2026. Running notes, with Claude as an asymmetric chevruta partner fetching references while I think aloud.
How conversation designers are evolving into LLMOps analysts, and why linguists are uniquely suited for monitoring, evaluating, and improving GenAI systems in production.
Three soft layers between a user and hallucination all fail silently. Three hard defenses make confabulation visible, measurable, and refusable without suppressing the answer.
Annotation of So, Cheng & Krishna Murthy (2026). CHI EA '26. Proposes a spectrum of LLM interface metaphors from anti- to hyper-anthropomorphism to encourage critical engagement.
Capturing ideas from the Guildford session and shaping them into a talk.
Annotation of Masure (2019). Interface Critique no. 2. Argues against user-centered design as a reductive framing, tracing from Xerox Star through Don Norman to experience design.
Extracting meaningful entities and relations, combining thematic analysis with TAO topic mapping in three passes.
Five short experiments validating new Claude features (March 2026 release) on the digital garden. Practical demos 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.
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.
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.
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.
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.
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.
Articles
Field Notes
Seeds
Weblinks
Videos
