#6 Automation, Dumbing-Down, Business Models, & 2 Idiots on the Balcony • Zettelkasten Live

Automation is a double-edged sword for digital knowledge work: offloading “thinking” to software can drain the user’s competence, increase error rates when concentration drops, and ultimately make people worse at the very craft they’re trying to improve. The central claim coming out of the discussion is that knowledge work needs an environment that keeps the user responsible for judgment—while still letting tools handle small, helpful actions—so the work stays a skill-building exercise rather than a button-pushing routine.

The disagreement begins with feature requests for automation inside writing software (including automated suggestions for links and tags). Christian’s critique, as summarized in the conversation, is that automation transfers intelligence from the person to the program. That transfer becomes a structural obstacle: when tagging, linking, or other decisions are delegated, the user stops practicing the judgment that makes the system useful in the first place. The practical warning is also blunt—when people are tired, they’re more likely to make mistakes, and a long day can unravel after a bad run. The proposed countermeasure is not “no automation,” but an interface and workflow that force deliberate decision-making while using software to make actions easier and more pleasant.

A second principle targets “clever” software design that becomes hard to escape. The approach being advocated is software agnosticism: even if the developers build their own tools, the underlying workflow should remain portable. The conversation repeatedly returns to plain text as the anchor—writing in text files, using external editors, and relying on widely available tools—so users aren’t boxed into a proprietary ecosystem. This portability isn’t framed as ideology alone; it’s positioned as business risk management. If a service like Evernote disappears, users can be left with missing data or dead workflows. In contrast, text-based archives can survive software churn.

The discussion also broadens into a critique of management-style “pre-canned solutions.” Drawing on Matthew Crawford’s Shop Class as Soulcraft, the argument is that centralizing process knowledge in manuals and systems lets organizations discard experts and replace them with cheaper labor. Applied to knowledge tools, the fear is similar: if software dictates the process, users don’t need to be intelligent, and the work becomes a degraded routine—like manipulating documents and spreadsheets without real understanding. The goal is to preserve “teeth” in the learning process: environments that show possibilities and best practices without forcing one rigid path.

The conversation then adds a real-world business-model angle. There’s news about a Ulysses writing-app clone allegedly copying Ulysses’ look, feel, and user experience—an example used to argue that copying a polished interface is a fragile strategy. The more durable strategy, the participants suggest, is building capabilities on top of simple, replaceable foundations (plain text and interoperable tooling), so the business doesn’t rely on a single proprietary experience.

Finally, the long-term plan is laid out: build a “scholarly machine” (a multi-tool system) that supports the full arc of knowledge work—from information gathering through publishing—without locking users into one tool. The project is framed as boring-but-serious business planning, with an emphasis on replaceability, community around text-based workflows, and cognitive surplus used for building tools that improve others’ lives rather than chasing maximum profit.