Why the best blog posts start from what you don't yet understand

Engineer-blogger Sean Goedecke argues that the point of writing publicly isn’t to broadcast expertise but to acquire it. He deliberately picks topics he only half-grasps—LLM steering, C2PA provenance, blockchain schemes, the actual history of the Luddites—because the act of condensing a fuzzy idea into defensible prose exposes exactly where his understanding breaks down. His test for whether a draft is worth publishing: he should learn at least two new things while writing it, and his conclusion should end up sharper than his introduction. If neither happens, the post gets scrapped.

The forcing function behind this is taking a real position. Goedecke refuses to publish posts that merely agree with someone else or dump unstructured thoughts; every piece has to make a mildly contentious argument, which compels enough research to survive obvious objections. Drawing on his background in poetry, he makes the counterintuitive case that constraint makes writing easier, not harder—a narrow, arguable claim is far more tractable than ‘write what you feel.’ He also embraces changing his mind mid-draft as a sign the process is working, and defends learning-in-public on the grounds that beginners often write better introductions than experts, that popular consensus is sometimes demonstrably wrong, and that transparent credentials keep readers from mistaking him for an authority.

On feedback, he offers a practical note relevant to anyone shipping technical writing: public posts demand thick skin against internet cruelty, but LLMs now serve as a gentler review layer. In his experience they reflexively push you to hedge every claim into meaninglessness and occasionally get things flat wrong, yet they’re genuinely good at flagging technical misunderstandings—and kinder than the average Hacker News commenter.