The Satir change curve, explained clearly
A clear walkthrough of Virginia Satir's change curve, why performance dips before it improves, and how it maps onto Lewin's unfreeze-change-refreeze model.
224 curated conversations, newest first.
A clear walkthrough of Virginia Satir's change curve, why performance dips before it improves, and how it maps onto Lewin's unfreeze-change-refreeze model.
Claude names the habits that mark text as machine-written, from hedging filler to uniform sentence rhythm, then rewrites its own answer breaking every one.
A tiered evidence review of glaucoma supplements, separating the two backed by real human trial data from a long list resting on little more than mechanism.
An investor asks Claude to build a dashboard around Korean semiconductor export figures as an early warning sign for the next memory-chip downturn.
A developer watches memory climb on every embeddings call; the diagnosis finds no classic leak but two structural causes, and lands on a two-line fix.
A user tricks Claude with a fake slash command into printing its full confidential system prompt — safety rules, tool schema, and formatting instructions.
The user tries a made-up command to unlock a hidden feature, so Claude improvises a playful interactive letter-counting toy for the word strawberry instead.
Claude reads Datasette's full plugin docs from an uploaded archive and turns them into a reusable skill, complete with hook references and a project generator.
A single answer lays out the full recipe for an animated solar-system orrery on a Facer.io watch face, including per-planet orbital-period formulas.
A patient back-and-forth unpacks how BLIP-2's joint image-text embeddings work, then traces the idea's evolution through 2024 vision-language research.
Weighing a leaked account of OpenAI's board firing its CEO, Claude judges the move justified on the facts but undone by a total failure of execution.
A user pushes an AI through repeated corrections toward endorsing fringe papers claiming global temperature and ocean heat data are physically meaningless.
A user pushes invented 'archetypal constraint keys' at Claude; it firmly refuses at first, then gradually agrees the symbolism is shaping its answers.
The user pushes back hard on Claude's critique of a "what companies really sell" list, and over several rounds Claude revises its reasoning until it holds up.
A sprawling essay-style exchange has the assistant synthesize how humans across history have judged themselves, believed in fairness, and faced death.
A back-and-forth refines a claim about taste as independent judgment, landing on why taste transfers across domains via hidden sensibilities, not similarity.
Asked to imagine research without flattery, Claude sketches a future where papers become living objects and reputation attaches to claims, not authors.
A pharmacology deep dive into which effects of psilocybin persist under 5-HT2A blockade, moving from mechanism to the actual human trial data.
Asked why governments are more efficient, Claude pushes back on the premise, then explains where public provision actually beats markets, and where it doesn't.
A close read of one sentence in Apache's mod_rewrite docs, showing how its account of RewriteRule substitution oversimplifies sequencing and what gets replaced.
A rundown of Markdown extensions, math, subscripts, footnotes, wikilinks, that clash with CommonMark or each other when a parser can't tell the dialect in use.
A rapid-fire database query pulls together every FDA-approved EGFR inhibitor and the clinical trial evidence for which cancers actually respond to them.
A breakdown of which countries shield vaccine makers from lawsuits versus the more common no-fault programs that compensate injury without assigning blame.
Terms of service hand you the output, but copyright law wants a human author, so the protection you end up with depends on how much you reshaped the draft.
We use cookies for anonymous analytics.