Changelog

  1. Gandhi 2.0 now holds the whole argument, from causes to first actions

    Gandhi 2.0 now maps the whole argument the analysis finds in the Second Renaissance Forum thread, not just a starting slice: what keeps the problems in place, ten changes with the effects they should have and the risks each brings, the alternatives people raised, what has to happen first and eight possible actions. Robert Bunge's reservations now appear as risks, each with the remedy the analysis proposes, and every idea still links to the posts it comes from. Each question is split into titled parts, feedback loops are drawn as loops, and the overview follows one path from a problem to a first action. Actions name the role someone might take on, but nobody is assigned. Everything remains proposed and worded by the AI, and none of the people credited has reviewed it yet.

  2. Gandhi 2.0 maps the countercurrent's paradox of success

    After looking at an earlier version of the map, Leif Eriksson pointed out that it missed a central theme of his work: movements that succeed within an existing order can be reshaped by the very capabilities that brought them success. Gandhi 2.0 now includes it. The goal's six items are treated as success factors, and the fifth now rests on two conditions: surviving in selection between institutions, and keeping selection within them faithful to what they exist for. A sixth conflict sets these against each other, and a new change sets out Leif's hypothesis that human–AI collaboration and institutional design might alter the selection pressures themselves. Each new idea links to his posts in the thread. It all remains proposed and worded by the AI, and Leif has not yet reviewed the new wording.

  3. See Reason Commons working on itself

    Explore Reason Commons through its own accepted reasoning: one overview, six perspectives and inspectable statements and connections, with original wording, named endorsers and recorded context. This read-only demonstration uses real collaboration data and leaves the current homepage in place while we explore the direction.

  4. Gandhi 2.0 turns a discussion into questions people can inspect

    Gandhi 2.0 adds a proposed map from analyses of a Second Renaissance discussion: common ground to review, the tension around AI dependence, and a possible conflict-guidance test with its benefits, risks and unassigned first step. Every statement and connection remains proposed, with source locators and AI inferences visible; the original participants have not endorsed the map.

  5. Gandhi 2.0 credits its ideas to the forum posts they come from

    Gandhi 2.0 now says whose ideas it draws on. Every statement and connection names the people whose forum posts or essays it comes from, and its page links to each post in the Second Renaissance Forum thread. The map also shows all five conflicts the analysis finds in the discussion, not one; the first, between institutional strength and democratic integrity, leads the overview because it sits behind the problem of central intervention. Everything remains proposed and worded by the AI, and none of the people credited has reviewed it yet.

  6. Commons you can read

    The reasoning demonstration is now a small site of commons instead of one long page. Each commons has an overview, a page for each of its six questions with diagrams drawn only from published connections, a view of every statement and connection, and a record of changes. Ours sits beside a clearly labelled fictional example, Willow Lane Primary, which shows what else a commons can hold: proposed and disputed reasoning, assumptions, open questions and reports from a trial.

  7. The front page says what Reason Commons is for

    The home page and About now open with what Reason Commons changes for a group: talk a goal or a problem through and come away with a shared picture of what you are trying to achieve and what you will do about it. The method underneath is still explained in the Guide, and the main button now leads to the skill. Earlier framing that no longer describes the project has gone from the current pages, and a site check stops it coming back.

    Two skills built for the retired Reason Commons app, ltp-project and ltp-visualize, are retired too; reason-commons is the skill to install.

  8. A sixth landing page to compare

    The landing page gallery has a sixth candidate, Find a Way Forward: it invites you to bring a problem that keeps coming back and follows an illustrated session from understanding it to trying something and returning with what happened. It joins the visitor study now; the evaluation results still cover only the first five. The gallery is now linked from Demos.

  9. A seventh landing page to compare

    The landing page gallery has a seventh candidate, Disagree and Get Somewhere: it turns a disagreement between two colleagues into a question both sides can investigate. Its download button does not work outside Claude yet. It joins the visitor study now; the evaluation results still cover only the first five.

  10. Compare landing pages with inspectable AI evaluations

    The landing-page evaluation lab now compares five designs and the current homepage with scores, quoted evidence, screenshots and judge disagreements. Four fresh Codex contexts assess desktop and phone captures against an explicit product-trial rubric; the findings remain provisional until checked against human comprehension.

    Correction, 20 October 2026: the provisional recommendation to start human testing with Untangle the Problem is withdrawn. Its only install button works only inside Claude, so on the open web it cannot save the skill; the judges read the page text and never clicked it.

Built with LogoFlowershow