Tech • AI • Robotics • Game
Debrief of LIE 1 and 2: 3 failures and one lesson
3/10TL;DR
The Emerging Intelligence Lab is being reworked after two uneven first sessions revealed that spontaneous facilitation was insufficient and that future workshops need clearer structure, asynchronous preparation and more concrete participant value.
KEY POINTS
A rough start for the lab
The initiative launched about three weeks earlier with the aim of identifying rules and practices that help collective intelligence emerge. Its organizer acknowledged going in largely unprepared, relying on improvisation that had worked in other fields such as IT, yoga and martial arts. That approach produced more friction than expected and led to the conclusion that enthusiasm alone was not enough.
The concept behind emergent intelligence
The lab’s starting point was the idea that institutions can design rules that let better ideas surface without constant top-down control. Two corporate examples framed the approach: Jeff Bezos reportedly favored a rule in which senior staff do not speak first in meetings, allowing junior employees to contribute before hierarchy narrows the discussion, while Elon Musk has promoted leaving meetings when attendance is not useful. Both examples were presented as simple mechanisms to reduce conformity and wasted time.
Two failed experiments in Session 1
The first workshop tested a speaking protocol inspired by the talking stick, in which the person speaking controls who talks next and interruptions are avoided. In practice, participants quickly reverted to overlapping discussion, and the method was judged to hinder rather than improve flow. A second attempt to collectively choose the next session topic at the end of the meeting also failed, as participants struggled to switch mental gears while still immersed in the previous discussion.
Three main lessons learned
The review identified three practical lessons. First, live exchanges need to be combined with asynchronous work, supported by simple, lightweight tools. Second, organizing the sessions required far more labor than anticipated before and after each meeting, including debriefing and redesign. Third, maintaining motivation depends on making each session concretely useful, since participants are trading time that could otherwise go to work, family or other priorities.
Ideas that stood out from participants
Despite the setbacks, several contributions were seen as especially valuable. One was a pre-mortem approach: anticipating disagreements and possible failure conditions while everyone is still aligned, in order to create confidence and clearer stopping rules. Another was the insistence that asynchronous collaboration is indispensable if the lab is to scale beyond real-time conversation.
A lightweight way to decide
A further contribution focused on decision-making. Rather than relying on majority voting, more elaborate systems such as quadratic voting or Condorcet methods, the preferred practical rule was non-objection. Under that approach, a decision moves forward if nobody actively objects, avoiding the burden of full consensus while still checking for resistance.
Session 2 drifted off its objective
The second meeting was meant to examine failures encountered on the path to sovereignty, drawing from roughly ten possible themes. Participants chose one subject, but the discussion widened into philosophy and spirituality, and the original goal of turning specific mistakes into useful lessons for others was largely lost. The outcome was judged weak because it became a general conversation rather than a structured learning exercise.
A new format and weekly rhythm
The lab will continue meeting once a week, for now on Wednesday at 8 PM, but not every live session will automatically be a workshop. The new approach is to avoid forcing a lab format when there is not enough material, reducing the risk of burnout and preserving room for smaller or more private sessions. Future programming will depend partly on participant responses to a short questionnaire.
Three calls for contributions
The next phase centers on three proposed formats. The first is a private prompt review workshop, where 4 to 6 participants would bring an exchange with an AI system to analyze what worked and what failed. The second is a series of structured debates designed to be more constructive than standard online argument. The third is a private exchange on digital second brains such as Obsidian and Notion, with a focus on note-taking practices increasingly coupled with AI.
AI as a practical tool
The upcoming third session is set to showcase an AI employee built through code, with the goal of demystifying programming and showing that non-specialists can now create useful systems. AI was also described as transformative for long-term knowledge management: after nearly 20 years of accumulated notes, AI-assisted search and synthesis were credited with making a book project possible by surfacing and organizing material from that archive.
CONCLUSION
The early setbacks of the Emerging Intelligence Lab have pushed it toward a more disciplined model built on preparation, selective formats and private practical workshops. Its central challenge is no longer simply generating discussion, but turning collective participation into repeatable and tangible value.
Ask a question