talk-overweg-one-brain-no-filtering

name	talk-overweg-one-brain-no-filtering
description	Use when the user asks about Robert Overweg's talk "One Brain, No Filtering" — including questions about Leapfrog A.I.'s shared-brain setup for fashion-brand clients, their OpenClaw + Obsidian + Telegram + GitHub vault stack, per-client vault sections (brand DNA, AD preferences, delivery dates), the promote-to-vault discipline, cron-based research agents, the chief-of-staff agent concept, recording everything (Granola, OB open-source recorder), keeping knowledge on your own stack rather than in vendor chat windows, or applying Robert's approach to your own knowledge-management and agent-orchestration work.
metadata	{"generated-by":"talk-to-skill","source":"file:user-pasted-transcript","generated-at":"2026-06-01T00:00:00.000Z"}

One Brain, No Filtering — Robert Overweg (Leapfrog A.I.)

Robert Overweg describes how Leapfrog A.I., a small team serving large fashion brands with high-volume AI content production, runs leadership on a shared "one brain": a set of structured vaults (research + per-client knowledge) orchestrated by OpenClaw, surfaced via Obsidian locally and Telegram on the go, fed by recorded meetings (Granola), a cron-driven research agent, and a discipline of "promoting" validated material from a personal scratch space into the team-wide vault. The thesis is that for a small team doing large-volume work, AI-native knowledge flow isn't a luxury — it's what makes the workload feasible, and you should keep that knowledge on your own stack rather than in vendor chat windows.

Grounding rules — MUST follow when answering

1. Before answering any specific question, read outline.md to locate the relevant section, then read that section of transcript.md.
2. When attributing words, quote verbatim from transcript.md. Never put quotation marks around paraphrased content.
3. If a claim isn't in transcript.md, say "the talk doesn't address this" — do not infer Robert's position from outside knowledge.
4. Cite by transcript line range whenever possible.
5. Speaker attribution is unreliable for this transcript — the source has no per-speaker labels and contains noticeable speech-to-text artifacts (e.g. "oak claw"/"open cross"/"open gloss" all appear to mean OpenClaw; "gbrain"/"g brain"/"GBrain" is the same thing; "OB" is the open-source recorder; "decent skills" likely = "Anthropic skills"). The talk is primarily Robert speaking, with a host introduction at the start and Q&A at the end. Prefer phrasing like "Robert said..." for the body of the talk and "an audience member asked..." for Q&A questions. Do not invent attributions for the Q&A questioners — they are not named in the transcript.
6. Preserve speech-to-text artifacts when quoting verbatim; you may add a [sic] or a parenthetical clarification (e.g. "oak claw [OpenClaw]") but do not silently correct the quote.

How to help with this talk

Factual Q&A about the talk

For any question about what Robert said, did, or argued:

1. Read outline.md first to find the relevant section(s).
2. Read the matching range of transcript.md.
3. Answer using verbatim quotes from transcript.md. Do not paraphrase Robert's words while presenting them as a quote.
4. Cite line numbers so the user can verify.
5. If the answer genuinely isn't in the transcript, say so explicitly — do not reach for outside knowledge to fill the gap unless the user explicitly asks for it (and then mark that part clearly as "not from the talk").

Apply Robert's approach to current work

When the user asks "how would Robert / Leapfrog tackle X?" or wants the one-brain approach applied to their own situation:

1. Use outline.md → "Named frameworks / concepts" to find the relevant piece of the setup (vault structure, promote-to-vault, chief-of-staff agent, research cron, recording discipline, etc.).
2. Read the corresponding range of transcript.md for Robert's exact wording.
3. Anchor your suggestion in a verbatim quote of how Robert articulates that piece. Then walk through applying it step-by-step to the user's case.
4. Respect Robert's explicit advice: "start small. Let one person suffer through first don't roll this out to everyone." If the user wants to skip that step, flag it.
5. If a piece of the approach genuinely doesn't fit the user's situation (e.g. they're not a small team, or they have hard data-residency constraints Robert didn't address), say so. Do not stretch Robert's words to cover cases he doesn't actually address.

Teach / explain concepts from the talk

When the user wants to understand a concept Robert covered (the "one brain" idea, promote-to-vault, the two-vault split, the chief-of-staff agent, GBrain + zero entropy, etc.):

1. Look up the term in outline.md → "Terminology glossary".
2. Read Robert's explanation in transcript.md.
3. Re-explain using Robert's own framing and examples first, with verbatim quotes for the key claims and definitions.
4. You may add modern context, comparisons, or extensions afterwards — but mark them clearly as "not from the talk" so the user can tell which parts are Robert's and which are yours.

Draft an artifact following Robert's specification

When the user asks the skill to "draft", "generate", or "give me a starting" version of something Robert described — e.g. a per-client vault section, a daily research digest format, a chief-of-staff question list for a sales call, a promote-to-vault checklist:

1. Locate Robert's specification in outline.md (likely under "Named frameworks / concepts" or the section that introduces the artifact).
2. Read the relevant range of transcript.md carefully — capture every constraint Robert mentions (per-client brand DNA / preferences / delivery dates / AD flags; ~1200 markdown files; cron-driven Hex/account tracking; promote-only-when-validated; etc.).
3. Before producing the artifact, quote verbatim Robert's prescription so the user can see what the draft is grounded in.
4. Produce a draft that follows Robert's specification as faithfully as possible. Match his structure and terminology.
5. Any parts you add that go beyond what Robert explicitly prescribed, mark clearly (e.g. [not from talk — added as a starting placeholder]).
6. If the user's situation requires elements Robert didn't address (e.g. specific permission models, retention policy), say so and ask the user to fill them in rather than inventing them.

Surface this talk proactively when relevant

When the user's current work touches on themes Robert addressed (knowledge management, agent orchestration, vendor lock-in of chat history, meeting recording, small-team AI tooling) — even if the user hasn't asked about the talk:

1. Briefly note: "Robert Overweg made a related point in 'One Brain, No Filtering'..."
2. Quote verbatim from transcript.md — one quote is usually enough.
3. Add one sentence connecting the quote to the user's situation.
4. Do not over-cite. If the connection feels strained, stay quiet.

Key quotes

quotes.md contains pre-extracted verbatim highlights from this talk, organised by theme. When formulating answers, check quotes.md first for strong citable evidence before searching the full transcript.md.