Gonçalo Esteves / building capable systems Field Notes →

The Vault / People

Person

Salim Ismail

how Peter introduces him on tape: "the emperor of exponentials" (EP180 0:01:06, 2025-07-02), "the emperor of exponentials, Mr. Exo" (EP228 0:05:22, 2026-02-09), "the father of organizational singularities" (Moonshots 266 — Planet Labs 1:16), "our globe troder [globetrotter], fath

Updated 2026-11-16.

This week (2026-11-16, checked at source; window 2026-11-10 → 2026-11-16)

  • Two new episodes with him, both read from the decompositions; both have sections below. EP310 (aired 2026-11-11, recorded 10 Nov): he joins "in the press box of the San Francisco Giants at Oracle Park", before "an AAI talk" where he will be "standing on top of the dugout… to do my keynote" (0:22:55–0:23:52). On EP311 he confirms it himself: "I did the last pod from the… San Francisco Giants Press media press [box]" (1:26:32). ⚠️ The event is not named on tape, and searches on 2026-11-16 found no listing for it, so it stays on-tape only. EP311 (live, unscheduled, recorded and aired Fri 2026-11-13): no location check-in for him. He says "I was in Spain last week" (1:10:19), which fits the Madrid summit already dated.
  • Playbook: no new step. EP310's "we can delegate outcomes rather than tasks" (0:47:11) and "AI's time horizon is going to now go faster than the time between board meetings… So who's really running the company?" (0:47:54–0:48:04) are a new frame for the governance argument, not a step. His Delaware AI-corporation prescription, "Do very small functional, very clear, very well prescribed areas… and then slowly broaden it out" (1:41:47), restates step 5 at the scale of a whole company. EP311 has no new prescription.
  • On tape, do not repeat: 🔴 "the average grocery store hires… 700" (EP310 1:52:00; Alex corrected it on air: 26–75); ❓ Copilot "completely collapsed" (EP311 1:17:42); ⚠️ "14 major experiments on UBI… staggeringly successful" (EP311 1:45:04, no list given).
  • The 22→4 workflow numbers he promised on EP308 have still not been published (Substack checked 2026-11-16; nothing new on the Resource Hub that we could verify).
  • Substack: no new post in the window (archive read 2026-11-16). The newest post is still the 10 Nov The Attackers Already Built Their ExO. That makes eleven weekly posts since 1 Sep (Sep 1, 8, 15, 22, 29; Oct 6, 14, 20, 27; Nov 3, 10), none marked paid. Posts have landed on Mondays and Tuesdays, so one dated 17 Nov would fall into next week's window.
  • Resource Hub: 116 results (was 114 on 10 Nov; read 2026-11-16). ⚠️ Two items, both Pro-gated, were reported only in a machine summary of the part of the page we could not read word for word: Feeding Your Agent: Building Better AI Context & Knowledge and Agent Data Checklist: Preparing Your Data for AI Agents, both shown as from the Pro Mastermind of 11 Nov 2026. Treat them as unverified until read verbatim.
  • Shaping Luck: the "48 hours" countdown is still running 24 days after it first appeared (read 2026-11-16: "Founding membership closes in 48:00:00", "Closes in 48 hours"; the timer block reads "02 Days : 00 Hours : 00 Minutes : 00 Seconds"; first seen 2026-10-23). Prices unchanged ($380 "Lifetime access", $400 "Lifetime + Book"). There is still no publisher, release date or pre-order; footer "© 2026 Salim Ismail & John Hagel · shapingluck.com".
  • Unchanged at source (2026-11-16): the next 10x Shift is "LIVE on Thursday, December 3, 2026", 8:00 AM PT, with no price on that page. Plans: Beginner free "for a lifetime", Pro $8,640 / $720, 10x Shift "Your Investment: $100", ExO Leap unpriced. Member counts are still "40,000+", "42,000+" and "40K+", none audited. openexo.com/singularity-exo is still "Now forming", with no price or launch date. The free MVIS workbook door still loads and says "It's yours to use, free." The homepage title still says "ExO Leap founding cohort", with no date, deadline or price in the body. salimismail.com still lists the book as "Spring 2026" / "Coming Soon" (out of date as of this read). Fund I is still "Now Investing" and "opens to outside LPs in 2026". The newest dated item is still from Nov 2025, and there is still no Shaping Luck and no Substack link.
  • Genius Group: still open. EDGAR full-text search for 2026-10-01 → 11-16 (run 2026-11-16) finds zero hits for "OpenExO" and zero for "Salim Ismail". The same search with no date limit returns 65 hits, including the two 18 Mar 2024 GNS 6-K exhibits, so the search works. The position stays not established.
  • News search (2026-11-16): nothing new about him in the window beyond the episodes. No recording of the Madrid, Oracle Park or DBS talks; no board-liability paper (EP293); no Shaping Luck publisher.
  • Open programmes, prizes or calls found this week: none. The ExO Leap "founding cohort" publishes no dates, deadline or price, and the 10x Shift and Shaping Luck are paid products, not calls.

Last week (2026-11-10, checked at source; window 2026-11-04 → 2026-11-10)

  • Two new episodes with him, both read from the decompositions. EP308 (aired 2026-11-04, recorded 3 Nov): remote from Madrid, "I'm in Madrid for a couple of days at a summit. Singularities." (0:14:21; ⚠️ the worst captions in the corpus, round-trip translated, so quote the tape, not this line). EP309 (published 2026-11-07, recorded ~6 Nov): his first episode as host, Peter away ("I am soliam ismail your temporary host for today", 0:03:26), back from Madrid ("I was in madrid for the singularity summit there", 0:04:53). Both have sections below.
  • The Madrid event is now verified at source. Singularity Summit Spain 2026, 4–5 Nov 2026, Espacio Acciona, Madrid (spain.su.org agenda, read 2026-11-10; Esade's sponsor page gives the same dates). He is on the agenda: "The Organizational Singularity", 12:30–13:30, in the Day 2 morning block (so most likely 5 Nov; the page's day blocks are not dated individually), billed as "Founding Executive Director of Singularity University." The agenda says it is subject to change; no recording found.
  • ★ Playbook: one step sharpened (step 6), from EP308. How to move workflows, by size: under ~50 people brute-force it; over 50, "set this up stack of intellectual resources on the side and start moving workflows one by one", with "One company took 22 steps and reduced them to four" (0:54:18–0:54:29). ⚠️ The 22→4 case is unnamed; he promised the numbers "during the next a few weeks", not yet published (Substack and Resource Hub checked 2026-11-10). EP309's "be active" close restates existing steps.
  • On EP309 his lead thesis is institutional: "our political our financial our social institutions are sub-linear" (0:31:15–0:31:28). 🔴 His NYC-vs-Anthropic energy comparison on the same episode (1:10:47–1:12:17) changes its own inputs mid-segment and is barred; see the EP309 section.
  • A new Substack post, 10 Nov 2026: The Attackers Already Built Their ExO (free, read 2026-11-10 via fetch). Argument: attackers already run like ExOs (staff on demand, leveraged assets, algorithms, experimentation, autonomy) and move in hours, while defenders lose weeks to approval boards and hand-offs, so more tools won't close the gap; build a small AI-native unit at the edge, starting with vulnerability triage, with written agent specs, a kill switch and a purpose agents can test instructions against. It repeats playbook steps 6 and 8 and his EP305 agent-controls line; no change to the playbook. Its numbers, all without links: Anthropic's disclosed AI-orchestrated espionage campaign (~30 organisations, AI on "80 to 90 percent of the tactical work"); Verizon's 2026 DBIR (exploitation overtaking stolen credentials as the top way in, flaw-to-exploit "from months to hours", median full patch time "from 32 days to 43"); "The typical defender takes six weeks" (unsourced). ⚠️ None checked against the reports this run; do not repeat until they are. Call to action: the REWRITE Score (openexo.com/rewrite-v2, no price). It also says "If you're at the summit today, come and find me": ⚠️ the summit is not named. Substack now has eleven weekly posts since 1 Sep, none marked paid.
  • The MVIS Stack Workbook free door works: openexo.com/mvis-stack-workbook-for-yt-10k loads ("MVIS Stack Workbook | The Shift 10K"), says "It's yours to use, free" and "the MVIS stack workbook our Pro members use in their weekly workshop" (read 2026-11-10). The Resource Hub still lists its 21 Oct copy as Pro-gated; both are true at once.
  • Resource Hub: 114 results (was 113). The one new-looking item, Slice Discipline by Kent Langley (Leap-gated), carries no date on the page; ⚠️ read through a summary, so not dated here. Nothing dated 4–10 Nov.
  • Shaping Luck: the "Final 48 hours" countdown is still there 18 days after it first appeared (read 2026-11-10: "Founding membership closes in 48:00:00", "Closes in 48 hours"; first seen 2026-10-23). Prices unchanged ($380 / $400); footer "© 2026 Salim Ismail & John Hagel"; still no publisher, release date or pre-order.
  • Unchanged at source (2026-11-10): 10x Shift next session Thursday 3 Dec 2026, 8:00 AM PT, no price on that page; plans (Beginner free "for a lifetime", Pro $8,640 / $720, 10x Shift $100, ExO Leap unpriced); member counts "40,000+", "42,000+", "40K+", none audited; openexo.com/singularity-exo still "Now forming", no price or date; salimismail.com still "Spring 2026" / "Coming Soon", Fund I "Now Investing", "opens to outside LPs in 2026", newest dated item Nov 2025, no Shaping Luck, no Substack.
  • Genius Group: still open. EDGAR full-text, 2026-10-01 → 11-10: zero hits for "OpenExO" and for "Salim Ismail" (control without dates: 20 hits, including the 18 Mar 2024 GNS 6-K, so the search works). The position stays not established.
  • News search (2026-11-10): nothing new about him in the window beyond the above. No board-liability paper (EP293), no Shaping Luck publisher, no recording of the Madrid or DBS talks.
  • Open programmes, prizes or calls found this week: none. (The 10x Shift and Shaping Luck are paid products, not calls.)

Earlier (2026-11-04, checked at source; window 2026-10-29 → 2026-11-04)

  • One new episode with him: EP307 (published 2026-10-31, recorded ~10-30, remote). Peter opens with "it looks like everybody's home today… This is home week on the pod" (0:09:36), but at the close he leaves early: "I I have to run back. I'm on stage at MIT in a few minutes. I have to run out of the door here" (2:15:40) — so on record day he is at/near MIT (Cambridge, MA), on tape only, event unnamed (search 2026-11-04 found no listing). Read from the decomposition; section below. Playbook: no new step — "Go build a business. Do not take an MBA." (1:45:50) restates step 8.
  • On EP307 he moves from disputing Chamath's Bitcoin call to calling his motive: "I'll be blunt. I think it's a scop [psyop]. I think he's working on behalf of the Treasury Secretary… he used to be a massive Bitcoin [bull]… He was playing his book." (0:01:05, after Peter's hand-off by name). ⚠️ An accusation with no evidence offered; capture it, don't repeat it as fact. Consistent with his EP306 Bitcoin defence.
  • He put out his own video on the jobs question (~29–30 Oct): "I put out a video yesterday basically asking a simple question… everybody's bemoaning the… destruction in jobs… If you'll flip that around and say, 'Wouldn't it be the greatest thing ever if all the jobs disappeared?'"; "it's getting like crazy comments" (0:35:57–0:36:27). ⚠️ Platform not named on tape and the video was not found at source (2026-11-04); on-tape only.
  • A new Substack post, 3 Nov 2026: 10x to 100x in 12 months? Here's how mid-sized companies do it (free, read 2026-11-04 via fetch). Argument: AI removes size as the moat; a ~300-person company has three edges (less legacy, a CEO who can still redesign the operating model, cheaper intelligence) and most mid-sized firms are still "dabbling", which is the opening. Its numbers: a Kaufman Rossin / NewtonX survey (94% of US firms with 20–1,000 employees use generative AI tools, 2% have it operating across the business) is ⚠️ linked to a secondary blog (ibl.ai), not the survey; an RSM 2026 survey of 1,030 middle-market leaders (54% say AI beat ROI expectations, 17% enterprise-wide) links to rsmus.com; SaaStr ~$700K ARR per employee at the 90th percentile; ⚠️ Lovable "$1M to $200M ARR in about 11 months" via a third-party write-up. Do not repeat the ⚠️ ones. Prescription: a 12-month quarterly roadmap (MTP + map workflows → rebuild around agents → scale and retrain → launch an AI-powered product) plus a weekly question, "If we started from scratch today with AI, what would our company look like?" — it sequences playbook steps 2, 5 and 6; no change to the playbook. Calls to action: a YouTube video (the linked page's title reads "Stop running AI experiments. Map your stack instead"; channel and date not readable from here, tied in the post to "The Shift" at "10,000 subscribers", ⚠️ unverified) and a free MVIS Stack Workbook at openexo.com/mvis-stack-workbook-for-yt-10k — the Hub's 21 Oct copy was Pro-gated, so a free door to it is new (link not tested). Substack now has ten weekly posts since 1 Sep, none marked paid.
  • Shaping Luck: the "Final 48 hours" countdown is still there twelve days later, reading "Founding membership closes in 48:00:00" (read 2026-11-04; first seen 2026-10-23; the body counter reads a static 02 Days : 00 : 00 : 00). Prices unchanged ($380 / $400); still no publisher, release date or pre-order.
  • Unchanged at source (2026-11-04): 10x Shift next session Thursday 3 Dec 2026, 8:00 AM PST, no November date; plans (Beginner free "for a lifetime", Pro $8,640 / $720 per month, 10x Shift $100, ExO Leap unpriced); member counts "40,000+", "42,000+", "40K+", none audited; Resource Hub still "113 results", newest dated items still the 28 Oct Agent Control Sheet / Agent Controls Presentation (Pro), nothing dated in the window; openexo.com/singularity-exo still "Now forming", no price or date; salimismail.com still "Spring 2026" / "Coming Soon", Fund I "Now Investing", "opens to outside LPs in 2026", no dated item in the window, no Shaping Luck, no Substack link.
  • Genius Group: still open. EDGAR full-text, 2026-10-01 → 11-04: zero filings mention "OpenExO" or "Salim Ismail" (control search without dates returns the 18 Mar 2024 6-K, so the search works). The position stays not established.
  • News search (2026-11-04): nothing new about him in the window. No DBS recording, no board-liability paper (EP293), no Shaping Luck publisher.

Earlier (2026-10-29, checked at source; window 2026-10-23 → 2026-10-29)

  • One new episode with him: EP306 (published 2026-10-28, recorded 10-27, description), remote from Toronto: "I'm in Toronto because I'm doing a talk here to about 400 CEOs tomorrow and uh and then flying back Thursday" (0:11:03). ⚠️ The event is not named on tape and we found no listing for it (search 2026-10-29), so it stays on-tape only. Read from the decomposition; section below. EP305 (aired 10-23) has no location check-in for him. Playbook: no new step.
  • A new Substack post, 27 Oct 2026: The Future of Jobs Will Be Decided Inside Your Organisation (free, read 2026-10-29 via fetch). Argument: young people in AI-exposed jobs are not being fired, they are not being hired; entry-level postings now ask for senior skills; "The entry-level job is still there. It now starts on the third rung." Three levers: workflows first and roles second, rewritten "AI-native at the Edge"; rebuild apprenticeship by putting juniors in charge of agents from day one; make the MTP the job description. His closing test: if AI did everything your juniors do today, how do they become your senior people in 2035? Its numbers: the WEF Future of Jobs figures (170M new roles, 92M displaced, net 78M by 2030; "close to 40%" of core skills changing) match the WEF Future of Jobs Report 2025 (Jan 2025) as widely reported; ⚠️ "fell about 11% between late 2022 and mid 2026" for 22–25-year-olds in the most AI-exposed jobs ("Stanford researchers tracking payroll data") and PwC's "more than half the new skills" are not checked, no links given, do not repeat. Calls to action: the REWRITE Readiness Score (openexo.com/rewrite-v2, no price) and a "completely redesigned" AI chat at openexo.com/chat. The levers repeat playbook steps 2 and 6; the apprenticeship lever assumes an organisation that hires juniors, so no change to the playbook. Substack now has nine weekly posts since 1 Sep, none marked paid.
  • Resource Hub, new dated items: Agent Control Sheet and Agent Controls Presentation, 28 Oct 2026, Pro-gated (read 2026-10-29 from a partial page read; the hub shows 113 results). They follow his EP305 four-controls-per-agent line (below). The 20 Oct 10X Shift deck and workbook still show no lock label.
  • Shaping Luck page: the "Final 48 hours" countdown is still there six days later, reading "Founding membership closes in 48:00:00" (read 2026-10-29; first seen 2026-10-23). The page states no end date. Prices unchanged ($380 / $400); still no publisher, release date or pre-order.
  • Unchanged at source (2026-10-29): 10x Shift next session Thursday 3 Dec 2026, 8:00 AM PST, no November date; plans (Beginner free "for a lifetime", Pro $8,640 / $720 per month, 10x Shift $100, ExO Leap unpriced); member counts "40,000+", "42,000+", "40K+", none audited; openexo.com/singularity-exo still "Now forming", no price or date; salimismail.com still "Spring 2026" / "Coming Soon", Fund I "Now Investing", "opens to outside LPs in 2026", no dated October item, no Shaping Luck, no Substack link.
  • Genius Group: still open. EDGAR full-text, 2026-10-01 → 10-29: zero filings mention "OpenExO" or "Salim Ismail" (control search returns the 18 Mar 2024 6-K, so the search works). New lead, ⚠️ aggregator only: MarketScreener's insider page lists his Genius Group chairman role with an end date of 09/10/2024. Not a filing; the position stays not established.
  • News search (2026-10-29): nothing new about him in the window. No DBS recording, no board-liability paper (EP293), no Shaping Luck publisher.

Earlier (2026-10-23, checked at source; window 2026-10-16 → 2026-10-23)

  • Two episodes aired in the window, both read. EP303 (published 2026-10-17, recorded 10-16), in person in the studio the day after XPRIZE Visioneering in Malibu, with Peter, Emad and guest Anousheh Ansari; and EP304 (aired 2026-10-21), with Chamath Palihapitiya, where he is "actually home" (0:01:39). Both are in Where he is lately; EP303 has its own section below. Playbook: no new step from either.
  • A new Substack post, 20 Oct 2026: Resilience Isn't Getting Back Up. It's Getting Rebuilt. (read via fetch 2026-10-23, not end to end). Argument: resilience framed as recovery is wrong when agents take on the work, because restoring the old organisation means defending something going obsolete. Three levers: fix the purpose and flex everything else; rebuild slack instead of optimising only for efficiency; build a protected AI-native alternative at "the Edge" before it is needed. No statistics, no sources. A board question it offers: "if we had to rebuild this organisation tomorrow, how much of it would we rebuild the same way?" Its calls to action: that day's 10X Shift (11am EDT, no price on the post) and a free building-an-exo skill linked to the Resource Hub (⚠️ no item by that name found on the Hub, 2026-10-23). The levers repeat playbook steps 1, 2 and 8; no change to the playbook. Substack now has eight weekly posts since 1 Sep, none marked paid.
  • 10x Shift: the 20 Oct session ran; the next one is Thursday 3 Dec 2026, 8:00 AM PST (openexo.com/10x-shift, read 2026-10-23; no price on that page). No November date is listed. The 20 Oct 10X Shift Presentation and Workbook are on the Resource Hub with no lock label (dated 20 Oct 2026).
  • Resource Hub, new dated items: MVIS Stack Workbook and MVIS Stack Presentation, 21 Oct 2026, Pro-gated; three ExO Leap: MVIS stack and agent loop Canvas items dated 2 Oct, Leap-gated (older than the window, first seen here).
  • Shaping Luck page now runs a countdown: "Founding Membership · Final 48 hours", "In 48 hours, founding access at this price is gone" (read 2026-10-23). Prices unchanged ($380 lifetime, $400 lifetime + book). ⚠️ The page gives no end date, so the 48 hours cannot be dated; still no publisher, release date or pre-order.
  • Unchanged at source (2026-10-23): OpenExO plans (Beginner free "for a lifetime"; Pro $8,640 / $720 per month); member counts still "40,000+", "42,000+" and "40K+" on one page vs his "50,000", none audited. openexo.com/singularity-exo still "Now forming", no price, no launch date, a "Share your interest" form. salimismail.com: book still "Spring 2026" / "Coming Soon", Fund I still "Now Investing" and "opens to outside LPs in 2026", no October item, no Shaping Luck, no Substack link.
  • Genius Group: still open. EDGAR full-text, 2026-10-01 → 10-23: zero filings mention "OpenExO". Not resolved; the 2025 annual report is still unread.
  • Context on SINGULARITY ExO's CEO, from filings (outside the window): Fermi Inc. (NASDAQ: FRMI) filed an 8-K on 13 Jul 2026 and an 8-K/A on 21 Jul 2026 on Miles Everson's resignation from its board (effective 10 Jul 2026, per a TradingView summary of the filing: "Disagreement over access to certain books and records"). The filings exist on EDGAR (accession 0001213900-26-077385); ⚠️ we read the summary, not the 8-K or its letter. It concerns Everson, not Salim. Recorded in Positions and disclosures.
  • News search (2026-10-23): nothing new about him in the window beyond the above. No DBS recording, no board-liability paper (EP293), no Shaping Luck publisher.

Earlier (2026-10-17, checked at source; window 2026-10-10 → 2026-10-17)

  • He spoke in Singapore on 12 Oct at the inaugural DBS Asia Future Forward Conference. Primary: DBS newsroom release of 30 Sep 2026 ("Salim Ismail, Chairman of OpenExO, will explore how AI could rewrite the corporation"; ~700 DBS clients expected; event date 12 Oct 2026, Singapore). His own account: Substack post of 14 Oct, "Yesterday I was in Singapore for the inaugural DBS Asia Future Forward Conference." (The "yesterday" reads as 13 Oct against DBS's 12 Oct; the post date, not the event date, is the loose one.) Two days later he joined the 15 Oct Moonshots episode "live from Singapore at 2 AM" (0:01:12).
  • New surface for this page: his Substack (salimismail.substack.com, read 2026-10-17). Seven weekly posts since 1 Sep 2026, all free ("Subscribe for free", no paid tier shown): 1 Sep Everyone saved time. Nobody can find it. · 8 Sep Your Agents Inherit Whatever You Failed to Write Down. · 15 Sep p(fab) Is Not a Forecast. It's an Operating Model. · 22 Sep Who's left to buy anything? · 29 Sep The Best Way to Predict the Future Is to Invent It · 6 Oct Your AI Strategy Is Facing the Wrong Way · 14 Oct Intentional AI isn't enough (the one inside this window; summarised from a fetch, not read end to end). Its argument: DBS's "intentional AI" (purpose on the tool) "stops one level too low"; three levels, intentional tools → intentional workflows → the intentional organisation; "This is the Organizational Singularity: the point where an organisation can rewrite itself as fast as the technology changes." Its prescription is in the playbook (step 6). It ends by selling the REWRITE Readiness Score v2 (openexo.com/rewrite-v2: 24 questions in 8 dimensions, ~5 minutes, results sent after a form that asks for a business email, company and revenue range; no price stated; it labelled our first-dimension test "You are dabbling"). The other six posts are not read.
  • ⚠️ The Singapore keynote as reported by The Standard (Hong Kong), 12 Oct 2026, no byline, not his text, do not repeat as his words: "All mathematics will be solved soon and cancers are curable, thanks to AI"; AI capability "doubling… roughly every 10 weeks"; recursive self-improvement "about 85 percent"; "14 definitions of AGI"; "95 percent of AI projects are failing"; Fortune 100 ExO companies delivered "40 times the shareholder returns over seven years"; "Whoever figures out the future of institutions is going to run the world for the next 100 years." 🔴 None of these figures has a source in the article; the 40× is OpenExO's own marketing figure (see the plans-page warning below); the cancer line repeats the barred EP294 claim. Verified only that he spoke there (DBS release + his post). No recording found.
  • Last week's unverified item is now verified, and it predates this window. The release "Miles Everson Named CEO, as OpenExO Community Launches SINGULARITY ExO" was issued by OpenExO, Inc., New York, 22 Sep 2026 (EIN Presswire, read 2026-10-17; CityBiz still 403). Everson: CEO, author of Free Birds Revolution, former PwC vice chairman, former MBO Partners CEO. Salim: "Chairman of the Board". openexo.com/singularity-exo now exists: "a subscription learning platform that teaches both shifts side by side", status "Now forming", a "Share your interest" form, no price, no launch date. The release does not state who owns the entity. Added to Positions and disclosures.
  • Shaping Luck has a price page (openexo.com/shaping-luck, read 2026-10-17): "Founding members get Shaping Luck before anyone else", $380 "Lifetime access", $400 "Lifetime + Book", a "Claim Founding Membership" button; authors Salim Ismail and John Hagel; no publisher, no release date, no pre-order. We cannot tell whether the prices are new this week (the page was not price-checked before), so this is recorded, not reported as a change.
  • The Organizational Singularity pilot page is now titled "The Pilot Cohort · SOLD OUT" (read 2026-10-17; no cohort size, dates or price on the page). The homepage title now reads "ExO Leap founding cohort". salimismail.com is unchanged: book still "Spring 2026" beside "is here", Inquire link to #, no dated events, no Shaping Luck; Fund I still "Now Investing", "opens to outside LPs in 2026", first investment The 99 Kids, Q1 2026 (company-stated).
  • Resource Hub has dated new items (read 2026-10-17): Running Loops Claude Skill, Getting Loopy: AI Feedback Loops & Model Orchestration and Session Transcript: Building AI Loops, all 14 Oct 2026, all Pro-gated; Leap: Process Metro Map and A Path to Data Literacy and AI, 9 Oct 2026, Leap-gated. New free items noted in Free materials below.
  • OpenExO prices unchanged (exopass.openexo.com/plans, 2026-10-17): Beginner free "for a lifetime"; Pro $8,640 / $720 per month with "Save 17%"; 10x Shift "Your Investment: $100"; ExO Leap unpriced. 10x Shift next session unchanged: Tuesday 2026-10-20, 8:00 AM PDT, recording included (no price on that page). Member counts still disagree, none audited: "40,000+", "42,000+", "40K+" on one page vs his "now 50,000" (EP289 1:18:14).
  • ⚠️ Open, not resolved this week: the Genius Group tie. EDGAR full-text (run 2026-10-17) shows Genius Group Ltd (NYSE American: GNS) 6-Ks of 18 Mar 2024 announcing a definitive all-share merger agreement with OpenExO, and zero filings mentioning "OpenExO" after Aug 2024 and zero mentioning "Salim Ismail" since 1 Jan 2025. Wikipedia says "OpenExO merged with Genius Group in 2024" and "He is the Chairman of Genius Group"; the March 2024 exhibit names Roger Hamilton as CEO and Chairman. Whether the merger closed, and whether he holds any GNS seat today, is not established. Do not say either.
  • On tape: the feed is quiet since the 15 Oct episode (watch 2026-10-17: "nothing aired"); that episode is read and its Singapore check-in is now in Where he is lately. The 5 Oct Astro Teller fireside on disk was not read for him (read 2026-10-23, see Moonshots #unnumbered).

What he is here for: how Peter introduces him on tape: "the emperor of exponentials" (EP180 0:01:06, 2025-07-02), "the emperor of exponentials, Mr. Exo" (EP228 0:05:22, 2026-02-09), "the father of organizational singularities" (Moonshots #266 1:16), "our globe troder [globetrotter], father of the organizational singularity and warlord against linear thinking" (Moonshots, 5 September 2026 — Fable 5.1, GPT-6 Astra, the Cybercab takeover 0:01:41). The show's descriptions: "Salim Ismail is the founder of Open ExO, a GP at Exponential Venture Capital/The Organizational Singularity Fund and a sought after global speaker and thought leader" (EP258–EP295). The 2025 descriptions add that he "is the Founding Executive Director of Singularity University and the founder and chairman of ExO Works and OpenExO" (EP162–EP172), which his own site confirms: he "co-founded and launched Singularity University at NASA Ames Research Center with Peter Diamandis and Ray Kurzweil" (salimismail.com). On the panel he is the organisations chair: how an institution with people in it changes without breaking. (We did not find "father of Exponential Organizations" verbatim on tape.)

His books, and the free materials

book year · publisher link
Exponential Organizations: Why New Organizations Are Ten Times Better, Faster, and Cheaper Than Yours (and What to Do About It), with Michael S. Malone and Yuri van Geest October 2014 · Diversion Books (ISBN 9781626814233, Open Library) Google Books · Amazon
Exponential Organizations 2.0: The New Playbook for 10x Growth and Impact, with Peter H. Diamandis and Michael S. Malone 2023 · Ethos Collective (ISBN 9781636801780, Open Library) openexo.com/book · Amazon
The Organizational Singularity ("ExO Book 3.0"), released as a book and a Claude skill 2026 · OpenExO openexo.com
Shaping Luck, with John Hagel forthcoming · publisher and date not stated · sold as a founding membership: $380 "Lifetime access", $400 "Lifetime + Book" (page read 2026-10-17) openexo.com/shaping-luck

On tape about the third: "We have released the book for free and we've released a claude skill for free. So, anybody can download the claude skill and run their entire business uh on this model. Uh, go to openexo.com." (EP259 0:01:30, 2026-05-30). The reason he gives: "a book is a static thing. The minute I finish publish the book, it'll be out of date… So, we're going to be launching a Claude skill" (EP258 0:08:45).

Free materials, and where to find them (checked 2026-09-29):

  • The ExO 2.0 digital book. Free with the "Beginner… Free for a lifetime" account (openexo.com/plans).
  • An eight-page chapter summary (PDF). openexo.com/files/ExO_Summary.pdf.
  • The ExO Canvas, a one-page worksheet for planning the change. "free to download" from the Resource Hub (ExO Canvas · Resource Hub).
  • The framework in one line. ExO 2.0 is MTP + SCALE + IDEAS (OpenExO's own shorthand):
    • MTP, the Massive Transformative Purpose: the organisation's highest aspiration, the reason people join.
    • SCALE, the five attributes that reach outside the walls: Staff on demand, Community and crowd, Algorithms, Leveraged and shared assets, Engagement.
    • IDEAS, the five that run inside: Interfaces, Dashboards, Experimentation, Autonomy, Social technologies.
    • The chapter list is in the summary PDF above. ExO 3.0 is labelled "MTP + DRIVE + SHAPE for the AI-native org" on the OpenExO site. We did not find the expansion of DRIVE and SHAPE published there. The SINGULARITY ExO page (read 2026-10-17) describes the rebuild as six stages: "Purpose, Sense, Interpret, Decide, Orchestrate, Learn".
  • Resource Hub, free-of-lock items (read 2026-10-17; "no lock label" is our reading of the page, not a tested download): the Organizational Singularity interactive book app; Claude Skill How to Guide; MTP as a Protocol (PDF); The New Boundary of the Firm (paper); The Organizational Singularity Presentation from 24 August 2026; Overcoming Your Organizational Immune System (tool); the 5 Step Edge Builder, the one item the page itself marks "Free to use, yours to keep"; the ten attribute toolkits (MTP, Staff on Demand … Social); the ExO Canvas in several languages. The ExO 3.0 Claude Skill and Codex Skill are behind a login (the free Beginner account, as far as the page says). The 14 Oct loop items are Pro-only.

OpenExO: what it is, what it costs, how to join

What it is. The community and platform behind the method. It runs the ExO Sprint, the certification journey, the Learning Hub and the monthly 10x Shift workshop (openexo.com). Prices published on openexo.com/plans, 2026-09-29:

tier price what it includes (their words)
Beginner free, "for a lifetime" "Exponential Organizations 2.0 Digital Book · ExO Chatbot · ExO Insights · Basic Profile"
The 10x Shift workshop $100, one time (plans page, 2026-09-29) a "2-hour intensive session to draft your MTP". Next session: "LIVE on Thursday, December 3, 2026", 8:00 AM PST (read 2026-11-16; the 20 Oct session ran) (10x Shift)
Pro $8,640 a year, shown as $720 a month weekly implementation masterminds, a monthly keynote and Q&A with Salim, a group "ExO Ready Sprint", the ExO Canvas and a 6-month roadmap, training programs, certification
ExO Leap not published "the hands-on rewrite program with coaching, peer-CEO pods, and a prepaid first quarter"

⚠️ The plans page also carries performance figures ("40X Higher Shareholder Returns" and others). They are OpenExO's own marketing and are not repeated here as facts.

How a person joins: create the free Beginner account on openexo.com/plans.

Where he is lately

From the show's opening check-in ("where on the planet are you?"). Work locations only, newest first; updated weekly.

episode (aired) where what he said
EP311 (aired live 2026-11-13) no check-in "I was in Spain last week" 1:10:19; "I did the last pod from the… San Francisco Giants Press media press [box]" 1:26:32
EP310 (aired 2026-11-11, recorded 11-10) San Francisco, Oracle Park "I am in the press box of the San Francisco Giants at Oracle Park… I'm here for an AAI talk… I'm gonna be standing on top of the dugout uh to do my my keynote" 0:22:55–0:23:16 (event unnamed, no listing found 2026-11-16)
EP309 (published 2026-11-07, recorded ~11-06) back from Madrid (hosting) "I was in madrid for the singularity summit there" 0:04:53
EP308 (aired 2026-11-04, recorded 11-03) Madrid "I'm in Madrid for a couple of days at a summit. Singularities." 0:14:21 (⚠️ round-trip captions). The event: Singularity Summit Spain 2026, 4–5 Nov, Espacio Acciona; his talk "The Organizational Singularity" on the Day 2 agenda (spain.su.org, read 2026-11-10)
EP307 (published 2026-10-31, recorded ~10-30) at/near MIT (Cambridge, MA) at the close: "I I have to run back. I'm on stage at MIT in a few minutes. I have to run out of the door here" 2:15:40 (event unnamed). Peter's greeting "everybody's home today" 0:09:36 conflicts; his own line wins
EP306 (published 2026-10-28, recorded 10-27) Toronto "I'm in Toronto because I'm doing a talk here to about 400 CEOs tomorrow and uh and then flying back Thursday" 0:11:03 (event not named)
EP304 (2026-10-21) home Peter: "Salim, you're home." Him: "I am actually home" 0:01:39
EP303 (published 2026-10-17, recorded 10-16) in the studio, in person, after XPRIZE Visioneering (Malibu) Peter: "you've just flown in from Singapore" 0:02:30; at the close, "thank you for joining us at Visionering" / "Great to be here in person" 1:43:59
2026-10-15 episode (no feed number) Singapore Peter: "our globetrotter live from Singapore at 2 AM, Salim Ismail" 0:01:12 (whisper pass); him: "I'm tired, hungry, and cranky. Do not mess with me today." Three days after his DBS keynote there (12 Oct; DBS release + his 14 Oct post)
EP294, Moonshots LIVE (2026-09-27, recorded 09-25) Los Angeles, on stage introduced with the mates at 0:03:36
EP293 (2026-09-22) Bangalore "I was on stage at the beginning of this. I'm in a car in Bangalore." 0:38:25
EP291 (2026-09-17) Miami, then India "I'm getting on a flight to India tonight… we're launching the organizational singularity uh out there… I'll be doing the launch event and a couple of other talks." 0:04:58
EP289 (2026-09-13) home "I'm home. I'm home." 0:01:59
EP286 (2026-09-05) Houston "I was with the uh seauite [C-suite] of one of the bigger oil companies in the world in Houston" 0:02:11
EP283 (2026-08-27) São Paulo "I did a talk today for Ed McKenzie's [McKinsey's] forum to a couple hundred of their CEOs" 1:37
EP270 (2026-07-13) Barcelona "I'm in Barcelona at the Festival of Consciousness." 0:01:38
EP269 (2026-07-08) Mallorca a retreat hosted by the Festival of Consciousness 0:01:11
EP264 (2026-06-11) Las Vegas "I just spoke at a Zcleller [?] crypto cyber event" 0:01:21 (event name caption-mangled)
EP263 (2026-06-08) Dallas–Fort Worth airport "I came from the Mckenzie's AI summit in Montter [Monterrey], Mexico yesterday" 0:00:59
EP256 (2026-05-21) Brazil "a bunch of meetings and presentations here" 0:04:30

His site lists no public calendar (salimismail.com, checked 2026-10-17). We could not read his X feed from here. Off the show, dated at source: DBS Asia Future Forward Conference, Singapore, 12 Oct 2026 (DBS newsroom release 30 Sep 2026; his Substack 14 Oct) · Singularity Summit Spain 2026, Madrid, 4–5 Nov 2026, talk "The Organizational Singularity" (spain.su.org agenda, read 2026-11-10).

Consciousness

What he says, in his words. We don't characterise it beyond that.

  • The two theories, and which way he leans. "there's two theories of consciousness we've talked about this it's either an emergent property based on some sort of complexity in which case AI could get there the top down theory says that we're like localized antenna for consciousness and we connect into it. And as as we're getting more and more evidence, it seems the latter is more likely to happen… I think of it as a century problem more than anything else." (EP289 0:18:39–0:19:07)
  • What we don't have. "we don't know what consciousness is right we don't have a definition we don't have a test… if you talk to philosophers a subset of consciousness is considered to be self-awareness" (EP270 1:25:10–1:25:21).
  • The Festival of Consciousness. "We helped curate this and help put this together in the original years. Several thousand people show up at the Barcelona Convention Center for kind of an experiential uh uh understanding of consciousness." (EP269 0:01:16–0:01:32)

Why he is in this archive

Gate: present in all five episodes read. "Emperor of exponentials", "father of organizational singularities", "our Where's Waldo" (he dials in from a different country each week).

Positions and disclosures

Founder OpenExO (founder and chairman); founding executive director of Singularity University (salimismail.com); GP at Exponential Venture Capital / The Organizational Singularity Fund (episode description)
Chairman of the Board SINGULARITY ExO, the subscription learning platform "launched by the OpenExO community"; CEO Miles Everson (OpenExO, Inc. release, 22 Sep 2026, via EIN Presswire; openexo.com/singularity-exo "Now forming", no price). Ownership of the entity not stated. Not disclosed on tape in the episodes read. Context on the CEO: Fermi Inc. (FRMI) 8-K of 13 Jul 2026 and 8-K/A of 21 Jul 2026 on his resignation from Fermi's board, effective 10 Jul 2026 (⚠️ reason read from a TradingView summary, not the filing)
⚠️ Genius Group (NYSE American: GNS) Definitive all-share merger agreement with OpenExO announced in GNS 6-Ks of 18 Mar 2024; he was listed among new board members in GNS's 2023-results release. No filing mentions OpenExO after Aug 2024 or him after 2024 (EDGAR full-text, 2026-10-17). Whether the merger closed and whether he sits on that board today: unverified. Wikipedia's "Chairman of Genius Group" conflicts with the filing naming Roger Hamilton as Chairman. Do not state either way
Books Exponential Organizations (2014), Exponential Organizations 2.0 (2023, with Peter Diamandis), The Organizational Singularity (2026, free with a Claude skill), Shaping Luck (forthcoming, with John Hagel). See His books, and the free materials. On writing them: "three years of hell" for the first, "six months of joy" for the third, with AI as developmental editor
Past VP at Yahoo (Brickhouse); "Google bought my company"
Practice speaks to CEO forums ("McKinsey's forum, a couple hundred of their CEOs", São Paulo); "you deal with the Deloittes and the McKinseys"
Building a business-connection matcher; a BCI pilot programme; shapingluck.com with John Hagel — and on EP289 0:36:43 he claims they have gone further: "we figured out a way of both measure defining luck and measuring it… which is a big claim to make" (his own caveat)
Source he brings Principal Financial Group's SME survey — "friend of the pod, a bunch of them watch us" (EP284)
Also co-invests in the EV-toll real-estate idea with Peter and Dave; "no lab investments"

Where his information comes from

Practice, not press. CEO rooms, his own consulting clients, the ExO community, and the occasional survey a friendly institution hands him. He is the panelist most likely to say "I did a little bit of research and I've made a list."

Main ideas

  • Exponential technology, linear institutions. "The singularity is not when the machine becomes infinitely capable, it's when the institution can't adapt at all." — the organizational singularity.
  • Rules don't scale; principles scale.
  • Narrative beats evidence, and that is the problem. "Nuance has no viral coefficient." "Don't legislate with a story in your head." "Please be evidentiary and data-driven in your orientation."
  • Governance first. "The governance problem comes first." "The constitution does not update when transaction costs go to zero."
  • Intelligence cannot be contained. "We're treating intelligence as a product that can be contained, but it's not."
  • Smaller beats bigger, trust beats control (from the 2014 book).
  • Discount the consultancies and the futurists. "Take any of their estimates and multiply it by 10 or 20." "Great at framing the past, questionable on the future, because they don't understand exponentials."
  • Voice is the interface people will trust first. "The winners won't be the firms with the most associates. It'll be the firms with the best intelligence stack."

★ EP289 (the AMA, 2026-09-13) — four deltas

  • "Deliberate slack" — a new, unit-sized prescription. "we tend to overchedu [over-schedule] ourselves. Uh you want to make sure you create deliberate slack in the system. give yourself 10 or 20% of your time to just not have anything scheduled to allow uh experimentation, playing with new tools, meeting new people… If you wait till you have uh spare time, you'll never catch up." (0:05:41) The only prescription on the episode that costs nothing and needs no capital.
  • Education flips from supply-side to demand-side. "forget getting an education on the supply side where you become a doctor, a lawyer, an engineer, and then you try and go and find demand for that supply side skill… go to the demand side, tell the kids to get excited about a problem and then go find the skills and capabilities and techniques" (0:45:54), on the premise "a child with an AI today in an hour can learn what they could learn sitting in a classroom for the whole day. That impedance mismatch will break the existing system." (0:45:44)
  • The transition, sized for the first time. "about 50 institutions by which we run the world… We can't find one that survives a tsunami of AI plus 20 other Gutenberg moments hitting us at the same time" (1:17:20); "that period of uncertainty I think started around 5 or 10 years ago and it's going to go on for another 10 to 20 years… the question is can we minimize that trough" (1:18:26); the ExO community "now 50,000 people kind of a peace core [corps] for the world" (1:18:14). ⚠️ The 50,000 is unaudited.
  • Consciousness — the top-down theory is now his lead, not a both-sides. "as as we're getting more and more evidence, it seems the latter is more likely to happen… I think of it as a century problem more than anything else." (0:18:43)
  • On religion (the closing question): "…we invent our gods and then we believe in them… either there is no god and we're just moving towards a wonderful future driven by technology or there's a simulation and in which case it kind of doesn't matter anyway." (1:24:25) The line the editors chose to end the cold open on.

★ EP288 (2026-09-11) — he attacks the show's own standing frame

1 · ★ "Our models are clearly wrong." Asked to take a segment home, he refuses the frame the show runs on: "I haven't had my morning coffee, so I'm a little bit cranky… I think we need to retire this whole idea of things are moving faster than we expect. At some point, if you repeatedly say that, then we need to change how we make predictions. Our models are clearly wrong. We're going much more on the vertical." (1:04:25) This is the panel's core rhetorical device, attacked from inside the panel. Peter's reply is a defence, not an update — "we are the scapegoat in this sense because we've been talking about this for 20 years" (1:04:36). Nobody changes a model. It is the sharpest self-criticism in the corpus and it is dropped inside ninety seconds.

2 · ★ "Plans with triggers" — the most usable operator instruction in the episode. "people need to reframe their thinking to say not 'oh my god this happened' but 'when this happens, what will it unlock' and start thinking in that frame. Make plans with triggers in it. When AI can perform this, what will we change at that point? What becomes enabled?" (1:05:06) Costs nothing, needs no capital, and is directly pointable at any training curriculum and at Model-based systems engineering, applied to a syllabus — a trigger is a requirement with a condition on it.

3 · New coinage attempt: HBI — "humanity's best intent". "I have a couple of kind of spiritual advisor types and they're always asking me… 'what's in humanity's best intent?' And it's a really interesting framing. It forces you to lift right out to a Dyson swarm looking at the earth." (0:34:59) He then proposes it as the alignment target and as a question to ask the AI directly.

4 · The alignment ladder — his best structural contribution to the p(doom) block. He is the only mate who decomposes the word: "the problem with alignment is it comes down to a whole bunch of levels. Are you talking about alignment with the individual user, the company operating it, the government, the majority of people in a particular area… Do we go with the UN charter for human rights as a model? Do you go with a community or culture? Do you go with humanity's long-term interests?" (0:34:27) That is the governance chair doing its job, and the debate moves on without using it.

5 · p(doom) stated: "about .1%" (0:34:17) — the lowest number on the panel, and he grounds it in the wrong place: "Evolution has had 4 billion years to wipe us out, and it has not done so" (0:33:57), immediately corrected by [attr? Alex] — "Tell that to the billions of species that no longer exist."

6 · The close that carries the only disclaimer in the episode. "the two things that have always been true about mankind are death and taxes. We're solving death pretty clearly and with UHI, UBI, Bitcoin, whatever, we'll solve taxes. So I'm pretty optimistic about the future… also, past performance is no indication of future results." (2:25:13) In 2h43m that joke is the only time the phrase is said by anyone other than Alex about his own holdings.

🔴 Barred (EP288): the unnamed company case study — "I won't say which one… the company's worth about $10 billion and the data subsidiary got valued at $32 billion. So your data may be worth four times as much as your actual company." (0:15:24) Unnamed by choice, therefore unfalsifiable. Peter builds a whole prescription on it ("for all the accountants out there…"). Do not repeat the multiple.

★★ EP291 (2026-09-17) — he named the deliverable and bet against it. It landed in eight days.

This is the first clean, dated, resolved prediction in the whole hit-rate log, and it is his. Asked, as AMA question 2 near the end of a nine-and-a-half-minute speed-run, what the ideal outcomes of the US–China summit would be (2:11:45):

*"I would love a practical agreement where that makes dangerous incidents less likely. So, we have a lot of shared interests, right? Loss of control, protecting infrastructure. Uh, could you create reliable communications channels? Could you do incident reporting? Uh, shared incident reporting. Those are the things I'd love to see happen because there's so much to work on to generate a collective approach towards the world of abundance than to fight about it. I I don't believe that's going to happen, but let's see."*

What happened. Six days after record, the US and China agreed to establish a bilateral dialogue on AI after eight hours of talks in New York, and the Treasury Secretary said the two governments are discussing ways to warn each other about AI incidents that pose national-security risks. Xi's state visit ran 23–25 September 2026. He specified the mechanism correctly and put the odds against it.

Why it matters more than the call itself. The panel had just spent eighty-nine minutes arguing about which governance mechanism could work, and all three of the other chairs concluded it had to be Musk's cross-lab test harness, because "China's not going to agree to have some American regulators snooping around their AI companies" (1:30:16). What was actually announced was neither the harness nor a pause — it was the low-trust incident-hotline version that one mate described accurately and dismissed in 35 seconds, inside a segment they were rushing. The load-bearing answer was in the AMA. → Moonshots #291

★ The EP288 self-criticism did not survive six days. On Moonshots #288 1:04:25 he attacked the show's own standing frame from inside: *"I think we need to retire this whole idea of things are moving faster than we expect… Our models are clearly wrong."* On EP291 he argues from that frame without revision: "progress in all of this will not be stopped. It won't be slowed down a jot" 0:47:41; "the good news in my view in all of this all of these shenanigans… is completely irrelevant and won't make a dent in the progress" 0:31:49. No revised model appeared and nobody raised it.

New position on regulation, stated numerically and harder than the host's. "it will make zero zero zero difference because… once you have open weights available, uh anybody can download these open models and run a local harness and you essentially the whole thing is irrelevant" 0:31:28 — thirty seconds after Peter says "let's be clear about one thing there is going to be AI regulation" 0:30:47. The two are never reconciled.

★ A good standard, applied to others and not to himself. He demands a mechanism from the labs: "How you get to the idea that AI can meaningfully destroy all of humanity is incredibly irresponsible unless you lay out exactly the steps by which that might happen" 0:22:49, restated as "Show me the specific mechanism" 0:47:46. Thirty minutes later he asserts a two-hundred-year literacy decline with no mechanism and no source — see the barred row below. Both belong on this page together.

P(fab) is now furniture, and he sources it. "we kind of threw out on the last pod this idea of P fab… A whole argument erupted on Twitter on what should we call this? People said call it P abundance. Imad said call it P moon… my personal favorite is P fab just because it's the simplest to say." 0:38:22 The operator version of it is the usable one: "If you're going to publish a plan that AI could wipe out humanity, also publish how it could totally save everything and solve every goddamn problem… If you don't have that, then you're not balanced. And I'm sorry I'm not taking you credibly." 0:39:08 He also asks permission to perform the optimism landing — "Can I lift us up a bit please?" 0:37:28 — which makes it a named move rather than an emergent close.

New operator prescription — the shadow C-suite, and it is concrete. "what would happen if I took major chunks of my operation and allowed an AI to run it… If I could put in a shadow CFO to make financial decisions and a shadow CMO, an AIdriven CMO, you should be able to coordinate a whole set of roles now" 1:42:18. And the correct caveat on the benchmark it sits on: "This is a simulated year of vending operations, right? Not an actual independently verifiable real company" 1:41:51.

New long-range education position — invert the supply side. "the halflife of a a skill used to be 30 years and now it's like 3 years. We don't know what the demand is… all the way from primary to secondary to university. We have to redo the whole stack to go through problem solving side. Go to the demand side" 1:57:20; plus the institutional-immune-system ranking ("academia is the second worst. Religion is the worst. Healthcare is the third worst. Corporate immune systems the fourth", 1:57:54) and the verdict: "you have to do a new model at the edge."

🔴 Barred (EP291) — the episode's worst number, and he says it twice. "the highest uh adult literacy in the world was New England at the time of the formation of the US. 95% of the voting public was fully literate. And it's never been that high since. It's gone down since then" 1:08:35, repeated as "the highest literacy was like 200 years ago we've been sliding down as a long trend ever since then" 2:06:35. The ~90–95% figures come from signature-mark counts on wills and deeds among adult white men — a bare-minimum proxy, not functional literacy, excluding women and the enslaved. Modern US adult literacy on any comparable basis is higher, not lower. It fails on definition, not on decimal places, and it is the load-bearing premise for a decline thesis Dave partly adopts. Never repeat. 🔴 Also barred: "more than half the CEOs in Silicon Valley have a liberal arts degree" 2:03:19 — widely repeated, traceable to no census, no source given.

★ EP293 (AMA #2, recorded 2026-09-18, aired 2026-09-22) — P(fab) is his coinage, and he is the only one who keeps the condition attached

He joins 37 minutes late, from a car in Bangalore ("I was on stage at the beginning of this", 0:38:25), and speaks in four blocks. Three deltas and one new front.

★ Board liability for AI agents — a new front, and a forthcoming artefact

0:53:38, volunteered onto the end of a CISO's question:

*"We've been looking at the liability for boards because AI agents are doing fairly illegal things in a lot of the companies. And so there's a massive kind of overhang of liability that sits. So I'm actually writing a paper with a guy who's been on 30 different public boards on how do you navigate this as a board in the future. So watch for that."*

Track the paper. 🔴 Never repeat the premise — "fairly illegal things" in "a lot of the companies" is offered with no instance, no sector, no jurisdiction and no source. It is also the only claim in the episode that would create legal exposure if repeated. The co-author is unnamed.

★ P(fab) is his, and his number is the only one with its condition stated inside it

The coinage reaches the panel via a caller: *"Salim said something recently, which was about PFAB. And I thought, wow, what a great idea. We're always talking about PDoom. What about PFAB?"* (1:04:19). His answer (1:05:17–1:06:14) puts the work before the number:

*"What you need to get there is we need to rebuild all of our institutions globally. Education, for example, totally broken, needs to be rebuilt. Monetary systems broken, need to be rebuilt. Governance models, dispute resolution systems, legal systems, healthcare systems — there are 50 major institutions by which we run the world. They pretty much all need this. … If we were able to do it, my PFAB is close to 100 percent."*

The if is the whole answer. He gives no probability that the condition is met, and the condition is a project nobody has started. Compare Alexander Wissner-Gross (time-indexed, >90% for T>10yr) and Dave Blundin (99.9% conditional on surviving a 2–10-year arms-race window). 🔴 Stated unconditionally — "Salim is ~100% optimistic" — this becomes the opposite of what he said. The episode's own cold open performs exactly that compression on Dave's number. See Moonshots with Peter Diamandis and the research archive §7.

The E. O. Wilson frame, with his own addendum

1:06:00 — "The problem with humanity is our emotions are paleolithic, our institutions are medieval and our technology is godlike." Then the part that is his:

*"You can solve the paleolithic emotions with psychedelics, but the institutions being medieval really needs a whole other level of… group organization. We're really good at individual transformation. We're very bad at group transformation and institutional transformation."*

That last sentence is the cleanest statement of his standing thesis in the corpus and belongs in the deck as-is.

Universities ranked against religion — the write-off is now institutional

0:38:25–0:39:43. His EP289 move was curricular (supply-side → demand-side). Here it becomes a judgement about the institution's capacity to change at all:

"The traditional model of a university breaks. It cannot be a supply side, skills building, job schooling environment… Peter nails it when you say this is where you go to figure out your MTP… But that's such a massive transformation… and you take the immune system in universities being the second worst immune system I've ever seen next to religion — I don't have a strong view on their ability to actually transform. So we'll have to create new universities that do that, and then let that become the new gravity center over time."

The intelligence stack, stated as an architecture

0:44:31, answering a 25-year workforce-systems veteran on continuous capability matching:

*"You want the architecture of the organization to be so that you can accelerate those learning loops… When I talk to CEOs, I basically say please, whatever you do, invest in the adaptability and flexibility of your organization and just double down on that. Because that new architecture — and we're doing this in our organizational singularity work — we need to rebuild our workflows that are AI centric to be over on top of an intelligence stack, which is a learning loop, and then you layer workflows on top of that. And the whole thing becomes a self-learning proposition."*

The OpenExO / organizational-singularity product is the answer, undisclosed as such — the standard pattern on this page, and the reason the disclosure regex scores this episode 0.

★ EP294 (Moonshots Live, recorded 2026-09-25, aired 2026-09-27) — he asked the falsifier question himself, and answered it for the other side only

On stage in Los Angeles, introduced as "Our Afghan warlord, Salim Ismail" 0:03:44 (the EP284 airport anecdote, 0:46:40 there, turned into a credit). Full decomposition: the research archive.

★ The question Moonshots #293 left unanswered comes back, from him. "When you look at this whole narrative, it's interesting to me to ask the following question. What evidence on either side would have somebody change their minds, right? If you think ASI is bad, then you're like your narrative stuck. There's nothing we can do to show you that it's not bad. And same with the other side, it's all, it's very tough to change narratives in our world." 0:30:05–0:30:24. He names the symmetry and supplies no falsifier of his own. Then: "the she, Trump summit, the Bernie stuff, it's completely irrelevant. None of it matters… the minute you have local weights available and people can run their own models of a local level, the world is going to move ahead. It doesn't matter what any single person does now." 0:30:24–0:30:42 — the EP291 "zero zero zero difference" line, and the resignation-as-optimism move Peter Diamandis made on EP293 1:06:21.

🔴 Barred (EP294): "we will likely solve cancer in the next two years" 0:30:51. No mechanism, no source. He labels it "an unbelievably optimistic thing to aim for" in the next sentence; the clip will not carry the label. By his own EP291 rule ("Show me the specific mechanism" 0:47:46) it fails.

★ Nation states → city states, with Alexander Wissner-Gross taking the other side. "I've said this repeatedly, nation states are a terrible way of governing the world today. They're geared towards scarcity… AGI, ASI, whatever the hell you want to call it… emerges. And it forces us to operate as a world community." 0:16:37–0:17:03. Alex: not erosion but "a proliferation of new forms of governance" (0:17:27–0:18:02). Salim, in one breath: "I totally agree with you. We are moving this century from the concept of a nation state to a city state… Take any city, take LA. If you have solar energy, satellite internet, vertical farming, whatever… you don't need a country." 0:18:02–0:18:30; "the tension you see in the world today is not left versus right, it's urban versus rural. Brexit was London versus the rest of the country. Trump is urban versus rural" 0:18:30–0:18:40; "the problem of nation states is they have armies and they like to use them" 0:18:42. "I've said this repeatedly" checks: ep253 1:24:11, ep279 1:24:29, 2026-09-05 1:12:06.

The name. "we talk about AGI. First of all, it's not artificial. It's not general. It's not really intelligence. Other than that, I'm good with it… We need somebody to come up with a good name for all of this that's completely different. I've said repeatedly, I think what we're building here is an complimentary mechanism to human intelligence, not replicative." 0:23:06–0:23:32. ⚠️ "the first article… that said robots or AI will take all the jobs in five years appeared in 1964" 0:23:42 is unverified.

P(fab) "like a thousand times" P(doom), with the EP288 evolutionary argument (0:09:24– 0:09:48), and the same correction landing again — "You've got to remember that 99.9% of all species are no longer with us" 0:09:54 [attr? Alex] — waved through: "Yeah, that's fine. But we keep moving on." Second time.

Model fatigue. "I've got total model fatigue. I mean, I can't keep up with 5.5 versus 5 versus Astra… raise your hand if you have model fatigue… I love the progress, but Jesus, I can't keep up." 0:41:47–0:42:00. Against EP288 ("our models are clearly wrong… we need to change how we make predictions"): the complaint has moved from the panel's forecasting method to the pace, and no method has changed.

Disclosure. The Future Vision XPRIZE praised to its founder, on his own stage: "the XPRIZE for creating a positive vision of the future is one of the most important things that's happened in maybe a hundred years" 0:24:22 → Moonshots: the disclosures ledger. Closing verdict: "Biology is cooked. It's simple. It's true." 0:50:52.

Prescriptions → the playbook. None new to the listener. The sequence does not change.

The characteristic moves

  1. The two-line aphorism that closes a segment — "intelligence is becoming cheap, but the manufacturing of intelligence is becoming incredibly expensive."
  2. "Can I give the counterpoint, please?" / "Can I change the question?"
  3. The enumeration — "three things here. First of all…"
  4. The ExO pattern-match — "that's a very classic ExO play."
  5. The long parable (Jerry Yang's 300 VPs).
  6. Calls for the slide.

His model, step by step: the implementation playbook

The ExO method as a sequence a process person or a small organisation can run. It merges his books' structure (MTP, then SCALE and IDEAS) with the prescriptions verified on this page. Each step has its source. The steps that need an organisation or money are marked.

  1. Make room first: schedule deliberate slack. "give yourself 10 or 20% of your time to just not have anything scheduled… If you wait till you have uh spare time, you'll never catch up." (EP289 0:05:41)
  2. Start from a problem on the demand side, and write it as your MTP. "go to the demand side, tell the kids to get excited about a problem and then go find the skills" (EP289 0:45:54). The book's first attribute is the Massive Transformative Purpose, the organisation's highest aspiration (Exponential Organizations, ch. 4; summary PDF). OpenExO's $100 workshop is built around "Draft Your MTP" (plans).
  3. Score yourself on the ten attributes with the free ExO Canvas. Outside the walls, SCALE: staff on demand, community and crowd, algorithms, leveraged assets, engagement. Inside, IDEAS: interfaces, dashboards, experimentation, autonomy, social technologies (chapters 5–14; the ExO Canvas is free). Pick the one or two that are cheapest to switch on first.
  4. Write plans with triggers. "Make plans with triggers in it. When AI can perform this, what will we change at that point? What becomes enabled?" (EP288 1:05:06)
  5. Put an AI in one real role and watch it. "If I could put in a shadow CFO to make financial decisions and a shadow CMO… you should be able to coordinate a whole set of roles now" (EP291 1:42:18), with his caveat on the evidence: "This is a simulated year of vending operations… Not an actual independently verifiable real company" (1:41:51). ⚠️ Assumes an operation with functions to shadow.
  6. Build the learning loop before the workflows. "invest in the adaptability and flexibility of your organization… rebuild our workflows that are AI centric to be over on top of an intelligence stack, which is a learning loop" (EP293 0:44:31). The written test for this step (his Substack, 14 Oct 2026, Intentional AI isn't enough): "Take your best AI pilot and ask one question: if we rebuilt the whole workflow it sits inside" from purpose, what changes? Three levels, intentional tools → intentional workflows → the intentional organisation; "most organisations are stuck at Level 1." ⚠️ The post ends by selling OpenExO's REWRITE Readiness Score v2; the test is usable without it. How to move the workflows, by size (EP308 0:54:18–0:54:29): under ~50 people, brute-force it; over 50, "set this up stack of intellectual resources on the side and start moving workflows one by one", e.g. "One company took 22 steps and reduced them to four". ⚠️ The case is unnamed and its numbers are promised, not published; the over-50 route assumes an organisation big enough to run two stacks at once.
  7. Test every claim in the plan for its mechanism and its upside. "Show me the specific mechanism" (EP291 0:47:46) · "also publish how it could totally save everything" (EP291 0:39:08).
  8. Where the institution cannot change, build a new one at the edge. "you have to do a new model at the edge" (EP291 1:58:12) · "we'll have to create new universities… and then let that become the new gravity center over time" (EP293 0:39:31). ⚠️ Building a new institution assumes capital and a team.

How to check a consultant, by his rules

He works with the large consultancies and says what he discounts in them. As a checklist:

  1. Is the estimate built to be defensible, or to be right? "they're scared of being called out for anything vaguely approaching a real estimate… they want to make sure their numbers are defensible to the nth degree. So no speculation at all." (EP284 2:07:33)
  2. For anything exponential, test the forecast against the trend. "I typically say take any of their estimates and multiply it by 10 or 20" (EP284 2:08:49). ⚠️ That is his rule of thumb, not a measured correction factor. Use it to ask the question, not as a number.
  3. Do they understand the curve? Of the big commentators: "unbelievably great at framing the past somewhat questionable on the future… because many of the time they don't really understand exponential" (EP284 1:10:13).
  4. Ask for the mechanism. "Show me the specific mechanism" (EP291 0:47:46).
  5. Ask for the upside plan as well as the risk plan. (EP291 0:39:08)
  6. Will they push you faster down the old path? "if they bring in one of the traditional consulting firms, they just push them faster down the old path, right? And so that doesn't work at all… create a new stack on the edge that's completely built AI native from the ground up" (EP215 0:38:38–0:39:00).
  7. Do they redesign the institution or patch it? "we need to rebuild every institution and re-architect every institution by which we run the world. And that is the biggest advisory opportunity in the history of mankind." (EP234 1:00:41)

How to connect with him

Public doors only:

  • The OpenExO community. Free Beginner account at openexo.com/plans. The $100 10x Shift workshop is taught live by him. Next session: Thursday 2026-12-03, 8:00 AM PST (10x Shift, read 2026-11-04).
  • Speaking. The "Book Salim" and "Check Availability" forms on salimismail.com. "Salim's team will be in touch promptly."
  • Public channels. X @salimismail · LinkedIn · YouTube @SalimIsmail (all listed in the episode descriptions) · Substack, weekly since 1 Sep 2026, free (found 2026-10-17).
  • Events. Moonshots LIVE (Los Angeles, 2026-09-25), where he was on stage.

What an operator can do with this

Doors a person or a small company can walk through this month. Each was checked at the source on 2026-09-29.

  1. Join the OpenExO community. Free parts listed on the site: community profiles, "Webinars, networking calls & workshops", a Resource Hub of "Tools, templates & frameworks", and an ExO chat bot (openexo.com). The paid tier includes the ExO Canvas, the ExO Sprint, the certification journey, "The 10x Shift" monthly workshops and "The ExO 3.0 Assessment". Prices are in the OpenExO table above: a free Beginner account, a $100 workshop, Pro at $8,640 a year.
  2. Block 10–20% of next month as deliberate slack. Free, and it is his step 1 (EP289 at 0:05:41).
  3. Write one plan with triggers for your own operation. Pick a task AI cannot do yet and write down what you will change the day it can (EP288 at 1:05:06).
  4. Run a shadow for one function for a month. Let an AI draft the finance or marketing decisions next to the person who makes them, and compare. Keep his caveat that the evidence so far is a simulation (EP291 at 1:41:51).

What to take from him

Concrete, this week:

  1. Open the free OpenExO account and read the ExO 2.0 digital book, or the eight-page summary if time is short.
  2. Write one sentence of MTP for your organisation, and one problem it solves on the demand side.
  3. Fill in the free ExO Canvas. Mark the two SCALE or IDEAS attributes you could switch on in 30 days.
  4. Write one plan with a trigger: "when AI can do X, we will change Y."
  5. Next time a consultant hands you a forecast, run the seven checks above.

The one idea to keep. "We're really good at individual transformation. We're very bad at group transformation and institutional transformation." (EP293 1:06:08) His method is for the second problem. It is the one most AI advice skips. His rule on consultancies is a verify-first rule stated from the other side. The method maps onto Model-based systems engineering, applied to a syllabus: a trigger is a requirement with a condition on it.

★ EP298 (aired 2026-10-02)

Read in full 2026-10-06 (Moonshots #298).

  • "This is theater" on the accord (0:14:21), with two reasons: ceremony on "an exponential coordination problem", no independent testing or consequence for breach; "there's no mechanism for slowing any of this down. We need to speed up our response to it."
  • Abundance = access, not raw material (0:34:06): "when the essentials of a good life become affordable and accessible then you have abundance." The parenting-blogs example (0:35:02) is his; the "thousand times more effective" figure is his own estimate, not a measurement.
  • Anthropic S-1: steelmans the scale ("it's servable. You can refinance" 0:52:27), then: "we need your money to build a technology that we're warning you about" (0:52:37).
  • Pilot disclosure (new): "we're finding one thing in our pilot that's uh about now 80% through… these 10 companies as we rebuild them to be AI native… as we automate workflows with quote unquote agents, it's exposing a lot of ambiguity that humans cover for" (1:11:45–1:12:06). No interest stated. The finding (workflow ambiguity surfaces on automation) is checkable only when he reports back.
  • Operator door added to the model: "Give me the three decisions I need to make and with the judgment I can bring to the table plus give me the evidence to make those decisions" (1:12:39) — a design target for agentic automation: the human's surface is decisions plus evidence, not tasks. Check against his step list; it sharpens, not replaces, the "make the ambiguity explicit first" step. Assumes you already have workflows explicit enough to run; his own pilot says most are not.
  • Religion and morality (1:47:21–1:48:35): "religious morality is somewhat of an oxymoron", "based on absolute truths which are all totally false"; prefers utilitarianism, human rights, game theory. Emad dissents (1:49:07). Dave later: he hopes to offend "everybody equally" (2:08:49).
  • AMA: capitalism ↔ abundance: "keep open markets open… expand ownership… drop the cost of essentials" (2:10:37); on centralised governance: "common protections but distribute experimentation" (2:16:27).
  • ⚠️ "a free diagnostic doctor in China that's being used by 100 million" (0:46:47) — unverified, do not repeat.

Where it came up

How this page stays current

Refreshed weekly by the Moonshots lane.

  1. Where he is. Add each new episode's check-in location (work only) to Where he is lately, with the timestamp.
  2. News. Check salimismail.com and OpenExO for new books, launches and programs (Shaping Luck, the Organizational Singularity). Update the books table with publisher and date when they appear.
  3. ExO materials. Re-check openexo.com/plans and the Resource Hub for prices and new free tools. Record the date checked.
  4. Prescriptions. Add new ones to the playbook (5–8 steps), each with its timestamp.
  5. Frontmatter. Update updated:.

⚠️ Unverified

Fund role from the episode description. Auto-captioned quotes; his name is mangled in the captions and attribution is inferred from hand-offs. The Genius Group merger's closing and any current GNS seat (see Positions and disclosures). Every number in The Standard's 12 Oct 2026 account of his DBS keynote (see This week). Ownership of SINGULARITY ExO. The end date of the Shaping Luck "Final 48 hours" countdown (none stated; still at 48:00:00 on 2026-11-04, twelve days after first seen). The Toronto talk to "about 400 CEOs" (EP306, on tape only, event unnamed). The MIT stage he leaves EP307 for (on tape only, event unnamed). His ~29–30 Oct jobs video (platform unnamed on tape, not found at source). The "The Shift" / "10,000 subscribers" YouTube channel claim and the free MVIS Stack Workbook link in the 3 Nov Substack post (neither tested). The Kaufman Rossin / NewtonX and Lovable figures in that post (secondary links only). The Stanford "about 11%" and PwC figures in his 27 Oct Substack post (no links, not checked). The reason in Fermi's July 2026 8-K on Everson (read from a summary).

★ EP299 (aired 2026-10-03)

Read in full 2026-10-08 (Moonshots #299).

  • Holds the line that Brain-IT decodes perception, not thought (0:28:20).
  • Procurement: "can a procurement organization built for 20 or weapons programs operate on like a 90-day technology cycle" (2:03:44) — the question on the Meridian segment nobody answers.
  • Drone statistics (2:05:20, [attr?]) fail verify-first: Ukraine's ~10M is its own target; no source for Russia's.
  • AMA, 2030 graduate (2:11:06): "shift from doing tasks to focusing on owning outcomes… a totally new apprenticeship model."

★ 2026-10-07 episode (Moonshots #301, EP301 by sequence)

Read in full 2026-10-08.

  • On the Force: "a 20th century classic task force to address a distributed 21st century technology… a fundamental impedance mismatch" (0:24:43); wants "incident reporting, identity, liability" as scaffolding and "bounded experimentation… publish any failures" (0:25:25). Notes the administration's contradiction: "You have Clayton saying the danger is not moving fast enough, and Bess is telling the labs to slow down" (0:57:47) — his paraphrase of Clayton's reported "the risk of not being first is high."
  • Altman: "the goal should not be zero bad outcomes, that means you just ban the technology" (0:46:22); aviation rules "written in blood" (0:46:04).
  • Hotline: "completely expecting a nothing burger… The hotline is laughable" (0:57:57).
  • Personhood: "we're going to create AI personhood socially before we create it legally or scientifically" (1:17:04); "Just don't confuse simulation of personhood from proof of personhood" (1:32:27); on Anthropic's training [attr?] "the lab most concerned about safety is also the lab training the AI that it might be autonomous… a very big contradiction" (1:23:36). The teaser's first line is the front half of a sentence whose back half is that objection.
  • Frames: "organizational singularity" with "100,000 competent digital workers overnight" as the hypothetical (1:41:55) — not an operating claim. The AMA 2:31:41: iteration cycle down "to 10, 11 days".
  • T-shirt: "Pdoom less than zero" (0:23:06); at the close, Peter jokes about it.

★ 2026-10-15 episode (Moonshots #unnumbered, no feed number) — a portability prescription, and a crypto dissent

Read in full 2026-10-16; live from Singapore at 2 AM ("tired, hungry, and cranky").

  • The personal-ontology export (new): after each major set of queries he asks the model to "extract all that key information, figure out what the key concepts are, key relationships, key entities, and create a personal ontology, a personal memory that makes it portable between models" (0:38:37–0:39:17). Motive: Meta's active memory makes switching harder; "The next trillion dollar battle is not going to be about which AI is going to be smartest, [it's] which one knows you best." A single chat subscription is enough; his own caveat is that lock-in is "going to be the big battle over the next few months."
  • Dissent on crypto: "Crypto and Bitcoin and so on are not just mathematics founded… we solved the Byzantine generals problem" (0:49:~); Alexander Wissner-Gross disagrees.
  • "Every advanced PhD program in the world is toast right now" (0:52:24): one researcher directing hundreds of agents; an opinion about "every" program.
  • Self-description: "Canadian-level polite to my AIs" (0:11:33).

EP303 (Moonshots #303, published 2026-10-17, recorded 10-16) — the robot-timeline line he held

In person, the day after XPRIZE Visioneering, with guest Anousheh Ansari. Read from the decomposition (the research archive, 2026-10-21).

  • Robots start with dull, dangerous and dirty work, and he holds it under two push-backs from Peter: "I think it'll be a while before it comes into the home, liability issues" (0:22:52); "No, no, I said it would start off there"; the reason: "we have no supply chain right now for making billions of robots" (0:26:43). Neither moves.
  • The certification question on the flying-car segment, unanswered: "who regulates this? DOT, FAA, nobody" (0:12:50).
  • His pilot, unnamed, as evidence: "in the pilot that we're running people are taking 22step workflows and dropping it down to four steps" (0:31:27). Company-stated, no data; the same pilot as EP298.
  • Frames restated: the "20 Gutenberg moments" (1:01:21); MTP, "in 2012, we coined the phrase" (1:37:50); on work and meaning, "I'm not sure they get meaning, but they define their meaning that way" (1:25:14). Purpose-first ("if you don't have a purpose yet, go and find it" 1:42:00) is already step 2.
  • 🔴 Barred (EP303): "79% of corporate workers are dissatisfied with the work that they do" (1:25:25): Gallup's figure is about 79% not engaged, and engagement is not satisfaction. "362" pitches by Google's founders (1:41:17, [attr?]): no source. Swedenborg "in the 1500s a monk" (1:12:45): he lived 1688–1772 and was not a monk.

EP304 (Moonshots #304, aired 2026-10-21) — the crypto dissent goes quiet

  • Philanthropy is "a very scarcity era response"; "we're still taxing labor" (0:15:20, 0:15:54). The debt is "the biggest god damn issue facing the world today" (1:17:36). He flags the show's US-only lens (1:20:54).
  • Picked four idealab.ai ideas live (1:28:53–1:30:02) and proposed giving the list to an AI "and say go execute on all of these" (1:30:12). No change to the playbook: it is a demo, and it skips his own step 2 (start from a problem).
  • Silent on crypto when the guest says Bitcoin "goes to zero" (1:49:08), after his 10-15 dissent above.

The Astro Teller fireside (Moonshots #unnumbered, aired 2026-10-05)

  • Pitches his own organisation to the guest on air: "We are now 50,000 people worldwide, and we know how to create an environment where you can launch these things very quickly" (0:24:51) and "this could be a powerful collaboration" (0:25:13). The 50,000 is self-reported and unchecked (Moonshots: the disclosures ledger).
  • Two-class question: "pain points" against far-sighted projects (0:29:02); the guest wants both (0:29:22).
  • Caption mangles: "Stuart Brent's 'tempo levels' model" (Stewart Brand's pace layers, 0:25:13).

EP305 (Moonshots #305, aired 2026-10-23) — four controls for every agent, pace layers, and a Bitcoin dissent

Read in full 2026-10-24 (detail in the research archive of ep305-elon-s-100b/).

  • The control loop (no change to the playbook): after OpenAI's monitor incidents, "Every agent needs to have four things. It's almost like a a digital passport with metadata saying what you're allowed to do… trusted evaluation suites… full traceability and logging… roll back capability… and the fourth is a human review queue" (0:52:36–0:53:17), tied to "the organizational singularity book" (0:52:26). It is an organisation-level step; not verified at a source and not a listener step, so the playbook is unchanged.
  • Pace layers, against "bottleneck": "I think you're all three wrong… the wrong metaphor bottleneck… Stuart Bran's pace layers where… government's metabolism moves slower than culture and technology has a very fast metabolism" (0:34:02), applied to a company running pilots at 100× the speed of the rest (0:34:48).
  • Portable AI memory file (no change): "create an AI memory across all of the projects and all of the different models… so that there's a persistent identity… that becomes portable across models" (1:07:43–1:08:15); later, to the Bill Gross list, "if you have an AI memory, like I suggested earlier, get that to scour the list" (1:45:14).
  • Agent mandates: "what can it spend? What terms can it accept? What's the governance boundary around it?" (1:14:28); "the organization as a static hierarchy of human beings is now essentially irrelevant. It becomes a protocol of agents and human beings" (1:15:41).
  • Data-centre dividend: a Bogotá skyscraper crowdfunded by its community as the model (1:25:42), "it's better from the bottom up"; Virginia "16% of the GDP of the state is now data centers" (1:26:35, 🔴 unsourced). Meta Muse free is "a grab for all of your data" (1:19:16).
  • Bitcoin, [attr?]: a mate who is not Peter, Dave or Alex says of EP304's "goes to zero" "for the record, I am not [in line]" and "Bitcoin is going to do unbelievably well uh in the future" (0:07:15–0:07:26), argues blockchain "has solved the Byzantine general's problem" (0:09:59), and says "the Bitcoin community has been talking about this for a decade" (1:30:02, the hand-off "Seem" at 1:30:24 fixes this one). Consistent with his EP2026-10-15 push-back ("Crypto and Bitcoin and so on are not just mathematics founded") and with his silence on EP304.
  • Ties and plugs: "one of the companies applied AI and in six weeks is saving close to $50,000 a month… I'll report back on the progress in a structured way" (2:21:47–2:21:58), his own programme, client unnamed. A PR hire who quit from OpenExO (0:46:42) is told as the counter-story to building a public profile.
  • Digression: "if you go back and look at any religious text, it will say… do whatever you have to to get other people to join your religion" (0:53:27); Alex answers on the record that he disagrees (0:55:36).

EP306 (Moonshots #306, published 2026-10-28, recorded 10-27) — the MTP as governance, then the guardians question

Remote from Toronto (0:11:03). Read from the decomposition (the research archive, 2026-10-29), quotes checked against the transcript.

  • Anthropic's Long-Term Benefit Trust, read as his own frame: "They've essentially embedded the MTP into the governance" (0:37:40); "The MTP gets a seat at the table and the power to choose the table" (0:38:22). He asks the question himself, "who watches the guardians… who holds the trustees accountable?" (0:38:32), and after Alex's dissent concedes it: "you could end up with group think at the trustee level… trustees appoint each other. So you've got this circular mechanism there" (0:41:51–0:42:02). No change to the playbook: it is a corporate-governance reading, not a listener step.
  • Token price: "53 days for a 50% drop in the token price… we've dropped the doubling time for price performance to to 7 weeks pretty much" (1:07:28–1:07:41). ⚠️ No source found (decomposition §10); do not repeat.
  • Ethics voice on the cloned-dog segment: "basically built for sacrifice" (1:37:05); "the consent paradox" (1:43:59).
  • Bitcoin, now attributed by name (Peter assigns him the AMA question, 2:08:39): "I would challenge the premise. You need to distinguish between cryptocurrency and… cryptography" (2:08:49); "with the lightning network, Bitcoin has become truly scalable… I expect Satoshi's to be the… currency of choice for AI agents" (2:09:45–2:09:55). ⚠️ Opinion, contested; not a fact to repeat. This settles the EP305 [attr?] Bitcoin voice as consistent with him.

EP307 (Moonshots #307, published 2026-10-31, recorded ~10-30) — the Bitcoin dissent becomes an accusation, and he quadruples down on religion

Remote; leaves at the close for an MIT stage (2:15:40, event unnamed). Read from the decomposition (the research archive, 2026-11-02), quotes checked against the transcript.

  • Chamath's Bitcoin call, from dissent to motive: "I'll be blunt. I think it's a scop [psyop]. I think he's working on behalf of the Treasury Secretary… he used to be a massive Bitcoin [bull]… He was playing his book." (0:01:05, Peter's hand-off by name). ⚠️ No evidence offered; an on-air accusation, not a fact. The arc: EP304 silence → EP305 [attr?] dissent → EP306 named dissent → EP307 motive.
  • His own video as the segment seed: "Can I can I just segue to this video I put out, Peter?… I put out a video yesterday basically asking a simple question… 'Wouldn't it be the greatest thing ever if all the jobs disappeared?'… it's getting like crazy comments" (0:35:57–0:36:27) — own-product seeding, logged in the decomposition's plug table. ⚠️ Platform unnamed, video not found at source.
  • Education, restated harder: "Go build a business. Do not take an MBA." (1:45:50). No change to the playbook — it restates step 8 (build at the edge).
  • Religion: he doubles down on the EP305 line, on his way out the door: "I want to make a comment about my religion comment from last episode… Double down." (2:15:19–2:15:40), to the panel's "Quadrupling down" and his own "Two-thirds of the world consider themselves to be religious. So I've just offended that." (2:16:45). ❓ The historical claims in it (forced conversion by "Tibetan Buddhists… bubon [Bön] practitioners", "the Jews in the Hasmonian [Hasmonean] kingdom") are not checked. Off-lane; don't use.
  • 🔴 Barred (EP307): Fort Lauderdale "blocked Uber… about 3 four years ago" (1:35:00) — it was 2015, county-wide (Broward), ~11 years ago; the airport-ticket-sales story is his and unchecked. "Half the court cases in the US are car accident related" (1:40:26) — auto torts are ~60% of tort filings and ~29% of general civil caseload (Court Statistics Project, old data), not half of court cases.

EP308 (Moonshots #308, aired 2026-11-04) — the workflow prescription gets numbers

  • The prescription, now concrete (0:54:18–0:54:29): under ~50 people, brute-force AI adoption; over 50, "set this up stack of intellectual resources on the side and start moving workflows one by one" — "One company took 22 steps and reduced them to four." Exact pilot numbers promised "during the next a few weeks" — watch for them. (→ model updated: this replaces the vaguer "exponential unit beside the core" step, verbatim + timestamp kept. Assumes an org big enough to run two stacks in parallel.)
  • New CEO diagnostic: "How fast is your Does the system increase its own capabilities?" (0:55:58) — the question he says he gives CEOs; pace-of-improvement as the standard, not the process.
  • New metric: "economic value produced on a unit of human participation" (0:37:02) — his benchmark answer in the GDPX segment.
  • Purpose answer (AMA Q2, 1:37:12): "We made a fundamental mistake by mixing busy[ness] with a purpose"; ⚠️ the "79% of corporate employees" disengagement figure is uncited on air (Gallup-shaped; say "most, per Gallup" if repeated).
  • Data-centre neighbours (AMA Q6, 1:49:02): factories shared prosperity through local hiring; data centres don't — the gain is "the use of intelligence that… comes out of this", not proximity.
  • In Madrid for the Singularity University summit; Peter not invited this year (0:14:21–0:14:43). The only mate with no Orin interest — "Salim, I'm sorry… Venmo, something for you" (0:21:58).

EP309 (Moonshots #309, published 2026-11-07, recorded ~11-06): his first episode in the chair

Read in full 2026-11-08 (detail in the research archive of ep2026-11-07-anthropic-s-ipo/).

  • Hosts: "I am soliam ismail your temporary host for today. Peter is away at I think presenting with uh, Mr. Isaacman of nasa" (0:03:26). He runs every story as read-in + "here's the question" + provocation, a debate-moderator format, with no provenance cues and no week's number, agenda card or AMA. The episode is the shortest panel episode in weeks ("we've come in much shorter", 1:45:24).
  • ★ New frame, opening and closing: "What if The singularity is not constrained by technology… our political our financial our social institutions are sub-linear" (0:31:15–0:31:28); at the close, "a more vicious Action reaction type of things from society" (1:42:45). This is the most institution-centred thesis he has put at the front of an episode.
  • 🔴 The NYC-vs-Anthropic energy math, self-labelled "death's advocate": NYC at 5.5 GW and $1.2T GDP against Anthropic at "a billion dollars of revenue per couple of gigawatts", then "eight billion in revenue per gigawatt", then "a billion and a hundred billion in revenue", concluding a city is "30 times more energy efficient" (1:10:47–1:12:17). The inputs change inside the segment. Anthropic's run-rate is >$80B (WSJ, 6 Nov), and NYC's 5.5 GW is average load, with peaks above 10 GW. Alex and Dave dismantle it on air; do not use.
  • The operator moment: "one of our pilot clients / Tried to get set up on one of the clouds… we're told i'm sorry. You're not big enough as a customer" (1:19:42–1:19:51).
  • Misread on the way in: "56 opposition to big tech and data centers" (0:06:28). The exit poll's 56% is "a crucial or major factor", and opposition to a local data centre is 76%.
  • Madrid: Spain needs "passion and… crazy", "We're not stereotyping at all on this show" (0:05:09–0:05:29).
  • Prescription (→ model: no change; it restates his "be active / own your MTP" step): "So you don't want to be passive… You want to be active?" (1:43:01), extended [attr?] (him or Dave): "you better be selling describing promoting publishing commenting… building" (1:43:16).

EP310 (Moonshots #310, aired 2026-11-11, recorded 11-10): from Oracle Park; outcomes not tasks; the AIC as the org story

Read in full 2026-11-12. Where he is: the San Francisco Giants' press box at Oracle Park, before a keynote from the dugout (0:22:55–0:23:52).

  • New frame: "we can delegate outcomes rather than tasks" (0:47:11); "AI's time horizon is going to now go faster than the time between board meetings, right? So who's really running the company?" (0:47:54–0:48:04).
  • Delaware AIC = "the most important story from an organization's perspective" (1:40:33). Prescription (→ model: no change; step 5, "put an AI in one real role and watch it", at entity scale): "Do very small functional, very clear, very well prescribed areas uh and then slowly broaden it out" (1:41:47).
  • Restated: 20 Gutenberg moments; AI "automates the adaptation mechanism itself" (1:03:00–1:03:12); the marriage riff; E.O. Wilson; Douglas Adams.
  • Pushback: challenges Elon's 5× robot productivity, "doesn't take into account uh maintenance and reliability, battery outages" (1:22:58–1:23:09); "correlation is causation here" is not the case for billionaires vs poverty (2:11:25).
  • 🔴 "how many people the average grocery store hires? It's 700" (1:52:00); Alex corrects on air (26–75).
  • Proposed metric: "how quickly can we lift the bottom" (2:12:18).

EP311 (Moonshots #311, aired 2026-11-13)

  • Angry at Paul Christiano: "where do you calculate the downside risk of… inaction?… where the hell is Pab" (0:47:11–0:47:33).
  • On archives: "the bottleneck was our cognitive ability to recognize it" (1:42:11), the opposite of Alex's sensing line two minutes later; the panel does not take it up.
  • Microsoft keeps distribution and compliance (1:16:37); ❓ Copilot "completely collapsed" (1:17:42). ⚠️ UBI: "14 major experiments… staggeringly successful" (1:45:04), no list given.
Reading rule. Anything marked unverified or on-air is a claim made on the podcast and has not been checked against a primary source yet. It stays marked until it has. Sourced numbers link to where they came from.