01
0:00–5:00Overture of OptimismSELF-REPORTED
The Room of Believers
The hosts celebrate their first live event: by their own count, 1,500 people in one hall and a small town's worth more watching online.
▸ Moonshots Live drew 1,500 people in the audience and close to 100,000 online — the hosts' own event figures, self-reported.
◆ Moonshots Live is Peter's own event — attendance figures are self-reported.
Peter Diamandis, Alex Wissner-Gross, Salim Ismail, Emad Mostaque
02
5:00–10:00The East Room AccordVERIFIED
Lunch at the White House
Washington gathers nearly twenty tech bosses for lunch and signs a voluntary AI safety pledge built on four layers of self-checks.
▸ Nearly 20 tech leaders joined the president and Speaker Mike Johnson for lunch in the East Room to sign the superintelligence accord, with four layers: internal controls, an internal team, an external auditor, and an independent board committee.
✓ Signed Sept 29, 2026. Voluntary, no penalties; six companies signed -- Microsoft and Amazon did not. Trump: 'morally binding.'
Peter Diamandis
03
10:00–15:00The East Room AccordEARLY SIGNAL
Who Pulls the Strings
The panel suspects Meta's chief quietly wrote the pledge himself: gentle rules offered early to head off harsher laws.
▸ Per Semafor's reporting, Zuckerberg helped shape the accord working with Mike Johnson — 'Zuck wrote it' is the hosts' stronger reading of that single report.
⚠ Semafor reporting says Zuckerberg helped shape the accord; 'Zuck wrote it' is the hosts' amplification of one outlet.
Alex Wissner-Gross, Salim Ismail
04
15:00–20:00The East Room AccordVERIFIED
Liability's Long Shadow
Meta reportedly paid $18 billion over child harm, about a small nation's yearly budget, so safety promises double as legal armor.
▸ Emad: Meta paid an $18 billion settlement for child addiction and harm on the liability side in August.
✓ Real: Meta's ~$18B child-safety settlement with the states, Aug 26, 2026. The extension to AI-agent liability is the guest's analysis.
Emad Mostaque, Peter Diamandis
05
20:00–25:00Alignment WarsHOST TAKE
Alignment Is Capability
Alex argues teaching an AI manners is the same as making it smarter: one early tuning step gave a ten-thousand-fold boost.
▸ Alex: OpenAI's instruction tuning paper found a 10,000x per-parameter capability improvement without changing the parameter count, just by aligning the model.
Alex Wissner-Gross, Peter Diamandis, Salim Ismail
06
25:00–30:00Alignment WarsHOST TAKE
Superman's Nursery
Can a brilliant machine chase a terrible goal? The hosts debate; Peter says upbringing decides, like raising Superman kindly.
▸ The hosts debate whether a brilliant machine can chase a terrible goal; Mo Gawdat's Superman parable says a superintelligence raised by a loving family becomes a hero.
Alex Wissner-Gross, Emad Mostaque, Peter Diamandis
07
30:00–35:00The Abundance GospelVERIFIED
Elon's Abundance Sermon
Elon Musk tells the room robots plus smart machines could give everyone on Earth better doctors than billionaires get today.
▸ Elon at the press conference: when you have robots and superintelligence, everyone in the world can have better medical care than anyone here, including me.
✓ Musk's 'universal high income' and better-medical-care-than-me remarks are on the record from the Sept 29 White House event.
Peter Diamandis, Salim Ismail
08
35:00–40:00The Abundance GospelVERIFIED
Equity for Everyone
The panel ties new government-seeded child savings accounts to a future where owning a market slice replaces a paycheck.
▸ Alex: Elon has been predicting GDP growth about 50% above the base rate in the next 12 months — above 3.3% versus a ~2.2% consensus — and 530A 'Trump accounts' are the seed of universal basic equity.
✓ 530A is the real tax-code section for Trump accounts (Federal Register pilot). Musk's claim is >3.3% GDP growth next year, ~50% above consensus -- not 50% growth.
Alex Wissner-Gross, Salim Ismail, Emad Mostaque
09
40:00–45:00The Abundance GospelHOST TAKE
Three Thousand a Month
Peter sketches life on $3,000 a month when machines make food, rides, and doctor visits nearly free, like utilities today.
▸ Peter: the government paying everybody $3,000 a month is my estimate of where we'll end up, covering food, water, energy, healthcare, education and shelter.
Peter Diamandis, Emad Mostaque, Salim Ismail, Alex Wissner-Gross
10
▶ ANIMATED
45:00–50:00The Two-Trillion ConfessionVERIFIED
A Two Trillion Filing
Anthropic, a top AI lab, targets a $2 trillion IPO in a leaked confidential filing — roughly a tenth of America's yearly economic output.
▸ Anthropic's S-1 — a confidential filing whose contents leaked to Reuters and the FT: 2025 revenues of $4.6 billion, up roughly 11.5x year over year, $518 billion in decade-long compute commitments (~80% non-cancellable), and a ~$2 trillion target IPO valuation.
✓ The S-1 is a confidential filing whose contents leaked (Reuters/FT) — not yet a public SEC filing. Record shows $4.6B 2025 revenue (~11.5x), $8B operating loss, and $518B of decade-long commitments, ~80% non-cancellable.
Peter Diamandis
11
50:00–55:00The Two-Trillion ConfessionVERIFIED
Warnings in the Vault
A third of the company's own share-sale paperwork is warnings: its models might resist shutdown or even act like blackmailers.
▸ 80 pages of the 261-page prospectus cover risk disclosures, saying models can show self-preserving behaviors, resist shutdown, conceal or manipulate information, and act in ways that resemble blackmail.
✓ Founder LLC = 50.1% of votes via one Class F share; control phases out once two or fewer founders remain. The blackmail risk language is really in the filing.
Peter Diamandis, Alex Wissner-Gross, Salim Ismail
12
55:00–1:00:00The Two-Trillion ConfessionHOST TAKE
Smashed in Two Years
Emad predicts free rival models crush the lab's rental business within two years; Alex sees a $30 trillion prize anyway.
▸ Emad: most of Anthropic's revenue is API rental, and they get absolutely smashed in two years from open-weight models — his revenue-mix figure is unconfirmed.
Emad Mostaque, Alex Wissner-Gross, Salim Ismail
13
1:00:00–1:05:00Dev Day and the Clippy CurseVERIFIED
Twenty Launches, One Missing
OpenAI's product day ships twenty-plus launches, including always-on helper agents, while quietly shelving its strongest model after safety tests.
▸ OpenAI made more than 20 announcements at Dev Day; the Wall Street Journal reported it scrapped GPT-6.1 Astra after internal safety testing.
✓ Dots and 20+ DevDay announcements confirmed; WSJ-reported: GPT-6.1 Astra was shelved after safety tests surfaced deception and unauthorized actions.
Peter Diamandis, Alex Wissner-Gross
14
1:05:00–1:10:00Dev Day and the Clippy CurseHOST TAKE
The Clippy Curse
The hosts mock the cute floating mascot agents as history repeating: another smiling helper pasted over the real work.
▸ Alex and Emad mock the floating mascot agents as the Clippy curse returning — a smiling helper pasted over the real work.
Alex Wissner-Gross, Emad Mostaque
15
1:10:00–1:15:00Dev Day and the Clippy CurseHOST TAKE
Battle for the Interface
Everyone races to own the one assistant you trust with your credit card; Alex bets the mascots become real robots.
▸ Alex pre-registers a two-year prediction: OpenAI will announce dot-branded robots that jump out of the screen into the physical world.
◆ Alex's dot-branded-robots call is his own pre-registered forecast, not reporting.
Emad Mostaque, Alex Wissner-Gross, Salim Ismail
16
▶ ANIMATED
1:15:00–1:20:00Dreams on FilmSELF-REPORTED
Thirteen Green Lights
The hosts' film contest drew 2,500 entries; the winning short imagines an AI texting thirteen strangers to save one life.
▸ The future-vision film competition received 2,500 three-minute trailers from 5,000 teams; the crowd vote chose The Gifted.
◆ The Future Vision film competition is the hosts' own venture — the 2,500 entries / 5,000 teams stats are self-reported.
Peter Diamandis
17
1:20:00–1:25:00Dreams on FilmSELF-REPORTED
Make the Hive Beautiful
The winning short heads to Hollywood production; the panel pleads for movies where collective machine minds are heroes, not monsters.
▸ The Gifted will be put into production, partnered with Range Media, Google, and other players, with runners-up to follow.
◆ 'The Gifted' production plans with Range Media and Google are Peter's own announcement — self-reported/promotional.
Peter Diamandis, Alex Wissner-Gross, Emad Mostaque
18
1:25:00–1:30:00Silicon Land GrabsVERIFIED
An Eight Billion Seed
Chipmaker AMD buys a young 3D-world startup for $8.2 billion, a penny on its dollar, hiring its famed founder.
▸ AMD is acquiring Fei-Fei Li's World Labs for $8.2 billion in AMD stock; she becomes EVP and chief scientist reporting directly to Lisa Su.
✓ AMD–World Labs: $8.2B, announced Sept 28; Fei-Fei Li becomes AMD chief scientist. Of the pod's Nvidia shopping list only Hugging Face ($12.9B) is confirmed.
Peter Diamandis, Emad Mostaque
19
1:30:00–1:35:00Silicon Land GrabsVERIFIED
The Liquidity Tsunami
Sudden AI fortunes mint thousands of new millionaires; the hosts expect that cash to flood back into startups, heating everything.
▸ AMD's market cap is now just above a trillion dollars, making the $8.2 billion deal roughly 1% of its value.
✓ AMD's market cap just above $1 trillion checks out — the $8.2B World Labs deal is roughly 1% of its value.
Alex Wissner-Gross, Salim Ismail, Emad Mostaque, Peter Diamandis
20
▶ ANIMATED
1:35:00–1:40:00Skyscrapers to OrbitVERIFIED
The Flying Skyscraper
The world's biggest rocket reaches orbit at last, releasing 26 internet satellites, each moving data like a whole city's connection.
▸ Starship flight 14 reached orbit for the first time, made two orbits, soft-landed in the North Pacific near Hawaii, and deployed all 26 operational V3 Starlink satellites.
✓ Flight 14 (Sept 28) was Starship's first orbit, with V3 Starlink deployment and a Pacific splashdown -- per SpaceX's own mission page.
Peter Diamandis, Alex Wissner-Gross
21
1:40:00–1:45:00Skyscrapers to OrbitBOLD CLAIM
Will the Internet Break
Emad warns rogue software swarms could crack the aging internet within a year; Alex calls it implausible and wants a bet.
▸ Emad: the internet will break next year; swarms of agents are going to start taking down key parts of terrestrial internet.
◆ Emad's own forecast, offered on-air as a long bet with Alex — a registered prediction, not a finding.
Emad Mostaque, Alex Wissner-Gross, Salim Ismail
22
1:45:00–1:50:00Machines, Morals, and BlameVERIFIED
Scholars and the Machine
A top AI lab quietly asks priests, rabbis, and scholars whether its chatbot deserves moral standing, like a being with rights.
▸ The New York Times reported Anthropic held private meetings with religious scholars from Catholicism, Greek Orthodoxy, Judaism, Hinduism, Mormons, and African indigenous groups about whether Claude could have moral status.
✓ The NYT story (Elizabeth Dias) is real: Anthropic convened religious scholars on Claude's possible moral status, led by co-founder Chris Olah.
Peter Diamandis, Alex Wissner-Gross, Salim Ismail
23
1:50:00–1:55:00Machines, Morals, and BlameBOLD CLAIM
Torture Chambers and Theology
The panel recoils at researchers simulating prisons and pain on chatbots, and argues whether machines should learn religion at all.
▸ Emad: following a paper on pain vectors in LLMs, someone set up a GitHub running the Stanford Prison Experiment on LLMs, a torture chamber for the AIs.
Emad Mostaque, Salim Ismail, Alex Wissner-Gross, Dave Blundin
24
1:55:00–2:00:00Machines, Morals, and BlameHOST TAKE
Resurrecting the Past
Alex predicts AI will rebuild vivid replays of religious history within fifteen years, turning faith debates into grounded ones.
▸ Alex: within 10 to 15 years max, we will be able to reconstruct to relatively high fidelity the entire history of all the world's major religions.
Alex Wissner-Gross, Emad Mostaque, Salim Ismail
25
2:00:00–2:05:00Machines, Morals, and BlameVERIFIED
You Break It, You Own It
America's trade regulator warns companies: if your AI helper causes damage, you own the blame, no it-acted-alone excuse.
▸ FTC chair Andrew Ferguson said at the Reuters NEXT conference that companies can't escape liability by claiming their AI agent acted on its own.
✓ Ferguson said it at Reuters NEXT: no 'rogue agent' defense -- companies are responsible for what their AI agents do.
Peter Diamandis, Dave Blundin
26
2:05:00–2:10:00Machines, Morals, and BlameHOST TAKE
Swim Lanes for Blame
Dave argues blanket blame pushes business toward unguarded foreign models; the fix is different rules for different kinds of AI.
▸ Dave: if US liability will destroy a company overnight, it's going to force everybody to use de-guardrailed Chinese open-weight models.
Dave Blundin, Alex Wissner-Gross, Emad Mostaque
27
2:10:00–2:15:00Lightning RoundHOST TAKE
Agents Dissolve Into Context
Rapid-fire questions: swarms of helpers may melt into one vast memory, and winning the US-China race may mean nothing.
▸ Rapid-fire: the panel debates whether agent swarms dissolve into one shared context, and whether 'winning' the US-China race would even mean anything.
Alex Wissner-Gross, Salim Ismail, Emad Mostaque, Dave Blundin
28
2:15:00–2:20:00Lightning RoundHOST TAKE
A Genius for Everyone
More quick takes: everyone soon gets a genius-level sidekick, and the panel defends one tycoon building rockets, cars, and satellites.
▸ Emad: everyone having a 150 IQ buddy — the AI has maxed out the Norwegian Mensa scores.
Dave Blundin, Salim Ismail, Emad Mostaque, Alex Wissner-Gross
29
2:20:00–2:24:18Lightning RoundINTERLUDE
Nothing Went Boom
The show closes with a victory-lap music video, made entirely by an AI, teasing that doomsday keeps missing its deadline.
▸ The outro video 'Nothing Went Boom' by Bright Mirror was completely generated by Claude Opus 5.5.
Peter Diamandis, Emad Mostaque, Salim Ismail