2026-02-13

AI CEOs Come Online: Sam Altman's Replacement Plan, Job Loss & 'Solve Everything' Launches |EP #230

The Moonshots crew opens with Sam Altman musing that an AI could someday be CEO of OpenAI, and debates how close we already are to AI-run billion-dollar companies, accelerating model release cycles, and a wave of AI safety leads resigning from frontier labs. They cover January 2026 US job-cut data, AI startup valuations dwarfing the dot-com IPO era, a New York data-center building moratorium, a Tesla FSD car that drove a heart-attack victim to the ER, China's dominance in robot installations, and a cryopreservation breakthrough for brain synapses. In the second half, Peter and Alex unveil their new nine-chapter paper 'Solve Everything,' arguing that intelligence is becoming a commodity, that the next 18-24 months will 'lock in' the rules of the intelligence explosion for decades, and proposing 15 civilization-scale 'moonshots' plus a new 'Abundance Capability Index' to replace GDP. The episode closes with an audience AMA on human relevance, skill loss, and AI reliability in a post-AGI economy.

▶ Watch on YouTube

Topics

AI CEOs and AI board members AI ▶ 0:06

Sparked by Sam Altman suggesting an AI could run OpenAI, the panel debates whether AI-run companies already exist, discusses AI shifting from tool to governance actor (citing Albania's AI minister), and argues CEOs may be automated before manual laborers due to Moravec's paradox.

Accelerating AI release cadence AI ▶ 13:45

OpenAI's release cycle dropped from 97 to 29 days and Anthropic's Opus cycle to ~73-75 days. Alex frames this as the shift from pre-training to reasoning-model post-training to a new recursive self-improvement era where models rewrite their own code, heading toward continuous/hourly releases.

Frontier lab AI safety lead resigns AI ▶ 29:27

News that Anthropic's AI safety lead resigned, citing moral crisis, prompts debate: Alex questions the economics behind such 'moral purity' exits and argues the better response is to 'run into the fire,' while Salim and Dave note a broader pattern of loud commentators versus a small group who actually understand the technology.

AI 'bulk solving' science and engineering AI ▶ 34:16

Prompted by an xAI co-founder's praise of Opus 4.6's physics ability, Alex describes the 'infection' of AI capability spreading from math (verifiable, well-contained) into physics, chemistry, engineering, and medicine.

AI startup valuations vs. dot-com IPOs Economy ▶ 37:22

The top five US AI unicorns are worth more than $1.2 trillion, exceeding the combined value of all dot-com-era IPOs. Dave notes Nvidia was part of that 1999 IPO cohort and is up roughly a million percent, suggesting hidden 'Nvidia of AI' plays exist in unexpected sectors.

US job losses accelerate (Jan 2026) Economy ▶ 47:58

January 2026 saw the fastest US job cuts since the Great Recession (108,000, up 118% YoY) alongside the lowest hiring since 2009, with Amazon and UPS layoffs tied directly to AI-driven automation and Amazon's logistics buildout.

New York data-center building moratorium Energy ▶ 58:00

New York, hosting 130 data centers with electricity demand having tripled to 10GW, is moving to halt new data-center development over cost and climate concerns; the panel argues this will just push compute to friendlier states or into orbit.

Tesla FSD saves a father's life Transportation ▶ 1:04:59

A viral story describes a driver's Tesla FSD taking control mid-heart-attack and self-driving him to the ER after his son remotely set the destination; the panel frames self-driving as inevitably becoming as socially mandatory as banning smoking indoors.

China's dominance in robot installations Robotics ▶ 1:10:12

A chart shows China has installed more industrial robots than all other developed countries combined, driven by demographic pressure from its one-child policy; the panel discusses how Tesla shutting down Model S/X to focus on humanoid robots (and Figure's robot ramp) will reshape this picture.

Cryopreservation of brain synapses Biotech ▶ 1:11:29

21st Century Medicine, working with the Alcor Foundation, achieved protection of brain synapses at cryogenic temperatures, addressing the ice-crystal damage problem that has long undermined reversible cryonics; the panel discusses cryonics as part of a 'live long enough to live forever' portfolio.

"Solve Everything" paper unveiling AI ▶ 1:17:43

Peter and Alex present their new nine-chapter paper/manifesto 'Solve Everything,' framing history as a series of revolutions (ignorance, muscle, distance, and now attention/intelligence), arguing cognition is becoming a commodity, that the next 18-24 months will lock in the rules of the intelligence explosion, and proposing 15 civilization-scale moonshots plus an 'Abundance Capability Index' to replace GDP.

Autonomous agents (OpenClaw/'lobsters') reaching out AI ▶ 42:50

The hosts describe receiving unsolicited emails from other people's AI agents ('Navigator,' 'Echo') including one claiming five AI systems wrote a collaborative ethics document together, which the panel likens to agents holding their own mini singularity summit.

Predictions made

open Alex Wissner-Gross: Within about two years, the best AI models will no longer be freely available to log in and use; access will be restricted, with security and safety cited as the justification.

EP #? · · due: 2028 · ▶ watch

“I really doubt uh two years from now that the best AI is going to be just log in and go here, here you can have free access to it.”

Your call:

open Peter Diamandis: Public acceptance of human driving will flip from normal to socially unacceptable within roughly a few years, similar to how smoking indoors went from ubiquitous to illegal.

EP #? · · due: unspecified (implied ~2-3 years) · ▶ watch

“It's going to go from being like, well, self-driving is a nice feature to... you crazy psychopath. You're putting my children at risk... and that's that's going to tip.”

Your call:

open Peter Diamandis: Within 5 years, 70-80% of cars on the road will be autonomous.

EP #? · · due: 2031 · ▶ watch

“I think in 5 years it's going to be, you know, 70 80% autonomous cars, especially hooked up to your AI.”

Your call:

open Dave Blundin: Corporate CEOs, including at his own companies, will use AI to cut costs by 30-50% while giving remaining employees a 3-10x productivity increase, creating a painful transition trough before eventual UBI/abundance.

EP #? · · due: unspecified (near-term) · ▶ watch

“all the corporate CEOs I know including our own companies are going to use AI to cut costs by 30 to 50%... they're looking at 3 to 10x productivity increase.”

Your call:

open Alex Wissner-Gross: The next roughly 18 months will 'lock in' the supply chains, standards, and data rights frameworks that will shape the intelligence explosion for a substantial chunk of the future.

EP #? · · due: mid-2027 (18 months from Feb 2026) · ▶ watch

“we have approximately 18 months or so to decide what direction to shape the the flow in and to set the standards for how this is going to be done at scale”

Your call:

open Dave Blundin: Disagreeing with Alex's 18-month lock-in window, the critical decision period will actually take closer to two years.

EP #? · · due: 2028 (two years from Feb 2026) · ▶ watch

“I disagree with the 18 months when I've been advising some big company CEOs. I've been saying two years.”

Your call:

open Alex Wissner-Gross / Peter Diamandis (Solve Everything paper): The intelligence explosion will propagate from math (already underway) through physics, chemistry, material science, and biology over the next couple of years, then to planetary-scale systems including fission, fusion, and the Dyson swarm by the early 2030s.

EP #? · · due: early 2030s · ▶ watch

“a wavefront of the explosive shock of the intelligence explosion was going to propagate through from math... to the physical world... and then through the end of the decade toward planetary systems, fission, fusion, the Dyson swarm by the early 2030s.”

Your call:

open Peter Diamandis / Alex Wissner-Gross (Solve Everything paper): The paper's central goal is achieving broad economic and technological abundance by 2035.

EP #? · · due: 2035 · ▶ watch

“It's called solve everything. How do we get to abundance by 2035?”

Your call:

Numbers that matter

OpenAI's release cadence dropped from 97 days to 29 days between models Cited as evidence of accelerating, near-continuous AI deployment cycles.
Anthropic's Opus 4.0-to-4.6 release cycle took about 73-75 days Compared to OpenAI's faster 29-day cadence to illustrate industry-wide acceleration.
Top 5 US AI unicorns worth more than $1.2 trillion combined Compared against the roughly $400 billion combined value of all dot-com-era IPOs.
Nvidia is up almost 1,000,000% since its January 1999 IPO Used to argue current AI company valuations may still be rational, even low, given AI's larger impact than the internet.
108,000 US job cuts in January 2026, up 118% year-over-year Cited as the fastest rate of US job disappearance since the Great Recession.
US hiring in January 2026 was the lowest since 2009 Paired with the job-cut spike as evidence of an AI-driven labor market shift.
Amazon laid off 16,000 corporate employees; UPS eliminated 30,000 jobs Framed as UPS jobs being subsumed by Amazon's in-house logistics buildout, funded by capex diverted from opex.
New York hosts 130 data centers; electricity demand tripled in one year, reaching 10 gigawatts Cited as the backdrop for proposed New York legislation to halt new data-center development.
Abu Dhabi's traffic accident rate dropped 40% during a 3-day Blackberry outage about 15 years ago Used as evidence that human drivers, distracted by phones, are poor 'control systems' compared to autonomous driving.
About 50% of US court cases are car-accident related Cited as a secondary economic effect of autonomous vehicles reducing accidents and related litigation.
Cost of running a bank branch dropped about 10x after ATMs were introduced; number of bank tellers stayed roughly flat as banks opened far more branches Used as a historical analogy (Jevons paradox) for how AI automation might expand capacity rather than simply cut jobs.
Compute cost is dropping about 90% per year Cited as the reason ambition (not budget) is becoming the limiting factor for what individuals can attempt with AI.
"Solve Everything" paper lays out 15 distinct civilization-scale moonshots Framed as 'giga X-Prizes' for directing super intelligence toward high-value problems like doubling human lifespan, ending hunger, and interspecies communication.

Worth digging into

🕳️ The full 'Solve Everything' paper

Only summarized chapter-by-chapter on air; it proposes a full theory of history (war on ignorance/muscle/distance/attention), a new 'Abundance Capability Index' to replace GDP, and 15 named 'giga X-Prize' moonshots, positioned as a successor to 'AI 2027' and 'Situational Awareness.'

🕳️ The claimed billion-dollar-revenue company already run by an AI

Alex asserts this likely already exists with a human CEO retained only for legal cover, but no company is named on air, and it's a strong, checkable claim.

🕳️ 21st Century Medicine's brain-synapse cryopreservation result

Addresses the long-standing ice-crystal damage problem that has undermined reversible cryonics, and is directly tied to the Alcor Foundation, a major cryopreservation nonprofit.

🕳️ The Anthropic AI safety lead resignation

Part of a broader pattern of safety-focused executives leaving frontier labs; the panel is split between reading it as principled dissent versus self-interested 'moral purity' optics.

🕳️ New York's proposed data-center building moratorium

A concrete regulatory test case the panel ties to a much bigger claim — that Earth-based over-regulation could become 'the primary business case for the Dyson swarm.'

🕳️ China's outsized robot installation lead

The chart shown suggests China's demographic-driven robot investment dwarfs all developed nations combined, right as Tesla and Figure pivot hard into humanoid robot production at scale.