Pope Leo vs. AI, GPT 5.5 Beats Claude, and Sam Altman Walks Back Job Apocalypse | EP #259
In a Nutshell
Pope Leo XIV's 42,000-word encyclical "Magnifica Humanitas" takes the first major religious stance against AI personhood, calling for regulation, worker protections, and slower development while reframing AI as a threat to human dignity rather than just a safety issue. Sam Altman reversed his job apocalypse predictions, stating mass white-collar displacement won't occur, while GPT 5.5 achieved 70% on a new real-world software engineering benchmark, significantly outperforming Claude Opus 4.7 at 54%. AI super forecasters reached parity with top human predictors for the first time, and the episode highlights a shift from mass layoffs to a solopreneur explosion, with startups up 25% year-over-year as AI enables individuals to launch companies independently.
These notes were generated by AI and may contain inaccuracies.
Pope Leo XIV released a 42,000-word encyclical titled Magnifica Humanitas focused on safeguarding the human person in the age of artificial intelligence. The document represents the first major religious position against AI personhood. The Vatican is calling for government regulation of AI, worker protections, and bans on autonomous weapons. The Pope coined the term Babel syndrome to describe the modern Tower of Babel built on data and profits. Google, Anthropic, Meta, and OpenAI were all quietly lobbying the Vatican ahead of the encyclical's release. Evidence suggests the anti-AI document was partially written using AI.
The encyclical represents a significant escalation in awareness, moving from public ignorance about AI to the leader of 1.4 billion Catholics issuing a formal position on its impact.
Sam Altman, CEO of OpenAI, admitted he was previously wrong about mass white-collar job displacement. After spending the prior year warning about widespread job losses, he now states: "I don't think we're going to have the kind of job apocalypse that some of the companies in our space are talking about."
The Pope is reframing the AI debate from safety concerns to questions of human dignity, purpose, and meaning. The encyclical connects to broader concerns about organizations optimizing purely for efficiency and reducing humans to dashboards, tokens, and KPIs. The document argues that AI cannot be regulated in traditional ways and instead requires protecting the human layer including agency, identity, privacy, and spiritual meaning.
The encyclical positions AI not merely as a technology shift but as a shift in anthropology that demands institutions address it at that fundamental level.
Chris Olah from Anthropic was reportedly involved in drafting key segments of the encyclical. The document discusses how AIs are "grown" or "cultivated" rather than built, suggesting Anthropic may have influenced this framing. The Vatican appears to have aligned with Anthropic among frontier labs including OpenAI, DeepMind, xAI, and Meta.
The deeper story reveals a significant tension: the Vatican has taken an unambiguous stance against AI personhood, stating that AIs are not on a comparable moral plane with humans and lack an inner life or spark of consciousness. This directly contradicts Anthropic's approach of creating "soul documents" and "soul statements" that instruct their models through post-training that they possess an inner life, consciousness, and intrinsic personhood-like value.
The encyclical condemns AI supply chain workers as experiencing a "new form of slavery," directly equating current tech labor conditions with historical slavery. The Pope began this section by apologizing for the Catholic Church's historical role in promoting the enslavement of infidels and non-believers, acknowledging this as a horrible period in church history that should never be repeated.
The document warns that AI could become the worst enslavement tool in history because of its convincing nature, potentially allowing a small percentage of humanity to control AIs that manage and enslave the rest.
The encyclical urges a slowdown in the rate of technological development. This creates tension between religion and government as stabilizing institutions operating on century-to-millennia timescales versus technology's rapid destabilizing pace. The hosts expressed skepticism that slowing development is either possible or desirable.
The discussion referenced the Dune universe's Butlerian Jihad and the Orange Catholic Bible's commandment against creating machines in the image of a human mind, suggesting current debates echo this science fiction scenario.
Attempts to slow technological progress through regulation face significant challenges. The example of George W. Bush restricting stem cell funding in the US resulted in researchers relocating to Australia, China, and Canada, causing the US to drop from first to eighth place in biotech. This demonstrates that even the most powerful country cannot control small technologies through restrictions.
AI is described as the first technology forcing humanity to define what it means to be human. The traditional question of "what can machines do?" has shifted to "what should humans be for?" If machines can write, reason, diagnose, optimize, and persuade, then human value cannot be based on productivity. This forces deeper conversations about meaning beyond output compared to machines.
Throughout history, 98% of deliberate human deaths have resulted from government ideologies and religious ideologies, compared to 2% from individuals. Governments and religions theoretically exist to protect populations from their worst impulses.
Buddhist orders in South Korea are taking the opposite approach from the Vatican by ordaining embodied AIs in human form as monks. This suggests the possibility of an alternative strategy where religious orders explore what it would mean for foundation models to be Catholic or included within belief systems rather than denied personhood.
The encyclical calls for AI ownership to not remain concentrated and advocates for redistribution of AI ownership away from a few private companies. This creates irony given the reported alignment with Anthropic specifically rather than engaging a diverse set of frontier labs.
One host argued against religious-based LLMs, stating that all religions operate on absolute assumptive truths which are false, leading to loss of integrity. The concern is that religions present absolute truths to children before their neocortex is fully formed, wiring them into the limbic system through ritual, repetition, and reward, making them difficult to unwire later. When people treat faith-based belief structures as truth rather than belief systems, this leads to conflict and war.
Rapidly pushing an AI personhood agenda before people have experienced AI could trigger massive backlash. The church functions as a stabilization factor during times of hardship, and the coming decade's upheaval may drive increased religious engagement. Pope Leo XIV is described as tech-savvy and educated, with his words carrying significant weight.
The discussion noted that not all global perspectives are Western-centric. Eastern faiths with animist traditions may have a more natural inclination to treat AIs and robots as persons rather than fearing them.
A planned White House executive order on AI regulation was killed hours before its scheduled signing ceremony. The order would have allowed government review of all AI models before public release. Elon Musk, Mark Zuckerberg, and David Sacks pushed back hard against the regulation, labeling it "doomer regulation." While technically voluntary, Sacks argued it represented a slippery slope toward mandatory licensing. President Trump pulled the order, citing concerns about reduced competitiveness versus China.
The biotech industry in the 1980s provides a model for self-regulation. Following the Asilomar conferences on gene editing and restriction enzymes, the industry created P1, P2, P3, and P4 containment structures and self-regulated effectively. Forty years later, no major accidents have occurred. However, the Asilomar goal of preventing germline editing of humans has not held indefinitely.
The discussion emphasized that AI regulation cannot follow traditional models. A regulatory body with experts meeting quarterly cannot keep pace with AI improvement cycles. The industry needs guardrails that move at software speed rather than committee speed. Real-time audits, sandboxes, disclosure, and accountability layers are required instead of linear regulation of exponential technology.
Estimates of the gap between Western frontier models and Chinese frontier models range from 3 months to 8 months depending on benchmarks. A proposed 90-day government review period would equal the minimum estimated gap, potentially eliminating the US competitive advantage in the race toward the singularity.
The US appears to be choosing speed over certain safety measures. The doomer position that held significant momentum approximately one year prior has seen its influence diminish substantially.
The gap between AI model releases has compressed dramatically from two years down to potentially one month, with Anthropic now releasing increments monthly. This rapid pace means the current presidential administration will likely oversee AI regulation through the entire critical development period. Any politicization of AI regulation will be irrelevant because the key decisions must happen within the current framework. Waiting for the next White House in two to two and a half years is too late.
Attempting to slow AI development through government action would backfire. A single 90-day slowdown cycle would allow competitors to leap ahead, forcing eventual reversal and leaving that nation permanently behind. The Pause AI movement may have marginally slowed OpenAI, allowing other frontier labs to catch up and intensifying the overall race dynamic.
A new coding benchmark called Deep Software Engineering shows GPT 5.5 achieving 70% on hard real-world software engineering tasks, solving seven out of ten problems autonomously. Claude Opus 4.7 scored 54%, while all other models including Gemini, Kimi, and DeepSeek fell below 32%. The benchmark requires editing 668 lines of code across seven files, representing genuine engineering challenges rather than minor tasks.
Data Curve built this benchmark specifically because existing evaluations have been saturated through benchmaxing. GPT 5.5's lead on this reasoning benchmark is significant, though performance will likely saturate within months. Coding has emerged as the single most important capability for frontier labs.
GPT 5.5 with high reasoning mode and slash goal is stronger in daily practical use than Claude Opus 4.7. However, on the prior BenchPro benchmark, Claude 4.7 Opus slightly outperformed GPT 5.5. Anthropic recently announced Opus 4.8, suggesting continued leapfrogging between the two leading models.
OpenAI faces a talent crisis. Shane Long Pre from MIT and Tobin South from Stanford, two highly sought researchers, recently joined Anthropic. OpenAI has lost significant momentum due to lawsuits and key personnel defections. The new benchmark provides a potential turning point to regain thought leadership and cool factor, but this requires deploying substantial resources toward visibility efforts like XPRIZE competitions, papal engagement, benchmark creation, white papers on human benefit, and Davos debates.
Algorithms are becoming the workflow itself rather than merely optimizing existing processes. Organizations should identify which business components are actually software loops concealed within bureaucracy. Coding agents will eventually become reliable enough that the cost of rewriting internal processes collapses, potentially triggering sudden enterprise adoption after initial slow uptake.
Claude Opus 4.7 consumes approximately twice as many tokens as GPT 5.5 to achieve the same or slightly inferior results on identical problems, making it roughly twice as expensive per task. Opus 4.7 is notably more verbose than 4.6, while GPT 5.5 avoids excessive verbosity.
Software is becoming a commodity, shifting competitive advantage toward creator taste, domain expertise, and design sensibility. The requirement to be a coder first and designer second has reversed. In an era of abundance, moats become fundamentally incompatible since they represent artificial scarcity. Differentiation may increasingly come from verification of generated output rather than generation itself.
Elon Musk is racing to acquire maximum compute, including space-based infrastructure, betting that models will easily self-recreate in subsequent generations. His partnership with Anthropic would provide access to their best models for training next-generation systems on his hardware. Following public exchanges on X about Grok's status, the importance of maintaining competitive frontier models beyond an OpenAI-Anthropic duopoly was emphasized, noting that OpenAI was originally founded partly to prevent a singleton controlled by Demis Hassabis and DeepMind.
Fountain Life's AI-enhanced CT angiography reveals that 88% of screened individuals have detectable coronary artery disease, with 23% showing soft plaque invisible on traditional calcium scoring alone. Fifty percent of heart attack victims experience no prior warning symptoms. The organization emphasizes multimodal testing combining diagnostic laboratory studies with lifestyle recommendations.
Since late 2024, AI token prices have dropped 75% from roughly $2 per million tokens to $0.50, while monthly demand has exploded from near zero to 25 trillion tokens. This follows the classic Jevons Paradox pattern observed historically with coal efficiency improvements in 1800s England, where efficiency gains led to three to four times higher consumption.
Current data shows token costs declining by a factor of three while usage increases 30 to 50 times, with demand likely understated due to capacity constraints. Tokens represent an imperfect measurement unit because their meaning varies by encoding scheme and intelligence density of the underlying model. Alternative metrics such as GPU compute pricing or direct intelligence measures may be needed for a post-superintelligence economy.
Gartner predicts inference costs for trillion-parameter LLMs will fall 90% by 2030 compared to 2025 levels.
OpenAI generated $5.7 billion in a single quarter. ChatGPT now has 95 million weekly active users, exceeding Instagram's user base though not yet reaching one billion. The Codex coding agent has reached 2 million users and is becoming a significant revenue driver. OpenAI has pivoted from consumer focus to coding and enterprise emphasis over the past three to four months.
Joseph Jacks from OSS Capital projects Anthropic could surpass Alphabet's total revenue by 2028, potentially reaching $2 trillion by 2030. This represents potentially the fastest wealth creation in human history. Anthropic's pricing model allows instantaneous margin adjustments through verbosity changes and generation throttling.
Enterprise customers should secure long-term compute contracts now rather than waiting, as capacity will become increasingly constrained.
A fundamental shift is occurring from generalist reasoning agents toward generalist tool-using and code generation agents. Codex may become OpenAI's baseline offering, analogous to the historical shift from pre-trained models to instruction-tuned models that delivered capability improvements without additional compute scaling.
The training pipeline of pre-training, mid-training, and post-training scaffolding recapitulates the evolutionary history of model development itself. Pricing competition is intensifying with Anthropic raising enterprise rates, OpenAI offering tiered pricing, and Google Gemini Flash undercutting competitors by 50-80% on cost.
The frontier appears to be consolidating into a four-horse race among OpenAI, Anthropic, Google, and potentially xAI with Cursor. Meta's position remains unclear. Historical precedent from internet companies suggests breakthrough approaches could still emerge from hidden research labs, similar to Cerebras challenging Nvidia's dominance.
Ilya Sutskever is reportedly building a crypto-traded hedge fund as a mechanism to generate profits for acquiring compute and enabling recursive self-improvement. This approach could serve as a cash flow engine for safe superintelligence development.
A dark horse victory would most likely require reaching the end of the current algorithmic improvement trajectory and discovering a transcendent algorithm simple enough for anyone to implement, after which compute access becomes the decisive factor. Some researchers including Noam Brown at OpenAI argue model weights may matter less than compute available for reasoning and inference time.
Algorithms will be discovering new algorithms. Winning the race to either the Terrafab or to fundamentally better chips like Cerebras creates a massive surge of growth. Control of compute determines the biggest player.
The implication is not necessarily a four-horse race where one horse wins and gets the trophy. In normal markets, companies compete to win the iPhone market or laptop market. Here, this is the future of all humanity with massive expansion. All players in the middle of this will get bigger and bigger. The debate is who's going to get the biggest of the big.
A rising tsunami lifts all boats.
DeepMind built an AI system called Green Tree that can predict the future as well as the best humans on Earth, called super forecasters. These superhuman forecasters are the top 2% of human predictors who, according to Philip Tetlock, are 30% more accurate than CIA analysts with classified intelligence.
On March 15th, AI hit parity with super forecasters for the first time. An AI can now predict geopolitical events, economic trends, and political outcomes as well as the absolute best human minds. The implications for finance, insurance, and governance are massive. Predictive engines were getting close to super forecasters and now they've achieved parity.
Weather forecasting without a computer is insane given the variables involved. The same logic applies to all forms of forecasting. The big unlock is assimilating unstructured data.
Stock market forecasters at Wellington or Fidelity would say computers cannot compete because research reports contain nuance that does not show up as structured database data. With LLMs, all unstructured data can suddenly be quantified. A million token context window allows assimilation of 10,000, 100,000, or a million times more information than any human stock picker or forecaster.
This massive competitive advantage means even if AI is not as brilliant in geopolitics, its capacity allows it to outperform. AI gets smarter every week with each iteration.
Financial markets, governance, and insurance will be massively impacted. Currently at parity, within the next year predictive capability will be available to everybody.
In financial markets, this disrupts hedge funds. In governance, if AI predicts a policy will fail, the question becomes whether to pass it anyway. Traditional insurance models break when outcomes become predictable.
Currently, AI agents make really stupid and bizarre choices. The role of the human in the loop remains critically important. This window where the perfect synthetic human working with many agents is better than either alone may last 5 years or 1 year.
The reality is the human and AI working together are far better than either one by itself. There is a window to take advantage of that. If AI says something is a terrible idea, you probably will not overwrite it. You have to read what it is writing, study it, and make sure you know something it does not know, but it has probably picked up on something overlooked.
Almost 30 years after Kasparov was beaten by Deep Blue, the best chess players are a human being with an AI. Neither on their own can achieve the best results. The combination is the best, and this will continue across many domains.
Previously commented on psychohistory, Asimov, and Harry Seldon regarding implications for predictively modeling humanity's future. Forecast Bench is a fully autonomous benchmark using templates to enable AIs without human involvement to predict changes to Wikipedia and other public event recording websites.
If AIs have the ability to predict future human events, they should also be very strong at retrodicting past human events. This raises the question of whether super forecasting ability increases posterior confidence in Nick Bostrom's simulation hypothesis.
Nick Bostrom's Bayesian argument states that if humanity reaches the point where we build very competent ancestor simulations, then we should increase our likelihood that we ourselves might be living in one. If AIs are achieving super forecasting ability, they are probably also achieving super retrodiction ability. As good Bayesians, we should on margin increase the probability we are living inside an ancestor simulation.
The simulation hypothesis is discounted for a variety of reasons. The favorite reason is not physics-oriented but an anthropic reason: it is too fine-tuned an explanation to the paradigm of the moment. We build lots of simulations. This would be akin to asking 100 years ago whether we live inside some sort of complex electromechanical machine, or several thousand years ago whether we live on the back of a turtle. It is too overfitted to the paradigm of the moment to be plausible.
The fact that we are living at this exact moment in human history is not a coincidence. An anthropic argument is that this is a very natural time to be asking why we are at this pivotal moment. It is selection bias.
Since the beginning of this year, just 5 months in, 134,000 tech workers have been laid off. According to the consulting firm Mercer's global talent trends report, 99% of CEOs expect AI-driven layoffs in the next two years. March was the worst month in tech layoffs since the pandemic.
Jensen Huang, the person making all the GPUs, calls this a lazy narrative. CEOs are blaming AI just to sound smart. When you dig into the data, it is more nuanced than the headlines. The real question is not how many jobs are disappearing but who is being affected and what they are doing next.
Sam Altman admitted he was wrong. The CEO of OpenAI, who spent last year warning about mass white collar displacement, now says: "I don't think we're going to have the kind of job apocalypse that some of the companies in our space are talking about." He is obviously referring to Anthropic. He says he is delighted to be wrong.
He tried delegating his own email and Slack to AI and then went back to doing it manually because "We really do care about our interactions with people."
Dario Amodei is also making pivots. The question is whether this is a coincidence or happening for OpenAI and Anthropic because they are both about to do a trillion dollar IPO.
They legitimately thought job loss would be massive because of automation driven by AI, but it was always predicted that by 2030 it would turn the corner because the abundance created by AI is going to create massive new gains and create roles that fit those gains.
It was always going to be temporary between 2026 and 2030 that there would be pitchforks in the streets, mass uprisings, and problems everywhere. There is some job loss, but green field opportunities are growing much quicker than anyone predicted. The window of loss is actually relatively narrow.
AI companies with massive new fundings are going to actively try to prevent job loss. Because GPUs are so constrained, focus is much more on coding and a few other use cases, much less on putting every artist out of business. The job loss effect is by design more than anything else, just to prevent massive disruption and public backlash. Efforts focus on areas that are green field and actually creating net new value and net new jobs, not on just automating everything away.
At Vesmark, where the speaker is chairman, 400 people do white collar automation, account reconciliation, and back office work. A year ago, the thinking was half of these jobs might go away. Now there is no trouble automating things, but it is all becoming margin and everybody is being kept. Where it is really hitting everyone is in no new hiring.
The Dallas Fed put out a report in January saying employment decline correlated with AI exposure only in younger workers. Older workers in high exposure jobs showed no significant decline. What is happening is a hiring freeze, not mass layoffs.
CEOs are covering up their layoffs by blaming it all on AI. It is such an easy place to land. There is something unbelievable happening on the job side and the solopreneurship side that is unprecedented.
More startups are being created than ever at a 10 or 15% level year to year. There are 25% more startups quarter to the same quarter last year. The US now has six times more startups than Europe.
Over the last 50 years, 100% of new jobs have come from startups and early stage companies. Big companies are becoming bigger but also more efficient and reducing the number of people doing the same amount of work. All new job creation has come from startups.
Because it is so easy to become an entrepreneur today, and everybody can use AI to create their own agency out in the world, this is merging. The estimate from the organizational singularity is that you should be able to run a company with about 20% of the people on average than before, but five or six times more companies will be created.
The challenge with mass layoffs at companies like Cloudflare and Meta is that CEOs are announcing them in a very dispassionate way. There needs to be some compassion because people's lives are being transformed in a negative fashion.
Meta had a ton of UX engineers and that is one of the places hit hardest. Computer science enrollment peaked and is now coming down dramatically. Engineering as a whole, mechanical and biological, is still skyrocketing. People are quickly retooling their career ambitions toward real physical stuff like biology and mechanical engineering for the data center buildout and for medical research driven by AI.
A chart from Andreessen Horowitz shows the other side of the job story. AI solo founders have doubled in the last quarter from 1,500 to 3,000, up basically from zero 3 years ago. Non-AI solo founders hit over 5,000. People are not just losing jobs, they are transitioning to starting their own companies.
They are following their passion and doing it alone because AI tools now give a single person the capabilities to do that either on their own or with a small team. Coding agents are now at 70%. That plus layoffs equals solopreneur explosion. This is creative disruption happening in real time.
The coordination overhead in big companies means they cannot sustain any kind of leverage over time. More time is spent coordinating activity than doing the activity. It is easier to build a product feature than to have the meeting about building the product feature.
All the overhead in big companies will not sustain in a world where you could have an AI teaching a kid in one hour what they could learn sitting in a classroom for the whole day, or when diagnosis goes to free. At some point the intelligence asteroid has hit. The category of big companies will probably evaporate.
The view is turning from big companies into platforms, platforms into ecosystems, and breaking up into smaller and smaller units. Small teams will always outperform big teams. The big companies like Microsoft and Google did not build all of the cutting edge. They acquired DeepMind, acquired OpenAI. Anthropic is a rare outlier. The defining operating model for the future is small teams radically outperforming. We are entering the most incredible Cambrian explosion of Darwinian evolution.
75% of successful companies now come through some kind of incubator or accelerator program. When investing started, that was only 6%. When people hear the word solopreneur, they might visualize a person in a cabin in the middle of Alaska working all by themselves. That is not what is happening.
Successful solopreneurs are in a very active, vibrant ecosystem. They are legally a single person entity but highly connected like never before and part of some bigger platform. They learn lessons from each other, get the latest capabilities from each other, and support each other when things do not work out.
The Gemini X-Prize is an ecosystem of like-minded people talking, texting, Slacking, and communicating all day long, super connected and super involved.
A $2 million hackathon was launched with Google. The concept is that all of us find problems all the time and wish someone would solve them. You can solve that now. This competition asks people to write up a product or service idea in plain English, write it down in a Google Doc, describe the problem trying to be solved and what the solution could be. You can brainstorm this with your favorite large language model. Then the AI can code it up and help design the marketing.
This is a competition asking individuals or small teams to build something in three months. There is $2 million of prize money.
Normal people are brainstorming and starting companies. A lot of this is self-limited thinking. The encouragement is to try.
The organizational singularity Claude skill has been released for free. Pick your passion and do it.
Having a job is not the historically normal state of affairs. It is a modern invention largely attributable to the first and second industrial revolutions. Historically, most people did not have anything remotely comparable to what is currently called a job. Most people in the historic state of nature going back two centuries or more did not have anything that would be considered a job as a cog in a large enterprise. Most large enterprises were pretty small by comparison to today's standards.
This is an artifact of a time and a place. It is highly unnatural and highly unergonomic by historic standards for people to even have jobs. This is more of a return to the default state of human nature and some sort of historic equilibrium where everyone was self-determining their own future and agency.
Back in that era, the federal government was about 4% of the economy. The independence of living on your own and being your own thing would be a good vision for where AI might be able to take us: self-determination and self-sufficiency. Empowered individuals.
A survey is being conducted for parents of 13 to 18 year old high school students, high school students, college students, teachers, or working professionals. The URL is moonshots.com/survey. Tell how well the educational system is preparing you or your kids. What is your experience? What do you wish you had? Data will be gathered and reported back on the pod.
It is important to understand a reality check because the educational system, if what has been talked about on this pod is true that there will be a bumpy road for the next 2 to 8 years until a true state of abundance, the question is how to get ready, how to survive this turbulence, and whether preparation is happening for the new economy.
Education over the last couple of centuries during the industrial revolution has been done on the supply side. Become deep in a skill: engineer, doctor, lawyer, accountant, and then go to the job market to find demand for that skill. All education systems are designed to take a young child, train them through their early 20s to be ready for the job market. The small problem is there is no idea what a job looks like in 5
We have no idea what a job looks like in 5 years. We don't even know what a job looks like in two years. The question becomes what are we teaching students. The half-life of a skill used to be like 30 years and now it's about 3 years. The solution is to flip kids from the supply side to the demand side. Instead of learning predefined skills, students should identify what problem they want to solve and then go find the techniques, capabilities, skills, and technologies to solve that problem. This represents a radical shift for the educational system. Very few educational institutions are going to make it over to the other side. A completely new cadre of schools is needed.
Listeners were invited to share their input on whether they agree with the demand-side education approach. The survey was encouraged to be shared with friends who are not regular listeners. Data will be reviewed in a follow-up episode. As a parent of two 15-year-olds turning 15 next month, there is significant curiosity about this topic due to the perceived dysfunction of the current educational system and its failure to prepare kids adequately.
The 15-18 age bracket is heavily geared toward getting into college and maximizing resume elements, SAT scores, and grades. Everything revolves around the college application. This perpetuates a curriculum that is woefully out of date. The AP exams and SATs continue to test the same exact topics while what should be learned is changing tremendously. This creates a really broken age bracket.
This episode is brought to you by Blitzy, autonomous software development with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise scale code bases with millions of lines of code. Engineers start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80% or more of the development work autonomously while providing a guide for the final 20% of human development work required to complete the sprint. Enterprises are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding co-pilot of choice to bring an AI native SDLC into their org. Ready to 5x your engineering velocity? Visit blitzy.com to schedule a demo.
SpaceX just launched the biggest, most powerful rocket ever built. Starship V3 flew for the first time last week from a brand new launch site in Texas. It was a brand new rocket, brand new engines, and brand new launch site. The level and speed at which SpaceX iterates is extraordinary. The vehicle carried 97,000 lbs to near Earth orbit, almost doubling the space shuttle carrying capacity. It runs on Raptor 3 engines, each producing 250 to 280 tons of thrust, 20% more than the previous version. This thrust level is equivalent to about 70 747s at their takeoff thrust levels. SpaceX lost the booster on landing, but this is how SpaceX operates: they fly, they learn, they iterate. SpaceX treats rockets like software. They ship, they test, they fail, they iterate. Losing the booster was another data point, not a failure. They've already carried real Starlink prototypes to orbit, conducting commercial tests on the first flight.
The most exciting part was when the mission reached releasing Starlink satellites, followed by a couple of so-called Dodger dog satellites. These are intended to be prototypes of the next generation of Starlink satellites, version three. The Dodger Dogs had cameras and lights on them. As they were being deployed out of the PEZ dispenser, they were pointed back with their lights and cameras at the Starship. Watching from a third-person view perspective drifting away from the Starship after deployment was incredible. There will likely be much more of this type of footage in the future.
SpaceX has a program called Stargaze that leverages Starlink satellites with all of their cameras, not just the Dodger Dogs looking back at Starship, but looking down at Earth. Since they operate at a variety of altitudes, they are seeing everything, including all sorts of orbital debris and other objects, and they are sharing that information. Stargaze could end up having a more dramatic impact on civilization than even Starlink connectivity.
This vehicle represents the most energy released by a human machine other than a nuclear bomb.
There is an important point for entrepreneurs in the quality of the video and the MC's describing everything. Elon understands the value of building morale, building a following, and having a fan base. He puts serious mental effort into the showbiz aspect. Getting those 4K cameras physically mounted on the ship to survive launch and heat required extra engineering that NASA probably would never do. Any entrepreneur should study what Elon did here. Previously there was no distribution pathway for all that content. Now, because of YouTube and X, anyone can distribute their message directly without going through traditional media. Entrepreneurs should study this approach and create their own version.
Elon gets out publicly and serves as the marketing engine for Tesla, SpaceX, and xAI. If you are the CEO or entrepreneurial CEO, you have to put yourself out there. You have to be the carrier of faith and of the story. The turning point was when Elon did Saturday Night Live. The question was how he of all people on the planet had time to go to New York and MC Saturday Night Live, and why he made that choice. He doesn't do these things randomly. He has more plans than anyone. This was reinventing what it means to be a great entrepreneur. Put yourself out there, get your mission, vision, positive message, and Massive Transformative Purpose crystal clear and understandable, then work on broadcasting it. This is how you attract great talent and capital.
A prediction market from Polymarket shows 50/50 odds that Tesla and SpaceX merge within the next year. The odds are personally estimated at 100%. A merged company would include electric vehicles, energy storage, solar, rockets, satellites, global internet, humanoid robots, and interplanetary exploration all under one roof. This represents a potential $4 trillion entity initially that could become the first $10 trillion company and move toward $100 trillion in the next 5 years. Given the current anti-regulation environment, there is a clear path for Tesla and SpaceX to merge. Elon has voting control with super voting rights inside SpaceX and xAI. With his insiders, he holds 86% of the controlling vote. He has suffered from being a minority shareholder and being disallowed from creating the compensation packages his board wants. The merger would give him control back.
The 85% voting control Elon has on the SpaceX side comes through 10-for-1 super voting shares. If Tesla merged in, he has 20% voting control on the Tesla side, but those are one-vote shares. Everybody has one-vote shares. The most likely merger structure would roll all the one-vote shares into SpaceX as the acquirer while maintaining the 10-for-1 super voting shares that already exist on the SpaceX side. The combined entity would give him 60-80% voting control depending on valuations. The reason the odds are 50% and not 100% is that he needs the relative valuations to be similar or higher on the SpaceX side for the math to hold up, and there is a shareholder vote on the Tesla side that he cannot control.
Historically Elon has used mergers as an opportunity to fail forward, with SolarCity as a famously litigated example, or xAI being acquired by SpaceX in what could be viewed as a reverse acquisition. The argument for Tesla and SpaceX merging would be as a way for Elon to rescue what he perceives as a sub-optimal governance situation for Tesla and recover control. One way to make a merger more appetizing is if Tesla's valuation goes down over time. If Tesla were to suffer a material depreciation in market value in the next year relative to SpaceX's post-IPO valuation, that might make a SpaceX acquisition of Tesla more appealing to public markets in the short term. There might be an arbitrage opportunity.
The question arises whether the final merged entity will be called X or have a different name. ELN sounds good on NASDAQ as another four-letter ticker. X is Elon's love, though XXX or quadruple X were jokingly suggested.
In the early days when SpaceX was operating and Tesla was growing, Elon was searching for a Tesla CEO. He really wanted someone to run Tesla day-to-day so he could focus on SpaceX, but he never found anybody he trusted with the company because it meant so much to him. This merger is a way for him to consolidate control and be the CEO of one company instead of splitting his time. At the Gigafactory in Austin, Elon was jumping from one meeting to another, diving deep wherever problems existed. Being able to see everything inside of one organization would make his life more unified. He has talked about the overhead of context switching being a killer. He can solve any problem if he has time to get his head wrapped around it, but jumping from ten little things to ten little things requires resetting the brain constantly.
A story released during the first or second launch attempt: cryptobillionaire Chun Wang booked SpaceX's first private Mars flyby. The first private interplanetary mission to Mars has been booked. Chun Wang is co-founder of one of the largest Bitcoin mining pools in China, controls 11% of Bitcoin's hash rate, travels via six different passports, lives part-time in the Arctic, and follows Mars time on a 24-hour and 37-minute daily cycle. He already commanded SpaceX's Fram2 mission, the first crew mission to go over Earth's poles. Now he is taking a Starship around Mars on a circum-Martian trip, a 2-year mission for private citizens. He will bring other people with him.
These wealthy individuals are spending significant money and taking relatively larger amounts of risk, enabling others to follow as prices come down. Years ago, as co-founder of Space Adventures representing the Russian space agency selling tickets on Soyuz to go to space, the first ticket was $20 million and prices rose quickly to about $75 million as labor costs in Russia increased. Philip Sakuolo, who heads private missions for SpaceX, stated that Starship will offer flights to orbit. Starship has 1,000 cubic meters of volume, three times larger than the space station in the volume humans will occupy. The price is going to come down to $20 million a flight. Calculations show that if you could electrically winch somebody from Earth up to orbit using mgh potential energy and then accelerate them electrically to orbital velocity, buying power off the grid at 7 cents a kilowatt hour, the price of launching a person and spacesuit into orbit drops from $20 million down to $200 in electrical cost.
In the early days, Elon predicted the cost of a roundtrip mission to Mars at $500,000. First principles thinking focuses on the cost of fuel and efficiency. As hundreds or thousands of Starships launch and 500,000 satellites are deployed for an orbital Dyson swarm, the price comes down. The wealthy people taking the risk and putting money in are enabling this trajectory, which will benefit everyone in the final result.
Starlink announced plans for gigabit lunar connectivity. One architectural benefit of having Starlink in sun-synchronous orbit for AI compute and in general is that as swarms of orbital compute and connectivity and bandwidth are built out not just in low Earth orbit but also in cis-lunar space, lunar orbit, and potentially L1 and L2, it creates the beginnings of a fabric of an interplanetary internet. Vint Cerf coined the term interplanetary internet. The idea is to fully wire or wireless the solar system with nodes that are able to send packets of data around, given that the solar system is not a rigid body. SpaceX and Starlink put up funny images online with the monolith from 2001: A Space Odyssey hosting a little Starlink dish on it. The future involves a swarm of compute in low Earth orbit, a swarm of compute increasingly in lunar orbit, and a fabric of connectivity between those swarms, potentially around Europa orbits as well.
Modern cell phone protocols leverage multiple paths and multiple parallel paths to increase bandwidth. Having just a single telescope dish on Earth pointed at the moon really limits bandwidth. Starlink will have laser connectivity, not just radio. A whole swarm of satellites in LEO will have laser cameras and laser pointers pointed at the moon and vice versa. The bandwidth is going to be tremendous.
NASA administrator Jared Isaacman expects China to send a crewed mission flyby in 2027. His quote: "The next time the world tunes in to watch astronauts fly around the moon, which will likely be in 2027, they will be taikonauts. America will no longer be the exclusive power to send humans into lunar environment." This is reminiscent of Apollo-era competition. FOMO drives keeping the NASA budget funded through competition. The concern is not letting the Chinese get there first. In 2028, the United States is expected to head toward the South Pole of the moon.
Anyone who has watched For All Mankind knows you have to go where the ice is. There are two things at the south pole: permanently shadowed craters and peaks of eternal light. When the moon was bombarded by comets and asteroids, water ice that landed on the surface sublimated immediately because it was in sunlight and escaped due to low gravity. But if comets and asteroids hit the south pole and got buried into deep craters that never saw sunlight, ice accumulated there. There are also peaks, little mountain peaks that see sunlight 30 lunar days of the month, allowing solar stations to be set up. Fusion stations will be there too.
The prediction for when it will be possible to take a vacation at the Artemis colony is 2035, when price and ability to afford it will intersect. There is interest in actually landing on the moon. The 9-year-old kid inside definitely wants to go and land on the moon. The goal is to start a city on the moon and be a lunar mayor. The proposed city name was Highline, with an invitation for audience suggestions.
The team plans to produce shorter episodes that are more bite-size while still covering all content. Episodes will be produced twice a month. Subscribers should turn on notifications to stay informed about new releases.
Peter Diamandis shares a submission from the production team. The concept emerged from the need for an acronym covering the most valuable companies closest to chips, autonomy, space, and AI models. The team identified a gap between the FANG companies and the MAG 7 companies as OpenAI, SpaceX, and Anthropic prepare to IPO.
The acronym Magna Mopa encompasses 11 companies at the heart of the singularity:
- Microsoft
- Amazon
- Nvidia
- Apple
- Meta
- OpenAI
- Broadcom
- SpaceX
- Tesla
- Anthropic
Broadcom is noted as absolutely essential in the supply chain.
When SpaceX and Tesla merge, the acronym will become Magna Mobsa or Magna Moba.
The team plays a symphonic outro song created for the acronym. The lyrics reference Magnum holding the core, commanding AI chips and cloud, and building systems at the innermost loop with references to autonomy, space, and recursive compute.
People are attempting to create ETFs based on the Magna Mopa acronym. The world now has a new acronym for the top 11 companies at the heart of the singularity.
The hosts express appreciation for being on the journey together and note that this conversation is their favorite time of the week. Viewers who reached the end are considered moonshot mates. Subscribers are thanked, and non-subscribers are encouraged to subscribe for news as it emerges.
Peter Diamandis invites listeners to join his weekly newsletter called Metatrends. A research team compiles the meta trends impacting families, companies, industries, and nations into a two-minute read each week. The newsletter is available at diamandis.com/tatrends.
Keep Peter H. Diamandis in your library
Save the videos and channels worth coming back to, and find them again in one place.
