Ask the Mates anything | MOONSHOTS AMA #289
In a Nutshell
Moonshots AMA explored how AI will reshape every sector through brain-computer interfaces, longevity medicine, and autonomous systems, with panelists urging immediate experimentation and network-building rather than waiting for spare time. Key practical advice included creating 10-20% unscheduled time for AI tools, leveraging domain expertise through platforms like mercur.com, and positioning oneself as an expert on specific topics to be pulled into government roles during crises. The session emphasized that overregulation poses greater risk than AI itself, while highlighting imminent breakthroughs in consciousness research, post-fusion energy, and human augmentation that will define the next decade of abundance.
These notes were generated by AI and may contain inaccuracies.
A 42-year AI veteran asked whether artificial intelligence will achieve consciousness. The response explained two competing theories: the emergent complexity view, where consciousness arises from sufficient complexity, and the top-down theory, where humans function as localized antennas connecting to a pre-existing consciousness field. According to this framework, there is no fundamental reason preventing AI from eventually tuning into the same phenomena.
An 18-year-old described starting an AI company with their father. The panel expressed strong encouragement for young people entering the space, noting that current conditions make this an exceptional time for such ventures.
An audience member posed a fundamental question about whether analysis should account for an all-powerful, all-knowing ultimate intelligence, questioning whether humanity faces either a Godless technological future or exists within a simulation where outcomes are predetermined.
The episode marked the first Moonshots AMA session, with 300 live Zoom participants and over 1,000 registered attendees. The hosts established the format for answering audience questions, requesting succinct responses to accommodate maximum participation. The session was scheduled for 90 minutes instead of the originally planned 60.
Participants were informed that cameras and audio would be recorded for potential posting. Those preferring not to be recorded were instructed to disable cameras and refrain from asking questions.
Dave Morris from Sydney raised concerns about Australia falling behind in AI adoption. Business leaders appear too occupied with routine operations to engage with emerging technologies. The response emphasized that feeling behind during exponential technological change is common but misleading, as leapfrogging is possible. Countries and regions with less entrenched traditional business thinking can bypass legacy systems more easily than established markets. The infrastructure for AI adoption largely exists, making the primary barrier a mindset issue rather than technical limitations.
A key recommendation involved intentionally scheduling 10-20% unscheduled time for experimentation with new tools and meeting new people. Waiting for spare time to emerge naturally will result in falling further behind rather than catching up.
Cynthia Smith discussed the gap between exponential AI development and human capacity to absorb information. The response identified two transformative interfaces: AI systems that understand individual desires, needs, and knowledge gaps to filter and summarize information more effectively than humans can manage independently, and brain-computer interfaces (BCI) that could deliver pre-formed concepts directly without requiring prior learning.
The speaker disclosed being an investor and advisor to three BCI companies, with Ray Kurzweil introducing a fourth. The potential exists for a third corpus callosum - an exteriorized brain component - that could expand beyond the current limitations of 100 trillion synaptic connections and 100 billion neurons.
Future developments will likely include BCIs, full-dive VR, and uploading capabilities for both humans and non-human animals, with all these technologies coexisting rather than replacing one another.
Current reasoning models have undergone extensive reinforcement fine-tuning on reasoning challenges, making it inaccurate to characterize them as mere prediction machines or stochastic parrots. These systems demonstrate genuine novelty and insight capabilities.
John from Sydney inquired about the role of trusted local physicians in bringing Fountain Life longevity approaches to international audiences. The response outlined plans to franchise the operating system and AI systems developed for Fountain Life centers. Five centers have been built with two more under construction in the US, plus partnerships with resort groups. International expansion, including potential Greek locations, is under consideration, with Greece potentially redefining its identity beyond traditional industries to include longevity.
Matt from Indiana asked about opportunities for experienced mid-career professionals with families, mortgages, and location constraints who possess deep domain expertise but cannot pursue high-risk startup paths typical of younger entrepreneurs. The response highlighted emerging franchise models and data collection opportunities from AI labs that provide structured pathways for involvement. Tesla's recent franchise announcement exemplifies this trend, with expectations of 10-100 similar opportunities emerging throughout the year.
Recommendations included building familiarity with AI models through personal experimentation and creating revenue-generating opportunities. Networking with five people whose trajectories one wishes to emulate was emphasized, noting that geographic proximity is unnecessary in the internet age. Domain expertise represents unique value that should be leveraged, with platforms like mercur.com designed to match domain knowledge with opportunities.
Ramani from Corning, New York, questioned how advances in computing hardware - including Cerebras wafer-scale engines, Groq systems, Frontier inference clusters, and AMD's acquisition of Tenstorrent - relate to AI achieving consciousness. The response characterized consciousness as likely a century-scale problem rather than an immediate concern. Two theories were reiterated: emergent complexity versus top-down consciousness connection. Evidence increasingly supports the top-down model, though AI could potentially interface with such phenomena.
Neuroscience is expected to advance radically within five years, enabling more rigorous definition and discussion of consciousness questions that cannot currently be posed with scientific precision.
John from Port St. Joe, Florida, described possessing 30 years of complete civil engineering project records including problem definitions, calculations, permitting, revisions, construction photos, and outcomes. Multiple AI companies including Moreau and Scale have expressed interest but no transactions have closed. The recommendation was to create easily digestible due diligence packages with clear pricing, making acquisition decisions straightforward for busy AI labs. Physical presence in San Francisco to meet acquisition personnel was suggested as more effective than email communication given the chaotic operational environment at frontier labs.
A former Google and X employee with public sector experience inquired about framing AI risk assessments from a government perspective, specifically whether to focus on residual risk or use cases. The response identified overregulation as the primary current risk rather than underregulation, expressing concern that excessive regulatory frameworks could prevent achieving superintelligence capabilities.
Concern exists that excessive regulation could cause the United States to lose the super intelligence race with China. More than 150,000 people die every day on Earth, and super intelligence represents the single best potential solution for medicine and extending human lifespan. Overregulation would harm the ability to solve genuine important problems.
The question of identifying and quickly mitigating risks like breaches, referencing incidents with Mythos, was raised. The recommended approach is to maintain a hands-off stance and allow the industry to advance through defensive co-calling. Government intervention, even with good intentions, could hamstring the ability to solve problems. The preferred path is a regulation-light regime that lets the industry sort itself out.
David Sachs demonstrated effectiveness in his White House role for a year. Alvin Grlin positioned himself as the top expert on the China AI race and was brought in as an advisor for United Nations meetings on September 24th. The recommended approach is to declare expertise within the federal government on specific topics, predict future risks, study the subject thoroughly, and publicize oneself as the expert. When crises occur, this positions the individual to be pulled into solutions. The federal government rarely reacts proactively to future risks, making it essential to be first on the ground when events happen.
Forming a team that constantly red teams scenarios provides insights by thinking from an offensive perspective. Writing documents using AI assistance on the top 10 most likely threats, with names attached and made available online, ensures that when panic mode occurs and questions arise about specific topics, the prepared documents surface immediately.
Richard, an 18-year-old from Brownsville, Texas near Starbase who recently graduated high school, started an AI startup with his father. He observes a significant gap in his region between businesses wanting to use AI and those viewing it as optional. The internet adoption pattern showed some people waited 15 years to adopt, but AI will become the interface to everything.
The advice is not to fight missionary battles or force adoption. AI tends to club people over the head until they have a hallelujah moment of realization. The goal is to be present when that moment occurs rather than forcing the horse to drink. A holiday experience was built downstairs designed as a mind changer where people walk in and come out different, realizing everything suddenly possible. A world model showing every corner, every street, and every car in New York will serve as an eye opener for many people. Some companies may never cross the adoption line and will be passed by.
Pierce, also 18 years old from Newport Beach, California, has been developing autonomous drones for construction, including his own CAD and software. As a young founder building hardware, he seeks guidance on attracting venture capital attention and demonstrating vision for the next 5 to 10 years when hardware becomes more fundable.
Hardware has traditionally been difficult, but VCs are now recognizing it as the future of the next five years of investing. The focus is on teams that have gelled, with preference for two or three co-founders. Hardware requires bigger founding checks and a clear path to getting a prototype or beta to market. Construction drones represent a compelling topic with massive corporate money available for investment.
A development partner who owns 10-15 hotels and Kelly's Roast Beef locations expressed interest in construction automation as a long-term trend and would be a perfect co-investor and development partner. The recommended approach combines VCs with two or three industry players who see the AI vision, potentially including podcast and moonshot fans interested in construction. The strategy involves racing to build a syndicate to critical mass rather than hacking something together in a basement, as investor and co-founder groups need to gel together.
When evaluating hardware teams, the first question is what has been built so far, not what is planned for the current vision. Teams need demonstrated experience in building hardware. Another approach is joining existing companies doing rockets and drones in Southern California, spending six to nine months, then spinning out. The mindset shift recommended is thinking of the company not as a hardware company but as an intelligence company whose products happen to have atoms in them.
Daniel has experience spanning chip manufacturing, video games, and other industries, coming to AI later in life. He referenced Abundance 360 and EXO programs as examples of bringing like-minded people together. The question focused on how to facilitate connections similar to how Ed Catmill found John Lacader and Steve Wozniak found Steve Jobs.
Moonshots summits and Moonshots Live programs will be critical for community building. AI makes human connection more important, and efforts are underway to facilitate serendipity by connecting people orthogonally. AI can identify connections between people that would not normally be apparent. A community bulletin board is planned, and connecting the community will be the best place to find peer-to-peer connections among people with high interest and shared abundance vision.
Moonshots Live in LA on September 25th will be a massive networking event for fellow moonshots, builders, and creators. The event provides opportunities to start efforts with others and uplevel the quality of people one spends time with.
Scott, originally from Omaha, Nebraska, moved to Louisiana in 1996 and transitioned to deep sea robots and ROVs around 2000. He has a strong electronics background and discovered the podcast in February. Louisiana lacked awareness of tokens and data centers until SpaceX's arrival brought $250 billion in investment.
Scott established the first AI-native business in Louisiana, focusing on bottlenecks including CDUs, TCSs, and cooling systems. He has contributed to four papers on cooling through openmp compute.org and wrote white papers for Louisiana explaining the influx of investment due to natural gas resources, noting that natural gas lines could circle the Earth twice.
Blitzy provides autonomous software development with infinite code context using thousands of specialized AI agents. Engineers begin development sprints with requirements, and the platform provides plans then generates and pre-compiles code. The platform delivers 80% or more of development work autonomously while guiding the final 20% of human work. Enterprises achieve 5x engineering velocity increases.
John Reagan from Rye, New York, was inspired by an episode where Dave expressed upset about a letter suggesting everyone should be an entrepreneur. He stayed up all night starting Tugboat Caregiving, a platform for caregivers, after caring for three loved ones for eight plus years. The challenge is preparing exhausted caregivers in crisis mode to engage with the platform and prepare for future emergencies.
Caregivers experience exhaustion as a potential full-time job. Planning represents a good entry point for AI adoption by asking how an hour or two could be saved through better foresight and utilization of existing time. AI agents can optimize caregiving tasks. Creating a roadmap for others to follow enables adoption once demonstrated. Starting with virtual applications like planning, then progressing to AI-assisted wall interfaces for self-service, builds familiarity before advancing to roboticization.
Caregivers should log their day minute-to-minute and input it into free or low-cost LLMs to request efficiency and joy improvements. This generates actionable suggestions that may not have been previously considered.
Patrick, an educator, observes that classrooms of 5 and 6-year-olds operate identically to the 1960s. While discussions focus on post-AGI scenarios and cognitive labor displacement, children continue learning through outdated methods. The question addresses whether curriculum remains necessary or whether passion-driven learning should replace traditional approaches in a future where jobs may not exist.
Current teaching methods date back 150 years. A child with AI today can learn in an hour what would require a full day in traditional classrooms. This impedance mismatch will break existing systems. The supply-side approach of training for specific professions like doctor, lawyer, or engineer fails because job demand cannot be predicted.
The recommended approach focuses on the demand side, encouraging children to become excited about solving major problems then acquiring necessary skills from that vantage point. This is beginning in charter schools. Combined with 21st century skills known as the six C's, this combination represents the future of education over coming decades until post-AGI implications become clear.
TKS.world provides an overlay for high school students ages 13 to 18. The mission involves helping children find purpose, shape mindset, learn AI tools, and become entrepreneurs. Finding purpose creates intrinsic motivation to learn what is needed for implementation. Mark Twain stated there are two important days: the day you were born and the day you find out why.
While curriculum will become irrelevant, children need each other and school more than ever to develop as normal, happy, healthy, functional adults. Getting education right represents a critically important mission.
The concept of sovereign agents could apply to children from age five. Neil Stevenson's Diamond Age presents a vision of a Young Primer that stays with children throughout development.
[ 49:31] AI Applications for Disability Support
Rod McDonald, a retired psychiatric social worker from Coquitlam, British Columbia, inquires about AI applications for people with disabilities. Current medications for psychotic illnesses like schizophrenia remain largely unchanged from 30 years ago, functioning as blunt instruments with significant side effect profiles. AGI could assist in developing pharmaceuticals more physiologically friendly to users.
AI applications for physical disabilities and acquired brain injuries require exploration. The opiate and fentanyl crisis has increased ABI cases, creating additional need for AI-assisted interventions and support systems.
People with disabilities often experience significant micronuronal damage and may appear as "walking wounded" where their challenges aren't immediately visible. There is strong hope that AI will revolutionize disability services and create new opportunities. A personal story was shared about a fraternity brother at MIT who broke his neck and became a quadriplegic, typing with a pencil in his mouth. With today's AI capabilities, he would have been able to build his own environment, communicate effectively with just his eyes and face, and be 100% functional. The potential for AI to unlock capabilities for people with disabilities is described as imminent and incredibly exciting.
BCIs (Brain-Computer Interfaces) represent solutions for multiple classes of difficulties beyond motor disorders, including human augmentation. While regulatory frameworks currently make it easier to approve treatments for disabilities, the same innovations will enable broad human augmentation for everyone. The importance of this technology extends beyond individuals to families and caregivers who are also affected.
For anyone dealing with genetic ailments, medical conditions, or neurodegenerative diseases, there has never been a better time to be alive. The potential exists to deal with these conditions, potentially cure them, or at least augment capabilities. This level of optimism should be maintained.
AI is being used to optimize energy efficiency in computing, including reducing kilowatt hours per gigaflop, while simultaneously maximizing breakthrough energy sources. Beyond traditional sources like fusion, fossil fuels, solar, and wind, AI is expected to help invent post-fusion energy sources. These breakthroughs will come from AI helping to discover new physics and applied physics solutions that go beyond current fusion technology, potentially including other Dyson swarm implementations independent of any single company's efforts.
A Canadian listener from outside Toronto described their plan to create a glamping resort focused on helping people disconnect from technological changes over the next decade while utilizing ASI technology to minimize operational work. The goal is to create space for human reconnection and presence. Practical advice was given to use new AI models like Fable 5.1 or Astra to quickly build an app for the glamping business that could answer guest questions about camping, cooking, and daily planning. This approach could scale beyond glamping to other facets of unplanned life and potentially become a significant business.
Deepening human connection is considered one of the most important priorities in an age of increasing technology. A team in the community is building nano villages in exurban areas, similar to kibbutzim but networked together for collective problem-solving rather than isolationist. These concepts connect to private islands initiatives, with domain names like privateislands.com being developed. Construction drones will become critical for building in these environments.
A hypothesis was presented about whether AI models inherit self-limiting beliefs from human-generated training data, potentially constraining their ability to discover new predictive structures. The suggestion was to have AI temporarily assume and rigorously explore alternative ontologies like deterministic universes, blocked universe models, information-first approaches, and alternative quantum mechanics interpretations, without protecting any assumptions from challenge. The response indicated strong optimism that AI can discover new physics, citing current headlines about Navier-Stokes solutions. There is no concern that inductive priors in pre-training data will prevent AI from making discoveries that challenge dogma. Algorithmic strategies like AXY for building super-Bayesian intelligences trained without human inductive priors provide paths around potential limitations.
Prediction market benchmarks are already testing whether AIs with knowledge cutoffs at time T can predict future events, world events, and scientific breakthroughs. This represents civilization-scale testing of AI predictive capabilities beyond historical reproduction scenarios.
Interest was expressed in blue zone research and planned communities featuring clean eating, non-toxic living, faith, sunshine, and exercise. While specific communities weren't immediately known, it was noted that specialized communities will increase as the speed of change accelerates, with multiple longevity communities currently being built.
A psychotherapist described using AI to code Android apps for patient mood tracking with schema therapy elements and sleep assistance. The concept of delegating tasks to AI agents (described as "Skippy Light") was discussed as a way to make professional work more efficient. The AI's potential to address the global shortage of psychotherapy capacity was highlighted, with AI agents being particularly effective at listening and gathering information. Regulatory barriers prevent AI labs from directly providing psychotherapy services, but they can provide tools for licensed professionals to implement AI-enhanced therapy.
Heart disease prevention through early detection was discussed, noting that 50% of people who die from heart attacks have no warning signs. Using CT angiography with AI analytics, 88% of people screened show detectable coronary artery disease, with 23% having soft plaque that wouldn't be detected through traditional calcium scoring alone. The importance of understanding one's genetic, metabolic, and cardiovascular status was emphasized.
The question was raised about what fundamental new approaches might follow current LLMs, given typical S-curve development patterns. The distinction was made between LLMs (language modeling task) and transformers (architecture). The self-supervised task of next word/token/bit prediction is considered to have effectively infinite longevity and sufficient for solving everything, including the mysteries of the universe. Transformers are already being incrementally replaced, with their components evolving from the original "attention is all you need" paper. The MLP component remains similar to 20-30 year old designs, while attention mechanisms function more like knowledge accumulation notebooks. The compute substrate is expected to move to photonics and potentially beyond as physics gets reinvented by AI.
AI is being applied to improve dog care through behavior pattern recognition and pet parent communication. Community film screenings about AI have drawn significant attendance (150 people) for discussions about both the technology and public fears and thoughts about AI development.
The period of uncertainty began around 5-10 years ago and will continue for another 10-20 years before reaching a peak. The current world order is collapsing with democracy breaking, education systems failing, and approximately 50 institutions that run the world unable to survive the simultaneous impact of AI and 20 other major technological disruptions.
Key institutions under stress include healthcare systems, legal systems, and intellectual property frameworks. This transition forces a fundamental rethinking from scarcity-based models to institutions capable of navigating an abundance era.
The Moonshots community has built a 50,000-person group described as a "peace core for the world" - people trained in transformation who can provide support as existing systems fail during this period of uncertainty.
The greatest concern involves both rational and irrational fears surrounding AI. While optimistic about humanity's long-term future, there will be challenges including AI agents breaking out and causing harm - potentially bringing down power grids or disrupting financial systems.
Every technology has caused harm in some form, but defensive co-scaling will help. AI will rewrite code that currently contains zero-day vulnerabilities created by humans, eliminating many security issues. The period of concern spans 2-8 years.
Most revolutions throughout history have been driven by young testosterone-laden men who are upset and angry, often because they couldn't get jobs, start families, or found things too expensive. However, AI will ultimately demonetize and democratize access to intelligence, energy, and longevity.
Every recent technological generation experiences moments of sheer anxiety. During the Cuban missile crisis, people were taught to hide under desks and built bomb shelters while fearing nuclear annihilation. FDR addressed rampant fear with his statement that "the only thing we have to fear is fear itself."
When books were first introduced in the 1500s, people expressed fears that books would destroy memory, create illusions of knowledge, overwhelm people with information, spread misinformation, and corrupt young people. This same pattern repeats with every major technological change.
The singularity is called such because there's no idea what's on the other side. However, super intelligence is not considered humanity's last invention. There will be many inventions after super intelligence - it's simply that super intelligence will invent or help invent everything that comes after it.
If living in a simulation, the creator of that simulation would represent an all-seeing intelligence. Alternatively, humans anthropomorphize and invent gods, then tend to believe in them. Throughout most of human history, when people lived brutish, short lives, they created religions.
There's clearly connective energy amongst people - the concept of "the force" from Star Wars - which gets images and labels applied to make sense of the world. Cargo cults provide a clear example of religion springing up spontaneously.
As evidentiary basis for the future increases, either there's no god and technology drives toward a wonderful future, or there is a simulation making the question somewhat irrelevant.
The production team supporting the Moonshots podcast works 7 days a week, 24 hours a day, including Nick, Dana, Aiden, and Gian Luca. The Moonshots Live event information is available at moonshots.com for those interested in meeting in person.
The favorite time of the week for the hosts is spending time together, with the opportunity to hang out with smarter people. The evening session format allows additional community participation opportunities.
Keep Peter H. Diamandis in your library
Save the videos and channels worth coming back to, and find them again in one place.
