Podcast Summary: Raoul Pal: The Journeyman
Episode: The AI Intelligence Tsunami Is Here
Date: February 26, 2026
Host: Raoul Pal
Guest: Ahmad Mostak
Overview
This episode features a deep and fast-paced conversation between macro/crypto thinker Raoul Pal and AI pioneer Ahmad Mostak. They delve into the explosive growth of artificial intelligence—especially autonomous AI agents—and unpack how these technologies are about to upend economic models, investment opportunities, and the very fabric of society. Together, they chart where value will accrue in this exponentially advancing landscape and debate the risks, rewards, and existential questions ahead.
Key Discussion Points & Insights
1. The AI Takeoff: Speed and Ubiquity
- AI adoption is happening at a rate beyond historic parallels—AI agents are rapidly proliferating, with everyone from tech leaders to regular users running dozens to potentially thousands for all manner of tasks.
- "It’s the great takeoff this year. Isn’t it? Really is, really is." (Ahmad, 04:05)
- New breakthroughs—such as transformer models etched onto silicon—are slashing compute costs and boosting speed to levels unimaginable before (e.g., real-time 15,000 tokens/sec).
- “There’s this wave of intelligence just about to hit us, it’ll cost basically nothing.” (Ahmad, 04:40)
2. Agents as Economic Game-Changers
- AI agents are set to become near-ubiquitous, radically increasing the intelligence and velocity of decision-making in the economy.
- As usability friction falls to zero, the ability to tap intelligence becomes universal and democratized, eroding traditional sources of profit.
- “All of those frictions are going to disappear. You know what frictions are? Frictions of profitability.” (Ahmad, 05:21)
- The paradigm shifts from monetary velocity and supply to "intelligence velocity and intelligence supply."
- “It's an intelligence velocity, intelligence supply... they removed a barrier and idiot tax to intelligence flow.” (Ahmad, 05:26)
3. Cost Collapse & Value Accrual
- Both intelligence (AI output) and energy (thanks to renewables like solar) are experiencing double-exponential changes: costs are collapsing while output skyrockets.
- “Solar electricity is lower than a dollar per watt now, right?... a dollar for all your words for an average person.” (Ahmad, 08:58)
- “This is not AGI from an economic perspective, it's ACI—actually competent intelligence.” (Ahmad, 09:18)
- Margins in software, tech, and services will compress drastically, since AI will destroy friction and rent-seeking in most digital value-chains.
- “Where does profit accrue in this?... it just becomes comes down to compute costs and energy costs and that's it.” (Raoul, 06:32)
4. Beyond Human Language—Maximizing AI Efficiency
- Current AI models rely on human language as an interface, but this will fade; AIs will soon communicate at machine speed and eschew language for direct binary or latent space protocols.
- “What's code? Code is a translation. It's a way for us to talk to machines. But machines can talk to machines without code.” (Ahmad, 12:38)
- Recursive, self-improving AI—where agents prompt and build each other—will drive even faster innovation:
- “It's prompted by humans. But Claude built Claude, they're all building each other now.” (Raoul, 13:18)
5. Memory, Context, and Continuous Learning
- Overcoming memory limitations (i.e., persistent contextual awareness) is the next step for AI to compound intelligence and autonomy.
- “Persistent memory is the big one... when they have more persistent memory, even context windows right now are still a bit small.” (Raoul, 16:49)
- “With Claude, when you run out of space, it compactifies... Continuous learning, I think, is almost cracked because you have so much hardware available.” (Ahmad, 17:12)
- Agents will soon maintain perfect, persistent knowledge of user context and tasks, making them even more powerful.
6. The Rise of Personal AI Agents
- Tools like Claude Bot and OpenClaw exploded in adoption partly due to their seamless, always-on integration with personal services (like WhatsApp), reducing friction for users.
- “You could send it any arbitrarily long piece of task... It became your coordinating agent, it became your chief of staff, your best personal assistant.” (Ahmad, 20:46)
- The importance of data sovereignty and security rises, as people look for localized, personal, ring-fenced agents—despite companies like Apple dragging their feet.
- “What don't you have right now? One click installer on your iOS and your MacBook for a claudebot/openclaw type experience fully locked down.” (Ahmad, 25:27)
7. Death of Switching Costs & SaaS Moats
- AI agents will automate and eliminate switching costs between digital ecosystems and services, threatening traditional SaaS and platform moats.
- “The moats are just going to disappear overnight… counter example to people who say that SaaS companies are safe.” (Ahmad, 27:48)
8. Data, Learning, and Synthetic Knowledge
- Quality of data now matters more than quantity; there’s already enough human knowledge to train AIs capable of mastering most tasks.
- “The data's been getting better and better... now we have higher quality data, we know better what to do with the data so we can train on it more optimally.” (Ahmad, 29:31)
- Synthetic data and few-shot learning mean AIs will improve quickly, even in new environments.
9. Simulated Worlds and Real-Time Content
- Generative models (like Google Genie, Sea Dance) can now produce interactive 3D games, movies, and experiences in real-time, tailored to a user’s preferences—with far-reaching implications for media, entertainment, and identity.
- “You can create video games that you can play real time... what comes first, GTA 6 or generative GTA 6?” (Raoul, 34:07)
- Persistent virtual universes (simulations) will fragment shared cultural realities.
10. Benchmark Saturation & AI Capability Ceilings
- AI models are breaking and saturating every benchmark rapidly—coding, superforecasting, creative tasks, and research. Soon, open-source models will rival the best closed systems.
- “By the end of the year, we've saturated all benchmarks... Google's latest deep thing model now is the fifth best coder in the world.” (Ahmad, 37:34)
11. Robots and Physical Automation
- AGI brains will be poured into a coming wave of general-purpose robots, turbocharging physical productivity. The bottleneck: hardware supply, not AI capability.
- “Once you put an AGI brain into a Android robot, humanoid robot, and that's it, it's game over.” (Raoul, 45:36)
- Robotics will disrupt labor economics, but ramp-up is constrained by manufacturing bottlenecks—true ‘Westworld’ scenarios are a decade away.
12. Finance, Crypto, and Economic Models
- AI agents will transact, pay, and optimize via blockchains and smart contracts—significantly expanding the total addressable market for crypto rails, stablecoins, and payment networks.
- “Agents will drive more of the economy than humans. Who is your end customer in this new world? Your new customer is the agent.” (Ahmad, 51:14)
- Raoul proposes the new GDP model: “GDP equals humans plus robots plus AI plus debt plus energy intensity plus compute efficiency.” (53:10)
- Reputation, insurance, lending, and asset management will all be redefined by agent economies.
- “There's an entire world of financialization to come for agents which will drive relative GDP growth.” (Ahmad, 54:14)
- Agent-to-agent marketplaces and payment rails are already underway; interoperability will become frictionless and increasingly non-human.
13. Investable Opportunities and Coming Upheaval
- Traditional tech, SaaS, and knowledge work moats are collapsing—value will accrue to the “agent economy,” platforms built for agent clients (like Stripe, payment rails, and robust API services).
- “Companies like Stripe... you look for the public equivalence of who basically is best set up for agents as customers. Traditional SaaS, that sell off continues, they're going to get absolutely gutted.” (Ahmad, 58:25)
- Hardware and data center buildout remains a short-term play, but will lose momentum as efficiencies improve and the digital workforce takes over.
- Mass layoffs and market disruption loom as “the Great Disruption”—with record business failures and reduced roles for humans in the digital value chain.
- “If a company fires you... they can take all of the emails you’ve written and documents you’ve created and create a digital clone of you that costs a thousand bucks, that's going to do your job better. That's the first time in history we've ever seen that.” (Ahmad, 59:40)
14. The Battle for the Primary Agent
- Big Tech (OpenAI, Google, Elon Musk with X, Apple) is racing to make their AI agent the primary, closest entity to every user, knowing utility and habit will define the winners in the agentic economy.
- “That agent is the one that matters because its utility function will determine everything... so much of the economy will be that.” (Ahmad, 63:33)
- Affinity, memory, and relationship with your agent (its ‘soul.md’ file) becomes the new switching cost.
Notable Quotes & Moments
- On paradigm shift:
“AI has been the fastest adoption of any technology... It’s Reid’s Law, which is Metcalfe’s Law squared. We’ve never seen anything like the change of technology we’re living through. And it’s only going to get worse.”
—Raoul Pal [02:42] - On friction removal:
"What Claw and Claude bot kind of did is they removed a barrier and idiot tax to intelligence flow... Profits are based on what Elon Musk calls the idiot tax."
—Ahmad Mostak [05:26] - On economics of AI:
"The cost to build a Hollywood level movie in a few years will be like less than a thousand bucks."
—Ahmad Mostak [36:50] - On digital clones:
"They can take all of the emails you've written and documents you've created and create a digital clone of you that costs a thousand bucks—that’s going to do your job better. That's the first time in history we've ever seen that."
—Ahmad Mostak [59:40] - On economic risk:
"We've never seen an exponential that fast."
—Ahmad Mostak [61:47] - On utility and agent loyalty:
“Its personality is in a file called Soul md. Soul Markdown. You can say export your personality, take it from one service provider to another.”
—Ahmad Mostak [66:58]
Key Timestamps
- AI Takeoff & Agent Explosion: [04:05]–[05:26]
- Collapse of Economic Friction: [05:26]–[08:57]
- Technological Costs & Universal Code: [08:21]–[11:03]
- Moving Beyond Human Language: [11:03]–[13:42]
- Self-improving AI Agents: [13:10]–[14:48]
- Persistent Memory & Learning: [16:49]–[17:12]
- Rise of Personal Agents (Claude Bot, OpenClaw): [20:26]–[25:27]
- Death of SaaS Moats: [27:01]–[28:00]
- Data, Synthetic Knowledge, and AI Training: [29:31]–[31:43]
- Generative Worlds, Genie, and Simulations: [33:31]–[35:54]
- Benchmark Saturation & AI Advances: [37:34]–[41:27]
- Robotics and Physical Automation: [41:27]–[45:36]
- Finance, Crypto, and Agentic Economy: [50:01]–[56:14]
- Investable Opportunities & Upheaval: [57:43]–[59:40]
- Battle for the Primary Agent Relationship: [63:33]–[67:09]
Conclusion
Raoul and Ahmad’s dialogue is equal parts thrilling and sobering, underscoring that we live amid an intelligence supernova where nearly all friction—economic, technical, even cognitive—is vanishing. The new “agent economy” will rewire everything: who earns profit, how wealth is built, what jobs survive, and how value is exchanged between machines on our behalf. As AI agents become everyone’s closest companion and the agent-to-agent economy explodes, incumbents will be swept aside unless they build for this coming tsunami. For investors and technologists, everything is up for grabs—and the exponential is accelerating.
Final words:
"On the other side, it's total destruction or complete abundance. One of the two."
—Ahmad Mostak [68:16]
