TECH013: Monthly Tech Round-up - Davos WEF, Claude Cowork, Macrohard, w/ Seb Bunney (Tech Podcast)

Episode Summary

Executive Summary: Preston and Seb explore how AI is rapidly transforming software, knowledge work, and infrastructure, arguing that computation is becoming a strategic asset in the AI economy. They connect this to Bitcoin’s scarcity thesis, tokenization limits, longevity breakthroughs, and the possibility of personalized apps, one-person companies, and massively more efficient AI-driven systems.

Main Topics: AI tools replacing traditional software workflows (Priority: 5/5): The hosts describe how Claude, Cursor, and related tools can generate apps, redesign UX, organize notes, and automate reporting far faster than conventional software workflows. Computation as the new form of wealth (Priority: 5/5): They argue that in an AI economy, access to compute and energy will matter more and more, and that scarcity-based assets like Bitcoin are better stores of value than issuable tokens or fiat. AI agents and persistent autonomous work (Priority: 4/5): The discussion highlights 'Ralph' and related agentic systems that can persist across failures, chain context, and complete expensive tasks overnight with little human oversight. Infrastructure efficiency and AI-optimized protocols (Priority: 4/5): They examine Tesla’s proposed transmission protocol and broader moves toward hardware-level efficiency gains that reduce latency and energy waste in massive data-center workloads. Tokenization vs true blockchain-native ownership (Priority: 4/5): The NYSE tokenization announcement is treated skeptically: they argue that unless legal ownership and issuance are natively on-chain, tokenization is still dependent on analog legal systems. Longevity, stem cells, and medical disruption (Priority: 3/5): They discuss Elon’s comments on aging and a stem-cell diabetes breakthrough as examples of medicine becoming an engineering problem, with major implications for pharma and healthcare. Geopolitical and institutional shift at Davos (Priority: 3/5): The conversation frames Davos as evidence that elites are increasingly accepting AI, Bitcoin, and a less centralized world order, while some older narratives (like climate-first rhetoric) fade.

Key Arguments: AI coding and agentic tools are now good enough to build polished functional apps in minutes, making many SaaS products vulnerable. Personalized, user-specific applications will replace generic software because AI can tailor workflows to individual needs at low cost. Computation will become a scarce and valuable resource in the AI economy, analogous to money, because it enables intelligence and production. Bitcoin stands apart from tokenized securities and fiat because it is scarce, issuerless, and resistant to clawbacks or state interference. AI systems may increasingly choose Bitcoin over fiat/stablecoins because they can reason about scarcity and sovereignty more objectively than humans. Persistent AI agents can accumulate context through repeated failures and eventually complete expensive or complex tasks with minimal human supervision. A lot of current 'blockchain' innovation in traditional finance is really just improved enterprise databases with tokenized wrappers, not true on-chain ownership. AI-driven biology and stem-cell therapies could shift medicine from symptom management to root-cause repair, potentially extending lifespan substantially.

Data Points: Citations in The Hidden Cost of Money: 400–500 - Seb said his book included roughly 400 to 500 citations. Time to write The Hidden Cost of Money: About 2 weeks - Seb said the pre-AI book was written in about two weeks using extensive notes. Apple notes migration time: About 5 minutes - Cursor organized 300 Apple notes into linked, categorized notes in minutes. SaaS underperformance vs Nasdaq 100: ~40% gap - Referenced a chart showing the Morgan Stanley SaaS index lagging the Nasdaq 100 by about 40% since a year ago. Ralph API cost: $297 - An example was given of a $50,000 contract completed using Ralph-like AI agents for about $297 in API costs. Tesla transmission efficiency gain: 5%–15% - The proposed protocol was described as potentially saving 5% to 15% in energy costs depending on scale and reliability. Transmission efficiency improvement: 100x to 1000x - The hardware/hardwired approach was described as potentially 100 to 1000 times more efficient than conventional TCP/IP in certain data-center contexts. Glycemic range improvement: 43% to 96% - In the stem-cell diabetes trial, time in target glycemic range improved from 43% baseline to 96% by month four. Post-transplant insulin independence: 75 days - The patient achieved sustained insulin independence starting 75 days after transplantation. Long-term glycemic control: >98% - At one year, the patient reportedly maintained over 98% time in target glycemic range. AI training infrastructure: 1 gigawatt - A commenter cited xAI as having the only one-gigawatt data center in the world at that time. Office and model development horizon: 3, 5, and 10 years - Elon’s forecasting onstage at WEF suggested near-term predictability, but major uncertainty by 10 years.

Pivotal Quotes: "We are literally going through the singularity right now." — Preston: He frames the pace of AI advancement as a singularity-like acceleration during the opening discussion. "AI is just eating all of these SaaS products." — Preston: Used while discussing Claude, Cursor, and the collapse of generic software moats. "I think that ultimately what we are seeing is our ability to support cellular regrowth, stem cells growth, you name it." — Seb: Seb summarizes why the diabetes stem-cell result matters for future medicine and longevity.

Implications: AI is compressing software cycles, lowering the cost of building, and shifting value toward compute, energy, and scarce assets like Bitcoin. Finance, medicine, and labor markets may all be reshaped by agentic systems and biotech breakthroughs.