The Long View
Weekend deep dives into long-term trends and strategic themes.
The Grid Said No: How AI Compute Is Being Forced to Become a Power Company
FERC's bring-your-own-power order, Meta's 1.6GW Crusoe deal, and Switzerland's nuclear reversal mark the moment AI infrastructure stopped waiting for the grid and started building around it.
The Water Question: AI's Quiet Reckoning With the Hydrological Constraint
Power gets the headlines, but the buildout's binding constraint over the next decade may be a resource the industry has barely learned to measure.
When the Load Becomes the Grid's Problem: AI's Voltage Reckoning
Texas's voltage ride-through failures expose a structural shift: AI data centers are no longer just power consumers — they are now grid-stability liabilities, and that changes everything about the buildout.
The Memory Wall Is Now the Cost Wall
Why HBM, DRAM, and the storage stack — not GPUs — are the binding constraint on the AI buildout, and what unbinds them.
The Social License to Build: Why the Datacenter Boom Hits a Zoning Wall
The AI buildout's binding constraint is shifting from silicon and substations to the consent of the towns being asked to host it.
The Long View: Open-Source AI and the Battle Over the Stack
Meta gives away its models. Startups build on them. Incumbents feel the pressure. Is this a sustainable equilibrium — or the early phase of something bigger?
The Long View: The End of the SaaS Era
The software model that defined two decades of technology — selling seats to dashboards — is being structurally undermined by AI, and the consequences will reshape the entire industry.
The Long View: The Trust Problem at the Heart of Artificial Intelligence
AI's biggest obstacle isn't capability — it's the fact that we have no reliable way to know when to believe what it tells us, and building that infrastructure matters more than building better models.
The Long View: The Quiet Revolution in Scientific AI
While public attention fixates on chatbots and image generators, AI is transforming the practice of science itself — from protein folding to weather prediction to mathematical proof — in ways that may prove to be the most consequential AI story of the decade.
The Long View: Digital Abundance and the Scarcity Problem
When AI makes the production of knowledge, content, and analysis nearly free, the economics of every information industry inverts — and the scarce resources that remain will define who captures value.
The Long View: The Infrastructure Supercycle
AI is driving the largest infrastructure buildout since the internet era — a multi-trillion-dollar supercycle in data centers, power generation, networking, and semiconductors with parallels to the railroad and telecom booms.
The Long View: When Models Become Commodities
GPT-4-level intelligence is rapidly approaching commodity pricing — and the consequences for where value accrues in the AI industry will reshape the entire technology landscape.
The Long View: The Last Human Advantage
An honest assessment of what humans still do better than AI — and why the list is shorter, more specific, and more fragile than most people want to believe.
The Long View: The Geopolitics of Compute
AI chips have become strategic resources as consequential as oil was in the twentieth century, and the scramble to control their production, distribution, and deployment is reshaping the global balance of power.
The Long View: The Attention Economy Is Dead. What Comes Next?
For two decades, attention was the scarce resource that organized the digital economy -- now AI intermediaries filter, summarize, and act on information without human attention, and the entire economic logic of the internet is shifting beneath our feet.
The Long View: Software Is Eating Software
Marc Andreessen said software would eat the world, and it did -- now AI-powered software is eating software itself, and the consequences for the people who build, maintain, and sell software are more profound than any previous platform shift.
The Long View: 2025, the Year AI Satisficed
Looking back at 2025 as the year AI crossed from impressive demos to good-enough-for-real-work, and why the gap between frontier capabilities and what enterprises actually need matters more than the frontier itself.