Theory is cheap. Everyone has frameworks. The question is whether those frameworks actually work — whether they help you see opportunities before they become obvious, and hold conviction through the volatility that shakes out everyone else.
I've been investing for over thirty years, starting on Prodigy as a kid in the late '80s. In that time, I've watched multiple technological revolutions unfold — the internet, mobile, social, crypto, AI. Each time, the pattern was the same: early believers seemed crazy, then suddenly they were vindicated.
The Exponential Age framework isn't abstract theory. It's a distillation of what I've learned by actually deploying capital through these transitions. This article is about proof of work — not in the cryptographic sense, but in the sense of showing receipts.
Understanding exponential dynamics — accelerating returns, network effects, group formation — creates asymmetric advantage. It lets you recognize opportunities early and hold through the volatility that eliminates most participants.
The Pattern Recognition Engine
Before diving into specific calls, let me articulate the pattern I've learned to recognize. It shows up consistently across technological revolutions:
- 1 Early dismissal: Experts call it a toy, a fad, or "technically impossible"
- 2 Exponential improvement: Core metrics double at regular intervals (usually 12-24 months)
- 3 Network formation: Early adopters begin connecting, building community and infrastructure
- 4 Group fragmentation: Subcultures and specialized communities form (Reed's Law activates)
- 5 Mainstream moment: Suddenly everyone agrees it was "obvious" in hindsight
The opportunity window is between stages 2 and 4. By stage 5, the asymmetric returns are gone.
Case Studies
What I saw: In 2010-2011, Bitcoin had all the exponential signals. The technology was improving (hash rate doubling every few months). A network was forming (early forums, meetups, developer community). Most importantly, the underlying insight — digital scarcity backed by cryptographic proof — was genuinely novel.
What experts said: "Ponzi scheme." "No intrinsic value." "Governments will shut it down." "Quantum computers will break it." The dismissal was nearly universal among traditional finance.
The exponential thesis: If Bitcoin followed the Law of Accelerating Returns for network technologies, the hash rate would continue improving exponentially, making the network increasingly secure. If Metcalfe's Law applied, value would scale with active users squared. If Reed's Law applied, the community fragments (Bitcoin maximalists, traders, developers, regional groups) would create additional network value.
What happened: All three dynamics played out. Hash rate increased by a factor of 10 trillion from 2010 to 2024. Active addresses grew from thousands to hundreds of millions. The community fragmented into thousands of overlapping groups. Price followed.
The lesson: When experts dismiss a technology with exponential improvement curves and network effects, their linear extrapolation will be wrong. The question isn't whether they're wrong, but how wrong and how fast.
What I saw: In the early 2000s, Google was one of many search engines. But the underlying technology — PageRank — was fundamentally different. It was using network structure (links between pages) to determine relevance. This meant the more the web grew, the better Google got. It was a network effect in disguise.
What experts said: "Search is commoditized." "Yahoo and AltaVista are already dominant." "How will they ever make money?" The skepticism was reasonable by the standards of the time.
The exponential thesis: If PageRank worked, Google would get better as the web got bigger. The web was growing exponentially. Therefore Google's quality advantage would compound. Users would shift (Metcalfe's Law), and eventually advertisers would follow.
What happened: Google went from search engine to global infrastructure. The advertising model emerged. The network effects proved durable. The company became one of the most valuable in history.
The lesson: Technology that gets better as its input grows is technology that follows exponential dynamics. When the input is also growing exponentially (the internet itself), you get compound exponentials.
What I saw: Tesla wasn't primarily a car company — it was a bet on battery cost curves. Lithium-ion batteries were following exponential improvement patterns similar to semiconductors. Every doubling of cumulative production led to a roughly 20% cost reduction. If this continued, EVs would reach cost parity with ICE vehicles, then beat them.
What experts said: "EVs are toys for rich people." "The range is too limited." "Tesla will go bankrupt." "Traditional automakers will crush them when they decide to compete."
The exponential thesis: Battery cost was the key constraint. If it continued declining exponentially, the rest would follow. Moreover, Tesla was building a network of superchargers — infrastructure that would have Metcalfe's Law dynamics. And the owner community was forming groups (Tesla clubs, forums, referral programs) that would create Reed's Law stickiness.
What happened: Battery costs dropped faster than most predicted. The supercharger network became a moat. The owner community became evangelical. Traditional automakers are still struggling to compete.
The lesson: When a constraint is on an exponential improvement curve, the current state of the technology is irrelevant. The question is the trajectory. And when that technology enables network effects, the compound dynamic kicks in.
What I saw: Around 2015-2016, it became clear that deep learning was real. The improvement curves in image recognition, speech processing, and game playing were unmistakably exponential. The question was: who captures the value?
The thesis: AI training required parallel processing. GPUs were optimized for parallel processing. NVIDIA dominated GPUs. Moreover, NVIDIA was investing heavily in CUDA — the software ecosystem that made their hardware the default for AI researchers. This was a network effect: more developers on CUDA meant more libraries, which meant more developers.
What experts said: "GPUs are for gaming." "Intel/AMD will compete." "Custom AI chips will commoditize GPU advantage."
What happened: The AI explosion created insatiable demand for NVIDIA's hardware. The CUDA ecosystem became entrenched. Custom chips emerged but couldn't match the ecosystem advantage. NVIDIA became one of the most valuable companies in the world.
The lesson: In exponential technology shifts, the infrastructure layer often captures disproportionate value. And when that infrastructure layer has network effects (developer ecosystems, training on previous models), the advantage compounds.
The Failure Modes
I've been wrong too. The framework isn't infallible. Here's when it fails:
False Exponentials
Sometimes what looks like exponential improvement is actually S-curve saturation in disguise. The technology is improving, but hitting fundamental limits. 3D TV, VR in 2016, and early voice assistants all showed improvement that looked exponential but was actually approaching ceiling.
Network Effects That Don't Materialize
Some technologies have theoretical network effects that don't manifest in practice. The switching costs aren't high enough, or the network isn't sticky enough. Many social networks and crypto protocols fell into this trap.
Timing Errors
Even when the thesis is right, timing matters. Being early is the same as being wrong if you can't survive the drawdowns. I've held positions that ultimately proved correct but caused significant pain along the way.
Regulatory/Political Disruption
Exponential technologies can be disrupted by non-exponential forces. Government bans, regulatory frameworks, or geopolitical events can break the trajectory. This is a risk the framework doesn't fully account for.
The framework works often enough to generate significant returns, but not always. The key is position sizing that accounts for the times it fails, while capturing asymmetric upside when it succeeds.
Why This Matters For You
You might read these case studies and think: "Great, you got lucky a few times." Fair skepticism. But here's the point:
The pattern repeats. It's repeating right now.
AI agents on crypto rails. Decentralized compute networks. Synthetic media production. Agentic financial systems. These are all showing the same signals: exponential improvement curves, network formation, community fragmentation, expert dismissal.
The specific opportunities change. The pattern doesn't.
What I'm offering in this newsletter isn't "follow my trades." It's the framework that generates the trades — the pattern recognition engine that identifies exponential opportunities before they become consensus.
Building Your Own Proof of Work
Here's how to develop this capacity yourself:
1. Track Exponential Metrics
For technologies you're interested in, identify the key metric and track its progression. Is it improving at a consistent percentage rate? Is there a Moore's Law equivalent?
2. Watch for Network Formation
When early adopters start connecting — forming forums, Discord servers, Twitter communities — pay attention. This is Metcalfe's Law beginning to activate.
3. Monitor Group Fragmentation
When the community starts fragmenting into subgroups with distinct identities, Reed's Law is kicking in. This is usually a sign of durable network effects.
4. Discount Expert Dismissal
When established experts dismiss a technology with exponential improvement curves, treat their dismissal as information — but not the kind they think. They're extrapolating linearly. You shouldn't.
5. Size for Survival
Exponential bets are volatile. Size positions so you can survive the drawdowns that shake out weak hands. The framework works over time, not on every trade.
Conclusion: The Ongoing Proof
This framework has worked for me across multiple technological cycles. It doesn't guarantee returns — nothing does. But it provides a coherent way to think about technological change that has historically generated asymmetric opportunities.
The best proof of work is ongoing. The patterns I'm identifying now — the convergences I'm tracking, the exponentials I'm monitoring — will either validate or invalidate the framework in real time.
That's what this newsletter is: a live demonstration of the framework in action. Not hindsight analysis, but real-time application of exponential thinking to current opportunities.
If the framework works, you'll see it in the calls. If it doesn't, that will be visible too.
That's the only proof that matters.