This is my private space, where I share thoughts and projects I am working on.
You can find me on X and LinkedIn. My DMs are open.
GITHUB
Vision
Tech is the most intellectually interesting industry in the world. It is the one category that rewards
continuous curiosity and optimism at an outsized scale.
Tech is a cyclical industry, driven by macroeconomic forces and technological disruption. Usually, highs are
pushed by a singular catalyst.
I believe right now might be the most interesting and dynamic environment ever for tech. This is because we are
not witnessing one singular technological disruption, but rather three major technologies that will work
together to reboot the entire tech stack and completely redefine the digital economy.
Value
Layer: Blockchains
The financial operating system gets rebooted on blockchains as the largest open and interoperable financial
stack. Every asset will live on this shared stack, which unlocks a global network with more human and AI
participants. In turn, new markets are created that unlock new opportunities. The flywheel of new assets,
new markets, and new users has a compounding effect that expands digital GDP massively.
We have witnessed the first waves with tokens, stablecoins, and money markets, but sooner or later, every
asset will live on blockchain rails.
Interface
Layer: Augmented Reality
Augmented Reality and spatial computing will fundamentally change how we interact with the digital world.
Interfaces are no longer restricted to a limited two-dimensional screen canvas. They expand to spaces and
positions. They can attach to information that is anchored in the physical space. They can add another layer
or alter the physical world.
The best practices for interfaces have to be rediscovered for every application, whether B2B SaaS or a
social network. What does a PowerPoint look like in this new paradigm? This unlocks opportunities for new
entrants that disrupt incumbents that adopt this new computing platform more slowly.
Productivity
Layer: AI
Throughout history, work has continuously shifted from laborious, nuisance activities to more abstract,
conceptual tasks. The first shift occurred during the Industrial Revolution, where machines increasingly
outsourced physical labor.
Artificial Intelligence is the ultimate evolution of this. Work in the future will look much more like the
work of CEOs of larger corporations: designing systems, orchestrating work, and delegating tasks.
Every productive task and action will be fulfilled by AI agents, whether in the digital (agents) or physical
(robots) world. What remains are humans who provide the vector.
Compounding Effects
These technologies do not work in isolation, but compound with each other and unlock new use cases. For
instance:
CE 1: Agents moving value on-chain
CE 2: Data collection through human POV
CE 3: AI agents that guide you through the physical world
...
CE n: Digital collectibles gain value as AR allows humans to wear them in the physical world.
VC.ai
In the year 2026, VC cannot be done the same way it was in 2020. Agentic loops have become so powerful that they
can replace junior VCs in many of the critical functions: sourcing, thesis building, and even first-pass analysis. Eventually, more
and more AUM will be directly or indirectly managed by agents. I think being at the forefront here can produce
significant alpha.
This agentic VC setup, at its current level, already completely changes how I research topics and
gain an understanding of tech trends. This agentic system is 24/7 on the hunt for new alpha.
Architecture
Fundamentally, this agentic system works on four layers: Data, Thesis, Evidence, and Opportunities. These
layers are connected and have feedback loops that can refine each other. All of this works continuously and
24/7.
The agentic system is powered by Obsidian and a Hermes agent harness that runs several subagents.
Data
The agentic system continuously monitors high-value data sources across the web, including social
media, academic research, and high-signal podcasts and YouTube shows. It studies articles, opinions, and
scientific papers. Initially seeded by a human-curated list, it quickly identifies its own data sources and
expands the network of data sources to monitor.
Human-AI loops strengthen this mechanism to ensure a high-dimensional dataset that remains
high-signal.
It scrapes this dataset continuously and constantly filters for high-quality data points and humans.
Thesis
Theses are the fundamental primitives of the agentic VC setup. They form before categories are formed and
chase narratives that might be 3-5 years from maturity. Unlike static memos that only capture a moment in
time, they are living and fluid concepts that are constantly maintained and refined by the agentic
system.
Each thesis is continuously refined through supporting arguments, counter indicators, and signals to watch in
order to provide an expert-level current read and conviction level on each thesis. Just like with human
analysts, seniors can provide feedback to steer and refine the direction of research the agentic system is
taking.
The agentic VC uses these lessons to refine its methodology and suggest its own theses.
Evidence
Evidence points are the main data points for agentic theses. They are early, unstructured, anecdotal proof points
that suggest the validity of a thesis. As such, their predictive power develops through their cumulative
mass.
The agentic system filters the data stream and matches evidence points to each thesis. Evidence can occur in
various formats: an academic paper that potentially removes a technological bottleneck, builders
experimenting with an interesting MVP, high-profile investors writing about points touching the thesis, or
new startups forming around a thesis. They can appear in different formats and across different
channels.
Individual evidence points can vary in weight and can even be counterpoints that question or reshape the
thesis.
Theses are the fundamental intellectual primitives. Explore a number of theses I had. Each thesis is
automatically created and maintained by my VC.ai system. Scroll through the theses and click on
one to explore the evidence VC.ai
has gathered behind each thesis.
