The thesis of the permanent underclass has two core knock on effects that are largely under allocated.  A free market economy is largely a competence hierarchy composed of individuals and their ability to generate value for others. Think of it as each human having a score that is some combination of ambition (hours worked) x genetic predisposition (ie IQ) x leverage (technology or resources).  While neither IQ nor ambition scale, technology does. A 100 IQ working 30 hours will overtime beat a 110 IQ working 40 hours given that the 100 IQ uses technology in a way to either double his output, or make him produce the results of a 120 IQ.  Now naturally in a free market economy, this experiment will lead to top percentile IQs that work every waking hour, trying to outcompete each other.  The permanent underclass is the population that falls below the line where neither them without AI, nor them with AI can compete with a standalone agent. As of today, this line does not really exist yet, as virtually all AI products, still require a human manager.  Thus the North Star for both those wishing to escape that ever rising base line and those wishing to compete at the very top will try to find ways to increase their score.  Where can people get advantages from?  Enhancing their own compute (brain power) both technologically (BCI/AR) and organically (biotech). Equally we may see physical power enhancements (mechanically and biologically), where the cost of enhancement is a fraction of a comparable humanoid.  Now assuming all the above is true, two knock on effects become unavoidable. Enhancing yourself will be a core tenant of the upper class, while finding experiences to fill the void left by exiting the productive class will be a core tenant of the underclass.

Most breakout tech today sounded like sci-fi ten years ago. From humanoids, to LLMs to longevity it’s all happening. So what then feels like sci-fi today that is ripe for breakthrough in the coming decade? One of those spaces we are focusing on is brain computer interfaces. From headphones that monitor your mood state like Neurable, to small devices that can “read your thoughts” by measuring the electromagnetic frequencies emitted by your muscles, all the way to full-on brain chips like Paradromics and Neuralink that connect directly to your neurons and can transfer information bit by bit from your brain to machines.  While BCI seems far out, we believe it’s one of the technological endgames for computing and human machine interaction. Even the most invasive companies requiring brain surgery have entered human clinical trials and are going commercial by 2030.  I’d take the bet that a decade from now, by 2036, at least 30% of your inputs (keyboard, mouse, voice) will be intent and/or BCI driven.

Where you invest may be just as important as what you invest in. While the SF/Bay Area only represents ~2.25% of the US population, it generates ~42% of all US unicorn startups. Naturally a debate emerges. If you have such a statistical edge investing in the honey pot of innovation, why would you touch any other geography/demographic when the hit rate appears to be much worse?  Markets tend to trend towards efficiency though, so while the sum of winners is larger, so is the density of deals as more founders flock to SF than any other city. Equally, when there’s more bidders, something has to give and you end up paying a premium for founders just because they’re based in SF.  To me the edge lies in just backing the best founders no matter where and being willing to pay market. That means in SF you will pay premium for a rockstar team. It also means you will be able to strike some incredible deals in overlooked places like Eastern Europe or Canada. And while the underdog may have a lower statistical hit rate, your pay-off may be 10x when you swap the premium for a discount.  Best of both worlds? Truffle-hunt outlier founders in overlooked geographies, back them early, then help them relocate to the Bay Area.

Cold email isn’t dead. If you do it right. If you ask most VCs they have done little to no deals that came in cold. Whether you love it or hate it, the game favors social signaling. That being said one of my most interesting deals actually came in cold last year. Can you make telepathy happen with BCI? Wild enough concept, and most wearables guys said “probably not legit”… in VC probably isn’t good enough so I booked a flight to NYC to check it out. Seeing the lab and early demo first hand, and spending a day with the founder made it clear there was something real here, so we kept tracking the deal. After four months of status updates, a lead came in and within a week or two, the company had an all star round, that was virtually entirely sourced cold, and then oversubscribed to a higher valuation. If you’re a VC: Stay curious If you’re a founder: Don’t listen to the gatekeepers and shoot your shots.

The progress of the last three years in AI will fundamentally disrupt and accelerate seemingly unrelated industries in the coming years. Robots and glasses are upgraded by receiving a new brain. World models can accelerate scientific research and bring us closer to living longer and better. Skill barriers across art and code have been obliterated, with creativity, ambition (and tokens, lol) being the primary limiting factors. Ironically, these appear to be the most exciting years in human history to be a venture capitalist, yet they coincide with the most underfunded VC environment in early-stage frontier tech outside the core foundation models. When the tide shifts, fortunes will be made by those betting on and building the future of how we work, play, and live, from BCI and biotech to XR and robotics.

What non-consensus sector are you most bullish on right now? VC is a power-law game. The highest-upside bets usually begin non-consensus.

On our latest pod episode, @Saxonbaum @FelixOHartmann discuss investing in frontier tech. They cover VR adoption cycles, the smart glasses opportunity, sub-vocal communication tech, why underfunded categories produce strong founders, and more. Listen: hubs.la/Q03-_4wB0

Long BCI

How we've built positions before consensus arrives: →3 BCI deals in 12 months →Power users before investors (sequenced embryos) →Spotting the 40-gram smart glasses breakthrough 24 months early Watch @FelixOHartmann's full convo with @Saxonbaum ⬇️ youtube.com/watch?v=4kaW4k0x…

❤️❤️❤️❤️❤️❤️

Paradromics just received FDA approval to start human clinical trials of its high-bandwidth brain chips. From solving speech disorders, mood disorders, and blindness to helping humans become superhuman, Paradromics is truly a generational company. Proud to be an investor!

Everyone’s talking about AI as a productivity tool. But what happens when it changes the entire economy? 🎥 @FelixOHartmann on the economic transformation powered by AI.

AI isn’t just a tool — it’s becoming the new game engine. In our AI in Gaming Q3 2025 report, @derzic_daniel explains how AI is reshaping every part of development — from testing to monetization — and where the real value is emerging.

The next decade won’t just expand economies — it’ll expand human potential. AI founders amplify intelligence. Longevity founders amplify life. 🎥 @FelixOHartmann explains how they're building what humanity can become: youtube.com/watch?v=oWzSXv93…

VR builds worlds. AR builds intelligence. The future of AR isn’t escapism, it’s empowerment. When your glasses translate, guide, and inform in real time, that’s not sci-fi — it’s the new OS for life. 🎧 Watch @FelixOHartmann on The Caplander Podcast: youtube.com/watch?v=oWzSXv93…

Oriane is THE AI video search engine 🚀 and we're proud to be backing their journey to build a platform for brands, media, and creators who need to find videos based on the content → not just metadata.