All In On Merge

If all you want is to maximize alpha, regardless of field, you should invest in materials.

Games and railroads

Within the field, though, to get your return in 10-12 years, I’d direct you to The Merge, where human and machine meet. The future of the individual is distributed. It is digital and it is analog. We’ll back up our minds and become masters of our physical form. We’ll travel between worlds on beams of light, in more ways than one. We’ll build Matrioshka brains and design alien species. But none of this happens without merging.

When it comes to longevity, the timelines to achieving radical life extension don’t fit within the life of a fund. So we’re looking for what I refer to as railroads plays. The foundational technologies upon which empires are built. SpaceX has funded the mission to Mars with railroads to earth orbit. And build upon that, railroads to the internet. Soon, AI. And next, lunar industries. It is a long and winding road. But it is the road that takes you where you want to go.

Here is the road to radical life extension and a future where humanity plays a part.

Merge is inevitable in the long tail (though this long tail is not guaranteed, more on this later). Lucky for us, Merge is also a game that has recently become available. It also precedes massively profitable follow-ons including virtual realities, upload, and the substrate of the new species. Other approaches that aim to preserve, explant, or destroy the brain are irrelevant in the long run. They may be profitable for some time, but this is far from a guarantee. The industry for horse doctors, horseshoes, saddles, was vibrant once. That is, until cars proliferated. And industries move even faster today. If it were up to me, I'd place an early bet on Ford. Merge, however, is inevitable should humanity prevail, makes massive money in the next decade, and is only the first rung on a very tall ladder.

Ride the waves

The game of Merge is available because it rides on current powerful waves of innovation: BCI, compute, and AI.

BCI

Brain interfaces are being derisked and solved right now. With massive funding and rapidly accelerating progress in the US, China, and beyond. As Neuralink, Merge Labs, and others grow, the pool of people capable of building these technologies grows too. Once Bell Labs and Texas Instruments built the first transistors, the pool of people who could make transistors exploded.

Compute

Running external cortices is compute hungry. As demand for AI grows, ignoring any local maxima, the cost of compute plummets. People like to reference the computational efficiency of the human brain, using on only 20 watts. Human compute was also once the most efficient at calculating orbital mechanics, and NASA filled rooms with people for this purpose. It is inevitable that the efficiency of nonbiological compute exceeds that of the human brain, though external cortices become profitable much before this inflection point anyways.

Digital neural networks

As AIs approach the capability for Integrated Recursive Self Improvement, designing vertically integrated, application-specific digital neural networks, the tooling and capabilities also benefit the design of external digital cortices.

Scope and adoption

When people think about cloud cortices and merging with AI, they sometimes leap to whole brain interfaces and less invasive BCIs. Neither of these are needed to build a massively profitable Merge product.

The functional plasticity of our brains allows us interface with vast external neural networks over only a small number of neuron-linked channels. The two systems don’t need full access to each others source in order to communicate over the API. Even in today’s early systems, external digital neural networks enable people to control computers and robots, play video games, and speak through microphones using only small number of channels.

While the surgery today is long and human-driven, the scaled system will be entirely automated and take only a few minutes from walking in the door, scanning your credit card, and leaving with your new chrome. This capability is not as far off as it may seem and is a tractable engineering problem.

It may be hard to imagine folks being willing to undergo this level of surgery on a large scale and there is some truth to this. But you don’t need most people to get the early versions in order to build a massively profitable business. The medical applications alone are enough to justify the venture. Once it becomes advantageous in business to get implanted, people will do it. From senior executives to entry level employees, your chip gives you an edge. It provides economic mobility. Businesses and governments will deploy vast swarms of edgerunners to manage their massive AI systems, steer them, and defend against cyber threats. And as product features expand from enterprise to consumer, individuals will increasingly adopt the technology for its benefits to communication, learning, safety, and entertainment. Of course, it will also become less invasive over time.

Conclusion

The forces of innovation favor this path and the core risks are being actively diminished by broader economic forces. There are still significant engineering problems to solve on several fronts. But the cost structures are coming into place, supply chains are being stood up, and a rich talent pool is growing.