3 future industry trends: space travel, life extension and AI

Originally published 13 March 2018. AI polished for a better reading experience: 11 August 2026.

If you have children, as I do, you probably wonder what kind of world they will inhabit when they grow up. Or you may have just finished high school and be choosing a university path. Prediction is hard, but we can still sketch plausible industry trends from current events and careful guesswork. For more than a decade I have followed forecasting and frontier technology—reading research and business updates at the edge of the field. Working at Amazon also keeps me in daily contact with new tech and product thinking. That mix is enough, I hope, to offer a short, reasoned list of industry trends for the next fifty years.

Space travel

This is likely the earliest of the three to mature at scale. We have been going to space since 1957, yet the basic model of travel changed little for decades: expendable rockets that deliver a payload and then become scrap. That approach wastes hardware, money, time (a new vehicle for every flight), skilled labour, and raw materials. Every launch destroys equipment worth millions of dollars—whether the rocket is Russian, European, or American. Prices differ; the destructive concept has been the same.

From a transport economics view, a lean operation keeps the price of the trip only modestly above the energy (in kWh) needed to give the object its kinetic motion. Maintenance, depreciation (how many launches one vehicle can fly), R&D, and fixed costs should be a small share of the bill compared with fuel.

For half a century that problem was barely addressed. Private firms are now racing to open space to a wider market with reusable rockets. SpaceX’s Falcon 9 and Falcon Heavy are the best known, but not alone: Jeff Bezos (my ultimate boss at the time of writing) has pursued reusability through Blue Origin. Private space is growing fast, the cost-cutting opportunity is large, and spectacular change in space travel looks likely. Over the coming years this industry may even rival the scale of disruption once associated with the internet.

Life extension

Life extension and ageing control are still early as an industry—and as a paradigm. As we understand the genome better, more personalised tools will appear to slow ageing and lengthen healthy life. Genetics differs from many earlier medical breakthroughs: it is not a single “one size fits all” pill for the whole planet. It works with individuality. Because we differ, single universal fixes are hard; genetic engineering points toward personal solutions.

When we will truly extend life is uncertain. Ageing is a complex, multi-factor problem. The practical answer will probably combine several strands: lab-grown organs, gene therapy, implants, and brain stimulation. It is reasonable not to expect a full halt to ageing much before mid-century (so plan on living toward around 2065—and on having the means then to access advanced care if indefinite life becomes possible). In the meantime the life-extension industry will grow on intermediate results that already improve healthspan.

Artificial intelligence (AI)

Before talking about AI, I want to separate it from “artificial humans.” I do not expect, in this century, machines that fully mimic human feeling and act with the full set of emotions—love and hate included. By AI I mean software that, with sparse and almost random input, can return a precise, complete solution to a hard problem. A few examples illustrate the level of ambition.

Opening a restaurant. You supply only basic data—how much capital you have, roughly where you want to operate—and the system replies whether the plan is feasible and, if so: which loan and bank fit (via public banking APIs); which premises are available now (live data from real-estate portals); layout, style, logo, and name (cuisine type, colours, local demographics, IoT footfall near the building, device types people carry, nearby reviews); suppliers, order cadence, and prices; full first-year financials with balances and cash flows; how many people to hire, when, and at what wages—and more.

HR systems. Continuous performance monitoring; when performance drops, proactive creation of a replacement job posting a month before a difficult exit so hiring has time; video interviews; talent spotting, development plans, and ad-hoc courses with little managerial input, driven by pre-set “instincts”; mass hiring and headhunting at scale.

Police investigation. From limited evidence, identify a likely guilty party and assemble a case strong enough to prevail in court.

That sophistication is hard to hand-code. Genuine machine learning must digest vast amounts of information already on the web and find useful applications inside the system. We are moving quickly, yet a truly universal, multifunctional, global, and multilingual AI will probably not fully arrive before the 2050s. As with gene engineering, the intermediate market will still be enormous for years.

So there is a map of what to watch. Other industries will grow—electric cars among them—but they may be less disruptive (and create fewer net-new roles) than the three above. Still other fields may appear that we barely name today (nanotech is an example); I will not claim certainty about those. Let’s see how the future unfolds. I think it can be bright and interesting.