Okay, Maybe 20 For Some Things
Business strategy needs to be short term, medium term, and long term.
Startups don’t care about the long term: for them, the medium term goal is usually to be acquired or even go public. But established companies need to think long term, because things can change and disrupt the company’s entire business.
Let’s think about ten years from now. That’s “long term” for most companies today. Long term might be a thousand years for an anthropologist, but for business people, ten years is about as far out as someone’s vision extends. After all, the average turnover for Fortune 2500 company CEOs is about five years.
Here’s a disclaimer: In this article I am going to dream a little. But just a little. I actually think all of this will happen, as laid out. I would not put money on any of it, but I do think it is accurate. As in all of my articles, it is based on facts and real research and tangible things. But no one can predict the future, and that is what I am going to do here. And some of it is admittedly quite “out there”, but I explain why I think it should be taken seriously. So take it for what it is: my own conjecture, as reasoned as I can make it, but it is conjecture.
Ten years is a long time today. It might not seem like it, but if one goes back ten years in any area of technology, things have completely changed. I refer to technology, because most business today is driven by technology. Think of one that is not! Restaurants rely on Yelp and Grubhub. Retail is now mostly in the cloud. Healthcare — well, that’s completely technology-driven. You get the idea. Transportation — have you fixed your own car recently? To do that, you now need a computer.
So what about ten years from today? That will be 2033.
Here’s one thing: AI will change. AI is on all our minds because of ChatGPT, which pushed AI into the awareness of the average person. Now everyone sees how far AI has come. It had to become a chat app that the average person could use before most people realized that AI has advanced to the point of super-human intelligence in some ways.
Most people have seen the videos of the Boston Dynamics robot Atlas dancing — better than most humans. Only five years ago the robot was clumsy and could not get up after falling. But while the sight of the robot dancing might be unsettling, most people dismissed it as a stunt. After all, it was “programmed to do that”. But the point is that the robot now has agility, highly coordinated action, and dexterity.
So fast forward ten years. We will have paired ChatGPT-like intelligence with mobility (like in Atlas).
Hmmm. Then we will have a robot that actually is intelligent, and that can walk around and do complicated and messy stuff — just like people.
And project how far AI will have further progressed in ten years:
We then have a robot walking around, and it is a lot smarter than ChatGPT — like a hundred times smarter.
If not, maybe you should be. The Terminator movie franchise was about scary robots from the future. I’m not worried about that — time travel is probably impossible. But there was a very different and more thoughtful TV series based on Terminator, called Sarah Connor Chronicles. The show was complex and examined many issues in a very thought provoking way, but one of the most intriguing was the unresolved question of whether the robot, Cameron, was sentient. The question is never answered in the show, and indeed, it could never be — because there is no way to prove or disprove sentience.
Science fiction contains no answers: it is made-up and fanciful. But it often asks powerful questions, and proposes profound ideas to consider. It is up to us to translate those ideas into our current and imagined realities, to help us to think about what is possible.
We will cure cancer. It will not be one cure. Cancer is a thousand different diseases. There are around 100 kinds of human cancer, but most of these come in myriad variations. This is because cancer is fundamentally an immune system failure: we all get cancer many times in our lifetime, but our immune system’s T cells destroy it. But eventually a cancer cell arises that has a mutation that evades the immune system, and it propagates unchecked.
Immune system signaling is extremely complex, and so cancers find a way to evade it. This happens through the process of evolution — just like in nature. For cancer, our body is a living ecosystem in which cells evolve over time, just like in the wild. Eventually a cell evolves that is able to propagate unchecked. For cancer, that is survival: as Jeff Goldblum said in Jurassic Park, “Life finds a way”.
But we are beginning to understand all of the intricacies of the immune system and the ways that cell DNA can evolve to become cancerous. Right now there are trials of a whole range of immune checkpoint targets — signaling molecules that tell the immune system to ignore a cancer cell. It is these checkpoint mechanisms that cancer often corrupts, preventing immunotherapy from working. We are learning about them and that will tell us how to fix things.
There is more to it. For example, it is often hard to get therapies to the cancer cells, particularly with solid tumors. But all of this is being uncovered, methodically. Based on current progress, and the immense acceleration of science brought about by so many advancements in methods, I expect that in ten years we will have fully decoded cancer.
This brings to mind an analogy from software programming. In programming, 90% of the effort is figuring out what the code should do. Once you know the details of what it needs to do, writing the code to do that goes very quickly. For cancer, figuring out what the code should do equates to understanding how cancer works, and deciding what pathways to block. Once you have that, today’s tools are so powerful that designing an intervention is actually the easy part.
That situation — being able to rapidly create interventions — is new. Do you know that it took only two days to create the Moderna SARS-CoV-2 (coronavirus) vaccine? (It then took a year to go through phase 1–3 trials.) Ten years ago we did not have today’s molecular design, synthesis, and delivery tools. (For delivery, think mRNA based nanoparticle vaccines.)
Medical Point of Care
Healthcare in the US has become a huge dysfunction. The fee-for-service model has corrupted medicine, creating an incentive for more services instead of curing people, and it has made insurers — bureaucrats — the gatekeepers of care. And treating healthcare as a for-profit enterprise has turned drug companies into drug pushers. The opioid epidemic is the latest example of that.
Drug companies don’t want cures: they want lifelong expensive treatments. It is therefore no surprise that most breakthrough medical discoveries come from non-profit funded research rather than drug companies. (Here is an example, describing “Five amazing health research breakthroughs in the last 12 months”.)
But the democratization of tools and methods might upend the current system — thankfully. We are seeing the “global south” arising in medical research, seeking to take ownership of drug patents and drug production. This applies to equipment too. And technology will dislodge the hold that for-profit regional hospitals have and make healthcare more of a globalized commodity.
In ten years, I predict that one will get most of one’s healthcare from the local pharmacy: you will walk into a clinic at the back (like CVS’s Minute Clinic), and an AI will ask you why you are there. Robotic equipment will perform tests right there, with results obtained within 20–60 minutes. You might then be connected to a doctor via a 3D display (no headset required), who is 8000 miles away.
(We would need to create an international care credentialing regime to make that possible, but companies like CVS will push for that.)
The remote doctor can instruct the robotic equipment to perform additional tests. Drugs and treatment will be delivered on the spot.
Healthcare point of delivery companies like GoForward are already creating this new model. And they focus on health and disease prevention, rather than fee-for-service. It is time to prioritize early detection, cures, and health maintenance instead of pushing life-long expensive drugs on people. We need healthcare to be more like the computer revolution where costs go down over time and performance goes up, and less like a bureaucracy.
If all that seems far-fetched, remember Atlas, and ChatGPT. And take note of the rise of microfluidics and “lab on a chip” technology. That is what Theranos was trying to do — they were ahead of their time. Unfortunately, instead of being transparent about things, they lied and misrepresented what they had achieved, hiding behind a proprietary wall instead of publishing papers. But the vision was sound, and it is happening.
The Cloud for Everything
We have seen the “cloud” take over computing. We are about to see it take over everything else.
In my company we are working with an organization to help them to incorporate a “cloud lab” into their strategy. Cloud labs are new. Emerald Cloud Lab (ECL) was the first. Now there are a few. ECL arose out of Carnegie Mellon University. The idea is to create automated — fully roboticized — laboratory equipment in a data center-like facility. Bio specimens are shipped there and then stored in an on-site biobank. Lab experiments are designed using a programming language that is based on Mathematica. (Thank goodness ECL did not use a “declarative” JSON/YAML language whose brittleness and lack of composability has so hamstrung today’s cloud computing providers.)
Experiments are initiated by running the program that you create, and if you want to tweak it, you just run it again. Results are automatically entered into a database, along with all of the experimental setup details. And samples go right back into the biobank, fully cataloged.
The advantages are immense. For starters:
- It enables “agile” experimentation, in which one runs small experiments with high repeatability, refines and reruns them, continually until one discovers what one is seeking.
- Equipment utilization is extremely high, because the lab is multi-tenant.
- Access to equipment is democratized: anyone with an account can use any of the lab’s equipment.
This is going to revolutionize medical research, as well as the manufacturing of drugs. It makes laboratory processes on-demand and programmable. Health science is becoming a purely computational and analytical activity.
The same will happen in other domains of work. We are seeing the early days of vertical farming. Hydroponic vertical farms can be more than ten times as productive on a square foot basis compared to traditional farms. And they can be entirely run by robots.
What about automobiles? Imagine a factory that can make any car you want, based on a design that you downloaded from the Internet.
Or a house — a team of Atlas-like robots show up and build the house that you selected from a catalog.
Or an airplane: imagine that you choose options and it gets made in an automated factory.
Okay maybe not airplanes for awhile, but that would be the next level. The point is that the trend will eventually encompass anything and everything that is manufactured. And we might even see containerized manufacturing facilities that can be deployed like shipping containers, to put them close to where resources or usage are. Medicine comes to mind: instead of stockpiling, deploy regional factories just after a pandemic strikes.
Energy and Transportation
I think that we are on the cusp of solving fusion energy.
It has been a long ride, but I think it is close. I am a former nuclear engineer, and I have a basic understanding of the physics. I think that the approach being used by Helion is very likely to work. It’s brilliant because they don’t need to maintain a stable fusion-level plasma. And it is simpler and much less expensive than other approaches.
As currently designed, the Helion design will not work for a number of reasons: (1) the reaction involving Helium-3 needs a much higher temperature to be efficient, and (2) side reactions involving protons will result in a high neutron flux, which must be absorbed by a containment coating material (such as Lithium) that will mutate into isotopes that can be recovered. But I think that these problems are engineering problems. Whether Helion will solve them is unknown, but I think that the basic design concept overcomes inherent problems with the other existing approaches.
Once a successful fusion reactor has been created, investment will flood in. We will see an acceleration of research into fusion for energy.
And then there is transportation. Imagine fusion-powered airplane propulsion. Far fetched? I don’t think so. A team at Princeton has a prototype fusion rocket engine, and that basic approach could also work for aircraft if the specific impulse can be increased eventually. If fusion energy proves itself, we will see a deluge of investment in every other potential application of fusion.
Predicting the success of fusion projects is perilous. I realize that my sanguine view is on shaky ground. But I really think it is a matter of time. We know the physics. The engineering is getting close. Hopefully it will not be like Zeno’s paradox, where we are always half way closer to the goal, but never get there.
Imagine how fusion energy and fusion propulsion will change society. Completely clean energy, at unlimited scale. Aircraft that have global range and trans-orbital hypersonic speed, and the ability to safely and economically travel to any part of the globe in an hour. And imagine an Airbus-sized vehicle that can take off from ground level with people and cargo, travel to and land on the Moon, and then return — on its own, without refueling, and at very low cost. Imagine what that would do for what can be created in orbit and beyond. It completely shatters many assumptions that we have about the difficulties of space travel. It makes space travel normal — like terrestrial travel. Again, the current designs cannot do that, but perhaps they are like the Wright Brothers’ airplane.
And imagine autonomous robotic activities beyond Earth. Are you aware of the raw materials that exist beyond Earth orbit? Atlas comes to mind again — and imagine that Atlas is trained how to make more Atlases — that opens the door to assembling things on a scale that is unimaginable today.
Okay, some of this might be 20 years away, but it’s not 50 — at least for these things to begin. Maybe some of these advancements will stall, but most will start much sooner than 50 years from now. And business is concerned with when change starts, because the start of change is the start of the end of investment in obsolete methods.
The technologies I have described each appear to be a few steps away, not many steps. No new science is needed. Only continuous improvement is needed — improvement that suddenly crosses a threshold. I am not saying that these improvements are easy or inevitable, but I am saying that they each appear to be attainable through steady incremental progress.
Once again, all this is conjecture. Conjecture is risky, so I am on a limb here. But I think that what I have predicted here is rational, and not far-fetched.
Do you know that water goes from being “perfect” body temperature of 99°F (37°C) to scalding 108°F (42°C) in a span of only 9 degrees? Technology is like that: continuous small improvements can suddenly cross a line from being impractical to world-changing, or from “doesn’t work” to “works!” During the 1800s ideas for how to make an airplane proliferated for decades until suddenly someone put just the right ingredients together in a workable way, and from that point on, we had conquered flight. And today our airplanes are nothing like that plane at Kitty Hawk, but instead are amazing metal spears that graze the stratosphere. At the time of the Wright Brothers, flight seemed so difficult to attain, and yet we were nowhere near the limit of what was possible.
For all of human history up to 1969 we never went to the Moon, and then we did, and ever since we have viewed going to the Moon as something that humans can do, whereas before, for all of human history, going to the Moon had seemed fantastic and unattainable. But today it is thought of as just engineering.
Some of the technologies that we have been discussing here are massively disruptive. They are inflection points that will pivot society, making the future and the past very different. The unimaginable will become routine.
Are you ready?