Nanosensors! Autonomous vehicles! [Yawn.] We’ve seen these before.
Organs-on-chips? Optogenetics?!? Now this is interesting.
The Davos-based World Economic forum released a list of 10 emerging technologies that made a tour of the press a few weeks ago. There is no guidance for the list. There is no set of recommendations or concerns. Each technology’s one page write-up is in a different style. Let’s look at the list in the same way we did past technology forecasts.
Questions to ask:
- How would we make this into a prediction with numbers and years?
- What scenarios could enable this future to happen?
- What challenges aren’t mentioned about this technology?
1. Nanosensors and the Internet of Things
Quantifying this prediction comes in the opening paragraph:
New IoT devices are announced almost daily, and analysts expected up to 30 billion of them to be online by 2020.
My challenge always with sensors has been; (1) what device or product will they be added to? (2) how will they get access to the Internet? (3) how will they draw power, and I was concerned in the past with (4) how will the data be analyzed? Any commercialization effort must address these constraints. The technology and capability of sensors is significant.
Several scenarios could accelerate adoption:
- Sensors for health. Hospitals are already full of sensors and there is immediate benefit and a method for analysis. Incorporation of more sensors on body, in hospital garments and beds could improve outcomes and alleviate pain and suffering for patients more quickly. (5 years – 2022)
- Smart vehicles. Beyond autonomous vehicles, could a fleet of Google Maps cars tricked out with the right sensors reduce the cost of road maintenance? Could integration of sensors reduce maintenance and downtime for aircraft? GE’s use of sensors on engines is well known, could the same use allow for more advanced materials to be used in wings and elsewhere? (15 years – 2032)
- Infrastructure. How would we sell ‘smart beams’ to contractors? Could they be written in to state purchase contracts to avoid fiascoes like the Bay Bridge? Would they be paid for over time with service and maintenance contracts? (10 years – 2027)
- With most products, the naive assumption is usually the right one – tomorrow’s sales will come from today’s markets. Sensors are used in phones. We’ve seen the Amazon echo achieve adoption in households. Could the future live in our kitchens and pockets, but with a greater set of capabilities? Many of the sensors in a phone are not now currently utilized. What must change for them to see greater adoption? (5 years – 2022)
2. Next Generation Batteries
Batteries are a source of personal interest and professional pain. The promise is real, the development and commercialization process is long, complex and difficult to improve on. A claim that a battery technology or use will become mainstream should achieve a benchmark of at least 1/3 of new installed capacity.
Batteries are hard to develop, require unique materials and novel manufacturing methods. For this technology to emerge, one of the following scenarios must occur:
- Nylon moment. A big company must pursue batteries with the same intent and dedication that DuPont’s Carothers used to pursue the development of Nylon.
- Software and simulation. Batteries require hours of tests, sometimes months, sometimes years. They are a slow, controlled bomb. To accelerate commercialization, there must be a way to short cut the testing and evaluation. Predictive materials software would shorten the development cycle.
- Pilot scale automation. High performance batteries have depended on materials made with uniformity and scale that is tough to scale. There are many battery technologies that made a splashy press release upon emerging from a lab, but never left a production site. Using technologies with a clearer path to scale up will accelerate development.
- Reverse compatible to supply chain. For batteries to emerge, they must be easily integrated into whatever supply chain where they will be used. The manufacturing must be easily done with existing processes. The use must allow for easy and convenient charging.
3. The Blockchain
I couldn’t understand what exactly was going to ’emerge’ from this discussion of practical application of bitcoin and etherium related cryptocurrency methods to areas like government and pharmaceutical development.
This portion could have been summed up with, ‘More honesty and transparency!’
4. Two-Dimensional Materials
The authors begin with graphene and list out all of the other “-phenes’; borophene, white graphene, germanene, phosphorene, silicene and stranene (from tin). Perhaps we should re-label polymerous nanofibers as, “porous polymerenes.”
The list of applications is what is often seen for new materials – batteries, filtration, pollution control, structural composites and medical applications. There are many scenarios for how these materials can be adopted, but they should all be measured at the end market level.
My forecast is that the -phenes won’t be used in more than 1/3 of the parts in a major consumer vehicle or passenger airplane within 20 years. 1/3 could be measured based on count or value.
A second forecast is that before this set of suffix-phenes achieve their goal, another set of materials will emerge with ‘wonderous’ properties.
5. Autonomous Vehicles
Autonomous vehicles are happening at a pace faster than anyone had expected. Uber and Volvo are rolling out a fleet of for-hire vehicles in Pittsburgh. Further, this timeframe beats my own past prediction that a fleet of self-driving vehicles would roll out by 2019. (This is the first prediction like this I’ve been right on!)
Autonomous vehicles will happen. Despite much discussion and warning, their impact will surprise many and disrupt major industries, economies and regions.
The report contemplates the miniaturization and automation of pharmaceutical animal models such that they can be replicated in silica.
Scenarios where such devices become readily available and/or disruptive could include:
- O-o-C use as sensors.
- Challenges around toxicity where such sensors are important in health and safety.
- Unfortunately, it is easy to see scenarios where use of such devices could enable new types of illegal drugs. Such scenarios often accelerate adoption of new technologies.
- Any device that aids in the reduction of obesity would show fast and immediate adoption. If these devices enabled use of personalized medicines in a fast detection / production setting, they could enable such methods.
- Lastly, any kind of pandemic or disease threat whose challenge was met with these devices would promote fast adoption.
7. Perovskite Solar Cells
This is a fancy name for solar cells made out of materials that are not silicon; not hard to make, not extensions of the semiconductor supply chain and hence not constrained by current systems and uses.
However, they are constrained by current systems. It is a common forecasting and technology trap to believe that a ‘different’ or ‘better’ version of a current thing will totally change the use case. The Perovskite solar cells will be subject to the same cost and use scenarios of current silicon based solar cells. Further, that cost structure will raise the barrier on where the new cells will be used. The performance:cost ratio will have to be much greater to achieve adoption.
Success would occur with annual production of 5% of the current silicon volumes. The scenarios where these solar cells achieve adoption could include:
- There must be a significant cost advantages throughout the supply chain for them to achieve adoption . The first production plants must be very cheap . The methods used must be very productive . Conversion to end sites must be very easy . If each of these methods is a factor of 10 or more better than current cells, then adoption could occur within 10 years of their first invention.
- These cells could enable a country or entity to leapfrog existing technologies.
- Lastly, any kind of unique advantage – use in orbit, use in the ocean, etc. could accelerate adoption.
8. Open AI Ecosystem
Russell and Norvig describe AI best – here the report discusses the value of an “Internet of AI”, where software systems are able to seamlessly interact with each other. The AI discussion starts to sound like the ways in which ‘Blockchain’ creates better honesty and transparency. AI will make everything better. In many ways AI is already here – in the guise of Excel, Minitab, and any other software that allows one persons work to scale and extend.
Could autonomous vehicles be our first mass adoption of independent AI?
An open AI ecosystem could be said to exist when; (1) 30% of the interactions are based on AI systems, (2) such that AI is clearly defined, and such that (3) those interactions happen in finance, personal management or food systems.
Scenarios where open AI will occur are in those arenas where they make a human operator better. Any scenario where everyone can acknowledge that an individual operator could be overwhelmed is an area prime for this kind of AI disruption. We find those in the areas where computers and the Internet were first developed – where there are (a) fast live operations, and (b) there are broad operations which can take a lot of time.
- Self driving vehicles.
- Ocean going vessels.
- Air traffic control.
- Complex production operations.
This will prove true at any point if a single major site has adopted an ‘Open AI Ecosystem’ in daily operation. That could be an industrial mine, a train station, some kind of crop harvesting – there are many examples. The implementation of the first ‘ecosystem’ will be enough to count (just like AI with a single city installation). In fact, the first ‘Open AI Ecosystem’ might be Uber’s Pittsburgh roll out.
Optogenetics is the use of light based systems combined with modified cellular components to create fast-paced biological change and inputs.
This is similar to two-dimensional materials and organs-on-chips. Cool ideas. Lots of potential. Measuring how such a concept makes it into routine use is difficult. Adoption could be found by measuring;
- The number of publications
- The number of patent citations
- The creation of a business focused in this area
10. Systems Metabolic Engineering
Our 10th and final concept looks at the evaluation of industrial supply chains in the same way we look at an ecosystem. The modern industrial supply chain is based on petrochemicals. Would a more sophisticated and nuanced approach allow greater use of recycled and engineered goods?
Could the use of specialty microbes at points in the supply chain improve yields?
Using materials across the supply chain other than petrochemicals enables the use of more diverse feedstock. In many ways #10 is a combination of; (1) greater mapping of the industrial chemical supply chain and (2) more sophisticated use of bio-fuels and other feedstocks.
Measuring success in this area will be complex, but would happen when:
- 1/3 of a top 20 material is made with this material. (Top 20 as measured by volume, mass or $s.)
- A new material enters the top 50 that is only possible with such an ecosystem.
This prediction is hard to see as *not* happening. As businesses become more efficient, as data becomes easier to access, our ability to understand sourcing grows. Why mine new material when it can be found in the recycling bin?
This is an interesting, diverse, difficult to systematically evaluate, list of technologies. Some were new to me. Some have already hit milestones that others on this list will never achieve. Such lists help populate ideation sessions, but a great deal of work lies between a technology’s presence on this list and actual broad impact.