I was honored when MIT Technology Review invited me to be the first guest curator of its 10 breakthrough technologies. Limiting the list was difficult. I wanted to choose things that will not only create headlines in 2019, but capture this moment in technological history – forcing me to think about how innovation has evolved over time.
My thoughts went to – of all things – the team. Teams are an excellent embodiment of the history of innovation. People have been using them since 4000 BC, when Mesopotamian farmers aerated land with sharpened sticks. We have since been slowly working on improving and improving, and today’s teams are technological wonders.
This story is part of our March / April 2019 issue
See the rest of the problem
subscribe
Read the introduction of Bill Gates
But what exactly is the goal of a team?
In addition to his introductory essay, read Bill Gates’ conversation with editor-in-chief Gideon Lichfield. Below are his choices for the 10 breakthrough technologies.
Dexterity of the robot
Nicolas Ortega
Dexterity of the robot
-
Why it matters
If robots could learn to cope with the messiness of the real world, they could perform many more tasks.
-
Key players
OpenAI
Carnegie Mellon University
University of Michigan
UC Berkeley -
availabilty
3-5 years
Robots teach themselves how to deal with the physical world.
Despite all the talk about machines taking on jobs, industrial robots are still awkward and not flexible. A robot can repeatedly pick up a component on an assembly line with amazing precision and never get bored – but move the object half an inch or replace it with something else, and the machine will be awkwardly awkward or thin with thin air.
But although a robot cannot yet be programmed to figure out how an object can be grasped by looking at it, as people do, it can now learn to manipulate the object itself through virtual trial and error.
One such project is Dactyl, a robot that has taught itself to turn a toy building block in its fingers. Dactyl, from the non-profit OpenAI from San Francisco, consists of a ready-made robot hand surrounded by a series of lights and cameras. Using what is known as amplification learning, neural network software learns to understand the block and to run it in a simulated environment before the hand actually tries it out. The software initially experiments randomly, so that connections within the network become stronger over time as it gets closer to its goal.
It is usually not possible to transfer such virtual exercises to the real world, because things like friction or the different properties of different materials are so difficult to simulate. The OpenAI team circumvented this by adding randomness to the virtual training, giving the robot a proxy for the messiness of reality.
We need further breakthroughs for robots to control the advanced agility required in a real warehouse or factory. But if researchers can use this type of learning reliably, robots can eventually assemble our gadgets, load our dishwashers and even help grandma get out of bed. —Will Knight
New-wave nuclear energy
Bob Mumgaard / Plasma Science and Fusion Center / MIT
Advanced fusion and fission reactors are getting closer to reality.
New nuclear designs that have gained strength in the past year promise to make this power source safer and cheaper. Among them are generation IV fission reactors, an evolution of traditional designs; small modular reactors; and fusion reactors, a technology that seemed to be just out of reach forever. Generation IV fission design developers, such as Canada’s Terrestrial Energy and Washington-based TerraPower, have entered into R&D partnerships with utilities, focusing on network supply (somewhat optimistic, perhaps) by the 2020s.
Small modular reactors usually produce tens of megawatts of power (for comparison, a traditional nuclear reactor produces around 1,000 MW). Companies such as the Oregon NuScale say that miniaturized reactors can save money and reduce environmental and financial risks.
Progress has even been made with merger. Although no one expects delivery before 2030, companies such as General Fusion and Commonwealth Fusion Systems, a MIT spin-out, are making some progress. Many consider fusion to be a pipe dream, but since the reactors cannot melt and cause long-lived, high-level waste, it should have much less public resistance than conventional nuclear. (Bill Gates is an investor in TerraPower and Commonwealth Fusion Systems.) —Leigh Phillips
Predicting preemies
Nenov | Getty
Predicting preemies
-
Why it matters
15 million babies are born prematurely each year; it is the leading cause of death for children under five
-
Main role
Akna Dx
-
availabilty
A test can be offered in the consultation rooms within five years
A simple blood test can predict whether a pregnant woman is at risk of giving birth prematurely.
Our genetic material usually lives in our cells. But small amounts of “cell-free” DNA and RNA also float in our blood, often released by dying cells. In pregnant women, the cell-free material is an alphabet soup of nucleic acids from the fetus, placenta and mother.
Stephen Quake, a bio-engineer at Stanford, has found a way to use it to address one of the most persistent problems of medicine: the approximately one in ten premature babies.
Free-floating DNA and RNA can provide information that previously required invasive ways to grab cells, such as taking a tumor biopsy or piercing the belly of a pregnant woman to perform amniotic fluid puncture. What has changed is that it is now easier to detect and sequence the small amounts of cell-free genetic material in the blood. In recent years, researchers have begun to develop blood tests for cancer (by finding the significant DNA of tumor cells) and for prenatal screening of conditions such as Down’s syndrome.
The tests for these disorders are based on the search for genetic mutations in the DNA. RNA, on the other hand, is the molecule that regulates gene expression – how much of a protein is produced from a gene. By determining the sequence of free-floating RNA in maternal blood, Quake can recognize fluctuations in the expression of seven genes he selects as associated with premature birth. This allows him to identify women who are likely to give birth prematurely. Once warned, doctors can take measures to prevent an early birth and to give the child a better chance of survival.
The technology behind the blood test, Quake says, is fast, easy and costs less than $ 10 per measurement. He and his staff have launched a startup, Akna Dx, to commercialize it. —Bonnie Rochman
Gut probe in a pill
Bruce Peterson
Gut probe in a pill
-
Why it matters
The device makes it easier to examine and study intestinal diseases, including a disease that prevents millions of children in poor countries from growing well
-
Main role
Massachusetts General Hospital
-
availabilty
Now used in adults; infant testing starts in 2019
A small device to be swallowed provides detailed images of the gut without anesthesia, even in infants and children.
Bowel disorders in the environment (EED) can be one of the most expensive diseases you have never heard of. Characterized by inflamed intestines that poorly absorb leaks and nutrients, it is common in poor countries and is a reason why many people there are malnourished, have developmental delays and never reach normal length. Nobody knows exactly what causes EED and how it can be prevented or treated.
Practical screening to detect it would help medical professionals know when to intervene and how. Therapies are already available for infants, but diagnosing and studying diseases in the gut of such young children often requires anesthesia and the insertion of a tube called an endoscope in the throat. It is expensive, uncomfortable and not practical in parts of the world where EED prevails.
That is why Guillermo Tearney, pathologist and engineer at Massachusetts General Hospital (MGH) in Boston, is developing small devices that can be used to inspect the intestine for signs of EED and even obtain tissue biopsies. Unlike endoscopes, they are easy to use during a primary care visit.
Tearney’s swallowable capsules contain miniature microscopes. They are attached to a flexible string-like chain that provides power and light while images are sent to a briefcase-like console with a monitor. This allows the caregiver to pause the capsule at points of interest and pull it out at the end so that it can be sterilized and reused. (Although the gag sounds inducing, the Tearney team has developed a technique that they believe does not cause any inconvenience.) It can also carry technologies that map the entire surface of the digestive tract with the resolution of a single cell or capture three-dimensional cross-sections some millimeters deep.
The technology has various applications; in HGH it is used to screen for the esophagus of Barrett, a precursor of esophageal cancer. For EED, the Tearney team has developed an even smaller version for use in infants who cannot take a pill. It has been tested on adolescents in Pakistan, where EED prevails, and baby tests are scheduled for 2019.
The small probe helps researchers to answer questions about the development of EED – such as which cells it affects and whether bacteria are involved – and to evaluate interventions and possible treatments. —Courtney Humphries
Custom cancer vaccines
Paper boat creative Getty
Custom cancer vaccines
-
Why it matters
Conventional chemotherapy demands a heavy load on healthy cells and is not always effective against tumors
-
Key players
Biontech
Genentech -
availabilty
In human testing
The treatment stimulates the body’s natural defenses to only destroy cancer cells by identifying mutations unique to each tumor
Scientists are about to launch the first personalized cancer vaccine. If it works as hoped, the vaccine, which activates a person’s immune system to identify a tumor through its unique mutations, can effectively stop many types of cancers.
By using the body’s natural defenses to selectively destroy only tumor cells, the vaccine, unlike conventional chemotherapy, limits damage to healthy cells. The attacking immune cells can also be vigilant in detecting stray cancer cells after the first treatment.
The possibility of such vaccines began to take shape in 2008, five years after the Human Genome Project was completed, when geneticists published the first sequence of a tumor cell.
Shortly thereafter, researchers began comparing the DNA of tumor cells with that of healthy cells – and other tumor cells. These studies confirmed that all cancer cells contain hundreds, if not thousands, of specific mutations, most of which are unique to each tumor.
A few years later, a German start-up called BioNTech provided convincing evidence that a vaccine containing copies of these mutations could catalyze the body’s immune system to produce T cells prepared for detecting, attacking and destroying all cancer cells in which they are located.
In December 2017 BioNTech started a major test of the vaccine in cancer patients, in collaboration with the biotech giant Genentech. The ongoing study focuses on at least 10 solid cancers and aims to register more than 560 patients at locations around the world.
The two companies design new production techniques to produce thousands of personally adapted vaccines cheaply and quickly. That will be difficult, as making the vaccine involves performing a biopsy on the patient’s tumor, sequencing and analyzing the DNA and transferring that information to the production site. Once produced, the vaccine must be delivered immediately to the hospital; delays can be deadly. —Adam Piore
The cow-free burger
Bruce Peterson / Styling: Monica Mariano
The cow-free burger
-
Why it matters
Livestock farming causes catastrophic deforestation, water pollution and greenhouse gas emissions
-
Key players
Beyond meat
Impossible food -
availabilty
Vegetable now; lab grown around 2020
Both laboratory-grown and plant-based alternatives approximate the taste and nutritional value of real meat without the environmental damage.
The UN expects that by 2050 the world will have 9.8 billion people. And those people are getting richer. Neither of these trends is a good omen for climate change – especially as people eat more meat as they escape poverty.
By that date, people are expected to consume 70% more meat than in 2005. And it turns out that breeding animals for human consumption is one of the worst things we do to the environment.
Depending on the animal, producing a pound of meat protein with Western industrialized methods requires 4 to 25 times more water, 6 to 17 times more land and 6 to 20 times more fossil fuels than producing a pound of vegetable protein.
The problem is that people are unlikely to stop eating meat soon. Which means that laboratory-grown and plant-based alternatives are the best way to limit destruction.
To make meat in a laboratory, muscle tissue must be extracted from animals and cultured in bioreactors. The end product is very similar to what you would get from an animal, although researchers are still working on the taste. Researchers at Maastricht University in the Netherlands, who produce lab meat on a large scale, believe that they will have a laboratory-grown hamburger available next year. A disadvantage of laboratory-grown meat is that the environmental benefits are still vague at best – a recent report from the World Economic Forum says that emissions from laboratory-grown meat would only be about 7% lower than those from beef production.
The better environmental case can be made for vegetable meat from companies such as Beyond Meat and Impossible Foods (Bill Gates invests in both companies), which use pea proteins, soy, wheat, potatoes and vegetable oils to imitate the texture and taste of animal meat.
Beyond Meat has a new 26,000 square foot (2,400 square meter) factory in California and has already sold more than 25 million citizens from 30,000 stores and restaurants. According to an analysis by the Center for Sustainable Systems at the University of Michigan, a Beyond Meat patty would probably generate 90% fewer greenhouse gas emissions than a conventional hamburger made from a cow. —Markkus Rovito
Nico Ortega
Carbon dioxide collector
-
Why it matters
Removing CO2 from the atmosphere can be one of the last viable ways to stop catastrophic climate change
-
Key players
Carbon Engineering
Climeworks
Global thermostat -
availabilty
5-10 years
Carbon dioxide collector
Practical and affordable ways to capture carbon dioxide from the air can absorb excess greenhouse gas emissions.
Even if we delay carbon dioxide emissions, the warming effect of the greenhouse gas can persist for thousands of years. To prevent a dangerous rise in temperature, the UN climate panel is now concluding that the world must remove no less than 1 trillion tons of carbon dioxide from the atmosphere this century.
In a surprising finding last summer, Harvard climate scientist David Keith calculated that machines could theoretically achieve this for less than $ 100 per tonne, using an approach known as direct air capture. That is an order of magnitude cheaper than previous estimates that have led many scientists to dismiss technology as far too expensive, although it will still be years before the costs get anywhere near that level.
But once you’ve captured the carbon, you’ll still have to figure out what you can do with it.
Carbon Engineering, the Canadian start-up Keith, co-founder in 2009, plans to expand its pilot plant to increase the production of its synthetic fuels, with carbon dioxide captured as the main ingredient. (Bill Gates is an investor in Carbon Engineering.)
The direct air capture plant of Climeworks, based in Zurich in Italy, will produce methane from trapped carbon dioxide and hydrogen, while a second plant in Switzerland will sell carbon dioxide to the soft drinks industry. The same applies to New York’s Global Thermostat, which built its first commercial plant in Alabama last year.
However, if it is used in synthetic fuels or soft drinks, the carbon dioxide usually ends up in the atmosphere again. The ultimate goal is to lock up greenhouse gases forever. Some may be embedded in products such as carbon fiber, polymers or concrete, but much more will just have to be buried underground, a costly job that does not seem to support any business model.
Removing CO2 from the air is technically one of the most difficult and expensive ways of dealing with climate change. But given how slowly we are reducing emissions, there are no more good options. —James Temple
Bruce Peterson
An ECG on your wrist
Legal approval and technological advances make it easier for people to continuously monitor their hearts with portable devices.
Fitness trackers are not serious medical devices. An intense workout or loose band can mess with the sensors that read your wrist. But an electrocardiogram – the kind of doctors used to diagnose abnormalities before they cause a stroke or heart attack – requires a visit to a clinic and people often fail to take the test on time.
Smart watches with ECG, made possible by new regulations and innovations in hardware and software, offer the convenience of a portable device with something that comes closer to the precision of a medical device.
An Apple Watch-compatible band from startup AliveCor in Silicon Valley that can detect atrial fibrillation, a common cause of blood clots and stroke, received FDA approval in 2017. Last year, Apple released its own FDA approved ECG feature, embedded in the view itself.
Shortly thereafter, health equipment company Withings also announced plans for an ECG-equipped watch.
Current wearables still use only a single sensor, while a real ECG has 12. And no wearable can still detect a heart attack while it is taking place.
But this can change quickly. Last fall, AliveCor presented preliminary results to the American Heart Association about an app and a system with two sensors that can detect a certain type of heart attack. —Karen Hao
TheDman | Getty
Sanitary facilities without sewerage
-
Why it matters
2.3 billion people lack safe sanitation and many die as a result
-
Key players
Duke university
University of South Florida
Biomass control
California Institute of Technology -
availabilty
1-2 years
Sanitary facilities without sewerage
Low-energy toilets can work without sewage and treat waste on site.
About 2.3 billion people do not have good sanitary facilities. The lack of proper toilets encourages people to dump fecal matter in nearby ponds and streams, causing bacteria, viruses and parasites to spread that can cause diarrhea and cholera. Diarrhea causes one in nine deaths among children worldwide.
Now researchers are building a new type of toilet that is cheap enough for developing countries and not only can throw away waste, but can also treat it.
In 2011, Bill Gates created what was essentially the X prize in this area – the Reinvent the Toilet Challenge. Since the competition was launched, several teams have put prototypes in the field. They all process the waste locally, so no large amount of water is needed to take it to a remote processing plant.
Most prototypes are self-contained and do not require sewerage, but they look like traditional toilets in small buildings or storage containers. The NEWgenerator toilet, designed at the University of South Florida, filters contaminants with an anaerobic membrane, which has pores that are smaller than bacteria and viruses. Another project, from Connecticut-based Biomass Controls, is a refinery the size of a sea container; it heats the waste to produce a carbon-rich material that can, among other things, fertilize soil.
A disadvantage is that the toilets do not work on every scale. For example, the Biomass Controls product is primarily designed for tens of thousands of users per day, making it less suitable for smaller villages. Another system, developed at Duke University, is only intended for a few houses in the neighborhood.
The challenge now is to adapt these toilets cheaper and better to communities of different sizes. “It’s great to build one or two units,” said Daniel Yeh, an associate professor at the University of South Florida, who led the NEWgenerator team. “But to give technology a real impact on the world, the only way to do that is to produce the units in mass production.”
Bruce Peterson
Flexible speaking AI assistants
-
Why it matters
AI assistants can now perform conversation-based tasks, such as reserving a restaurant reservation or coordinating a drop-off of the package instead of just performing simple assignments
-
Key players
Google
Alibaba
Amazon -
availabilty
1-2 years
Flexible speaking AI assistants
New techniques that establish semantic relationships between words make machines better at understanding natural language.
We’re used to AI assistants – Alexa plays music in the living room, Siri sets alarms on your phone – but they didn’t really live up to their supposed smarts. They would have simplified our lives, but they have hardly made a dent. They only recognize a limited number of guidelines and are easily stumbled by deviations.
But a few recent developments are about to expand the repertoire of your digital assistant. In June 2018, researchers from OpenAI developed a technique that trains an AI on non-labeled text to avoid the cost and time of manually categorizing and tagging all data. A few months later, a Google team unveiled a system called BERT that, by studying millions of sentences, has learned how to predict missing words. In a multiple-choice test, it did just as well as people when filling gaps.
These improvements, coupled with better speech synthesis, allow us to give simple assignments to AI assistants to conduct conversations with them. They can deal with daily details such as taking notes at meetings, finding information or shopping online.
Some are already there. Google Duplex, the eerie human-like upgrade from Google Assistant, can take your calls for spammers and telemarketers. You can also call to schedule restaurant reservations or salon appointments.
In China, consumers get used to Alibaba’s AliMe, which coordinates the delivery of packages over the telephone and negotiates the price of goods via chat.
But although AI programs have become better at figuring out what you want, they still can’t understand the point. Lines are processed in a script or statistically generated, which indicates how difficult it is to provide machines with real language comprehension. As soon as we cross that obstacle, do we see another evolution, perhaps from logistics coordinator to babysitter, teacher – or even friend? —Karen Hao
