In the first season of Crazy/Genius, The Atlantic's podcast on tech and culture, I asked experts to help me answer some of the hardest questions I could imagine. Would the U.S. economy be better off if the government broke up Amazon? Is smartphone use a behavioral addiction? And, seriously, where are all the aliens? In our upcoming season, the focus shifts from hard questions to radical answers, featuring a ragtag cast of scientists, tinkerers, and artists: a Harvard professor who's convinced that aging is just another curable disease; the chief engineer behind the world's most advanced self-driving car technology; climate scientists who study volcanic eruptions and see a lesson for slowing global warming; a startup couple developing the future of "meat" (it chirps); a concert pianist who plays duets with an algorithm, and whose work might be the future of creativity.

The recent explosion of interest, hype, and fear about artificial intelligence, data science, machine learning, and robotics has focused a spotlight on software engineers. The business magnate Elon Musk has called for regulation and the President of Russia Vladimir Putin has declared that world domination will result from mastering AI. Are software engineers responsible for these outcomes? Here, I argue that software engineers have less control over their designs than they likely realize. Instead, software technologies are evolving in a Darwinian way, or more precisely, they are co-evolving with human culture.

Japanese farmers are testing a new drone that can hover above paddy fields and perform backbreaking tasks in a fraction of the time it takes a labourer. The drone applies pesticides and fertilizer to a rice field in 15 minutes - a job that takes more than an hour by hand and requires farmers to lug around heavy tanks. Developers of the new agricultural drone say it offers high-tech relief for rural communities facing a shortage of labour as young people leave for the cities. They plan to release the $36,000 (£28,000) Nile-T18, which farmers can control through an iPad app, next year. Japanese farmers are testing a new drone that can hover above paddy fields and perform backbreaking tasks in a fraction of the time it takes a labourer.

More than one million Hongkongers are at risk of losing their jobs to artificial intelligence over the next two decades, according to a new study by a local think tank. The One Country Two Systems Research Institute on Tuesday unveiled research which estimates about 28 per cent of the city's 3.7 million jobs are vulnerable to automation. These workers, which include secretaries, accountants and auditors, face a 70 per cent chance of being substituted for machines before 2038, the pro-Beijing research unit said. But Hong Kong employees face a lower risk of encroachment from AI compared to their counterparts in other advanced economies such as the United States, Britain and Japan. The brighter forecast for the city was due to its economic structure, which included a smaller manufacturing sector, research officer Kristine Yang said.

The battle against propaganda bots is an arm's race for our democracy. It's one we may be about to lose. Bots--simple computer scripts--were originally designed to automate repetitive tasks like organizing content or conducting network maintenance, thus sparing humans hours of tedium. Companies and media outlets also use bots to operate social-media accounts, to instantly alert users of breaking news or promote newly published material. But they can also be used to operate large numbers of fake accounts, which makes them ideal for manipulating people.

HAIFA, ISRAEL (August 22, 2018) – Dr. Kira Radinsky and Shahar Harel of the Technion-Israel Institute of Technology Computer Science Department have developed a smart system for the development of new drugs. Founded on artificial intelligence and deep learning, the system is expected to dramatically shorten and reduce the costs of drug development. It will be presented this week during the KDD 2018 conference in London. Drug production is a costly and lengthy process. Costs of half a billion to 2.5 billion dollars per drug, over 10-15 years are common numbers in the world of pharmacology.

Java-wise, I sometimes pick up a medium decaf at the Pastry Chef, where I might also grab a plain croissant, though most days I have coffee and a toasted poppy bagel at Bella's Restaurant. I do not get the bagel hollowed out, viewing this as an affectation. I freely grant this information; all you data miners fooling around inside my car can do with your findings what you please. The deeply paranoid national obsession with hiding personal info from online snoops is misplaced and counterproductive. These guys are going to get it anyway, so why obsess about it?

With experience spanning academia, the private sector, government and even a start-up, Kalev Ruberg came to the mining sector rather late in his career. When he arrived 12 years ago, he found a "very fractured landscape" from a technology systems point of view. "Independent development may result in a one-off success, but being entrepreneurial and working in silos is the biggest challenge we all have," said Ruberg. "We need to govern the way innovation proceeds into production. The innovation success we've had at Teck has come from a very disciplined platform approach because it has staying power."

Single-particle tracking involves tracking the motion of individual particles, such as viruses, cells and drug-loaded nanoparticles, within fluids and biological samples. The technique is widely used in both physical and life sciences. The team at UNC-Chapel Hill that developed the new tracking method uses particle tracking to develop new ways to treat and prevent infectious diseases. They examine molecular interactions between antibodies and biopolymers and characterize and design nano-sized drug carriers. Their work is published in the Proceedings of the National Academy of Sciences.

Over the past decades, researchers and ML practitioners have come up with better and better ways to build, understand and improve the quality of ML models, but mostly under the key assumption that the training data is distributed identically to the testing data. In many real-world applications, however, some potential training examples are unknown to the modeler, due to sample selection bias or, more generally, covariate shift, i.e., a distribution shift between the training and deployment stage. The resulting discrepancy between training and testing distributions leads to poor generalization performance of the ML model and hence biased predictions. We provide novel algorithms that estimate the number and properties of these unknown training examples---unknown unknowns. This information can then be used to correct the training set, prior to seeing any test data. The key idea is to combine species-estimation techniques with data-driven methods for estimating the feature values for the unknown unknowns. Experiments on a variety of ML models and datasets indicate that taking the unknown examples into account can yield a more robust ML model that generalizes better.