Last week, Elon Musk dashed off 125 characters announcing a remarkably ambitious plan to send Amtrak to an early grave. "Just received verbal govt approval for The Boring Company to build an underground NY-Phil-Balt-DC Hyperloop. NY-DC in 29 mins," he proclaimed in a tweet. Sign up to get Backchannel's weekly newsletter. Yet something about this particular moonshot seemed off.

This spring, as part of their coursework, four Stanford University students found themselves in Coronado, California, doing pushups on the beach and charging into a 61-degree surf while overseen by Navy SEAL trainers. They performed this extraordinary homework to better understand the process of inculcating recruits into the elite corps of military frogmen and women. The end result of their (literal) immersion was a solution to an inefficiency in evaluating prospective SEALS: the time-consuming process of analyzing the mountains of comments made about each candidate. Tackling the problem like the internet entrepreneurs they hoped to become, the students created a mobile app to streamline the process. Their reward was thanks from a grateful military establishment--and college credit. Dan Raile is a freelance journalist based in San Francisco.

The plug-in – loved and hated across the world – won't actually be put out of its misery until 2020. But the company that makes it has signalled it will come to an end. Flash was once the technology powering the many games and videos of the early internet. As an animation platform it allowed for the creation of clickable games and videos on places like YouTube, and in so doing helped create the web as we know it today. The I.F.O. is fuelled by eight electric engines, which is able to push the flying object to an estimated top speed of about 120mph.

It's time for every product manager, entrepreneur or business leader to get up to speed on machine learning. Even if you're not building the next chatbot or self driving car, you'll probably need to use machine learning in your product sooner rather than later to stay competitive. The good news is you don't need to invent the technology (though kudos if you do), just leverage what already exists. Tech companies have open sourced tools and platforms (Amazon AI, TensorFlow, originally developed by Google, and many others) that make machine learning accessible to virtually any company today. When I started in machine learning I knew next to nothing about it, yet in a relatively short time I was leading the development of products with machine learning at their very core (such as this).

I found www.fast.ai in April 2017 and was a bit blown away. An AI course focused on actually getting things done? I was just finishing Yaser Abu-Mostafa's CS1156x'Learning from Data' on edX, and while a great theoretical course, it did cut down a lot of my enthusiasm for Machine Learning. I guess learning to code in Python while writing Linear Regression models by hand has that effect. What really got me about Jeremy Howard's'Practical Deep Learning I' (which I'll call FAI01/FADL1) was that, over and over again, he'd explain a thing, you'd go do it, and all of a sudden you're catapulted to the forefront of applied ML.

One of the major hurdles preventing the full exploitation of information from online communities is the widespread concern regarding the quality and credibility of user-contributed content. Prior works in this domain operate on a static snapshot of the community, making strong assumptions about the structure of the data (e.g., relational tables), or consider only shallow features for text classification. To address the above limitations, we propose probabilistic graphical models that can leverage the joint interplay between multiple factors in online communities --- like user interactions, community dynamics, and textual content --- to automatically assess the credibility of user-contributed online content, and the expertise of users and their evolution with user-interpretable explanation. To this end, we devise new models based on Conditional Random Fields for different settings like incorporating partial expert knowledge for semi-supervised learning, and handling discrete labels as well as numeric ratings for fine-grained analysis. This enables applications such as extracting reliable side-effects of drugs from user-contributed posts in healthforums, and identifying credible content in news communities. Online communities are dynamic, as users join and leave, adapt to evolving trends, and mature over time. To capture this dynamics, we propose generative models based on Hidden Markov Model, Latent Dirichlet Allocation, and Brownian Motion to trace the continuous evolution of user expertise and their language model over time. This allows us to identify expert users and credible content jointly over time, improving state-of-the-art recommender systems by explicitly considering the maturity of users. This also enables applications such as identifying helpful product reviews, and detecting fake and anomalous reviews with limited information.

Artificial Intelligence is reshaping your relationship with the world and it's just getting started. Tesla's autopilot, job automation, the products you'stumble upon' online - it's entering our daily lives, careers, businesses, even our homes with such blistering pace you probably haven't even realized it. There's a reason Andrew Ng, the founder of $100m company Coursera said "Artificial Intelligence is the new electricity" - soon it'll be as much a part of your daily life as your smartphone, except without the off button. But here's where things get really crazy. This time round, the revolution will see machines taking on tasks no human intellect could ever perform.

Digital technologies like the internet and smartphones are transforming our lives and society. They are proving to be powerful tools for liberating individuals' creative and entrepreneurial potential, as well as providing new educational opportunities and higher wages for marginalized people, both in the U.S. and around the globe. Unfortunately, in the U.S., outdated government regulations and weak consumer protections are undermining these opportunities. What's more, the Trump administration has not yet made significant moves to address this growing crisis: As of this writing, five key White House positions are vacant, without even acting directors or interim leaders to help the executive branch formulate U.S. science and technology policy. As the founder of both the Open Technology Institute and the X-Lab policy and innovation organization, I have spent years at the heart of many Washington, D.C. battles over technology policy, fighting for ideas that would best serve American workers and the general public.

Conventionally, the questions on a test are assumed to be kept secret from test takers until the test. However, for tests that are taken on a large scale, particularly asynchronously, this is very hard to achieve. For example, TOEFL iBT and driver's license test questions are easily found online. This also appears likely to become an issue for Massive Open Online Courses (MOOCs, as offered for example by Coursera, Udacity, and edX). Specifically, the test result may not reflect the true ability of a test taker if questions are leaked beforehand. In this paper, we take the loss of confidentiality as a fact. Even so, not all hope is lost as the test taker can memorize only a limited set of questions' answers, and the tester can randomize which questions to let appear on the test. We model this as a Stackelberg game, where the tester commits to a mixed strategy and the follower responds. Informally, the goal of the tester is to best reveal the true ability of a test taker, while the test taker tries to maximize the test result (pass probability or score). We provide an exponential-size linear program formulation that computes the optimal test strategy, prove several NP-hardness results on computing optimal test strategies in general, and give efficient algorithms for special cases (scored tests and single-question tests). Experiments are also provided for those proposed algorithms to show their scalability and the increase of the tester's utility relative to that of the uniform-at-random strategy. The increase is quite significant when questions have some correlation---for example, when a test taker who can solve a harder question can always solve easier questions.

Rich Sutton is old school king of RL.