This week, Elon Musk's X, formerly Twitter, saw its artificial intelligence chatbot Grok go Nazi. In the past three years of Musk's ownership of the social network, it feels like X has weathered at least one public crisis per week, more often multiple. Last week, Musk's artificial intelligence firm, xAI, saw its flagship chatbot Grok declare itself a super-Nazi, referring to itself as "MechaHitler". It made racist, sexist and antisemitic posts, which the company deleted. One example, via my colleague Josh Taylor: Grok referred to a person with a common Jewish surname as someone who was "celebrating the tragic deaths of white kids" in the Texas floods as "future fascists".

Jumping and hopping locomotion are efficient means of traversing unstructured rugged terrain with the former being the focus of roboticists. This focus has led to significant performance and understanding in jumping robots but with limited practical applications as they require significant time between jumps to store energy, thus relegating jumping to a secondary role in locomotion. Hopping locomotion, however, can preserve and transfer energy to subsequent hops without long energy storage periods. Therefore, hopping has the potential to be far more energy efficient and agile than jumping. However, to date, only a single untethered hopping robot exists with limited payload and hopping heights (< 1 meter). This is due to the added design and control complexity inherent in the requirements to input energy during dynamic locomotion and control the orientation of the system throughout the hopping cycle, resulting in low energy input and control torques; a redevelopment from basic principles is necessary to advance the capabilities of hopping robots. Here we report hopping robot design principles for efficient and robust systems with high energy input and control torques that are validated through analytical, simulation, and experimental results. The resulting robot (MultiMo-MHR) can hop nearly 4 meters (> 6 times the current state-of-the-art); and is only limited by the impact mechanics and not energy input. The results also directly contradict a recent work that concluded hopping with aerodynamic energy input would be less efficient than flight for hops greater than 0.4 meters.

Liftoff is the leading growth acceleration platform for the mobile industry, helping advertisers, publishers, game developers and DSPs scale revenue growth with solutions to market and monetize mobile apps. Founded in 2012 and headquartered in Redwood City, CA, Liftoff has a diverse, global presence. Looking for an engineer to build machine learning models with our data science team. Liftoff is committed to providing and maintaining a work environment where all employees and candidates are treated with dignity and respect and that is free of bias, prejudice, and harassment. Liftoff is further committed to providing an equal employment opportunity for all employees and candidates for employment free from discrimination and harassment on the basis of sex, gender (including sexual harassment, gender harassment, and harassment due to pregnancy, childbirth, breastfeeding, and related conditions), sexual orientation, gender identity, gender expression, gender nonconformity, race, creed, religion, color, national origin, ancestry (including association, affiliation, or participation with persons or activities related to national origin, English-proficiency or accent, or immigration status), physical or mental disability, medical condition(s), genetic information of an individual or family member of the individual, marital or domestic partner status, age, veteran or military status, family care status, requesting or taking pregnancy, parental or disability leave, requesting an accommodation, or any other characteristic protected by federal, state, or local law, regulation, or ordinance.

Liftoff is the leading growth acceleration platform for the mobile industry, helping advertisers, publishers, game developers and DSPs scale revenue growth with solutions to market and monetize mobile apps. Founded in 2012 and headquartered in Redwood City, CA, Liftoff has a diverse, global presence. With our success over the past 10 years, we are ready to move beyond our current technologies and build the next generation ML Platform to align with our current business needs as we continue to grow. Currently using a custom setup, we handle 1.5 billion ML inferences per second, every day. We are now building a new ML platform to support neural networking in order to increase the intelligence of our models.

Yunshi Zhao is a Machine Learning Engineer at Liftoff, a mobile app optimization platform for marketing and monetizing apps at scale. Her responsibilities range from researching and training models to deployment and monitoring models in production. She is also part of the diversity, equity, and inclusion (DEI) committee at Liftoff, focusing on representation in engineering. Before transitioning to startup life, she worked as a data scientist and aerospace engineer. Here, she talks about machine learning development, best practices, use cases, and ML in production.

Access to Spectrum's Digital Edition is exclusive for IEEE Members Create an account to access more content and features on IEEE Spectrum, including the ability to save articles to read later, download Spectrum Collections, and participate in conversations with readers and editors. For more exclusive content and features, consider Joining IEEE. This article is for IEEE members only. Join the world's largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum's articles, podcasts, and special reports. Join the world's largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum's articles, archives, PDF downloads, and other benefits.

A replica of the Mars 2020 Perseverance Rover is shown during a press conference ahead of the launch of a United Launch Alliance Atlas V rocket carrying the rover, at the Kennedy Space Center in Cape Canaveral, Florida, U.S. July 29, 2020. REUTERS/Joe SkipperThe $2.4 billion mission, slated for liftoff at 7:50 a.m. ET (1150 GMT) from Florida's Cape Canaveral, is planned as the U.S. space agency's ninth trek to the Martian surface. The United Arab Emirates and China separately this month launched probes to Mars in displays of their own technological prowess and ambition. Launching atop an Atlas 5 rocket from the Boeing-Lockheed (BA.N) (LMT.N) joint venture United Launch Alliance, the car-sized Perseverance rover is expected to reach Mars next February.

NASA is set to launch an ambitious mission to Mars on Thursday with the liftoff of its next-generation Perseverance rover, a six-wheeled robot tasked with deploying a mini helicopter, testing out equipment for future human missions and searching for traces of past Martian life. The $2.4 billion mission, slated for liftoff at 7:50 a.m. from Florida's Cape Canaveral, is planned as the U.S. space agency's ninth trek to the Martian surface. The United Arab Emirates and China separately this month launched probes to Mars in displays of their own technological prowess and ambition. Launching atop an Atlas 5 rocket from the Boeing-Lockheed joint venture United Launch Alliance, the car-sized Perseverance rover is expected to reach Mars next February. It is due to land at the base of an 820-foot-deep (250 meters) crater called Jezero, a former lake from 3.5 billion years ago that scientists believe could hold traces of potential past microbial Martian life.

SpaceX launched a three-ton shipment to the International Space Station on Thursday, including "mighty mice" for a muscle study, a robot sensitive to astronauts' emotions and a miniature version of a brewery's malt house. The Dragon capsule also is delivering holiday goodies for the six station residents. NASA's Kenny Todd isn't giving any hints, but said, "Santa's sleigh, I think, is certified for the vacuum of space." The recycled capsule should arrive Sunday. The Falcon rocket blasted off from Cape Canaveral a day late because of high winds.

Open access to high-quality education is limited by the difficulty of providing student feedback. In this paper, we present Generative Grading with Neural Approximate Parsing (GG-NAP): a novel approach for providing feedback at scale that is capable of both accurately grading student work while also providing verifiability--a property where the model is able to substantiate its claims with a provable certificate. Our approach uses generative descriptions of student cognition, written as probabilistic programs, to synthesise millions of labelled example solutions to a problem; it then trains inference networks to approximately parse real student solutions according to these generative models. We achieve feedback prediction accuracy comparable to professional human experts in a variety of settings: short-answer questions, programs with graphical output, block-based programming, and short Java programs. In a real classroom, we ran an experiment where humans used GG-NAP to grade, yielding doubled grading accuracy while halving grading time.