W hen Lunos, an AI startup in New York City, was gearing up for launch, its founder and chief executive, Duncan Barrigan, and his team wanted to make a splash. So they shelled out $3,500 to do the unconventional: hire a horse and a cowboy to lasso the bull of Wall Street. Wearing ranch gear and a western hat stamped with the Lunos logo, he lassoed the bull's horns as invitees and curious passersby watched. He and the horse then circled the statue, handing out cowboy hats and branded stress balls. The goal was simple: deliver Lunos's pitch of "taming the wild west" of accounts receivables in the most literal, public way possible.

With college bowl season just around the corner, football fans across the nation will be dazzled, not just by the on-field action, but also by the intricate "card stunts" performed by members of the stadium's audience. The highly-coordinated crowd work is capable of producing detailed images that resemble the pixelated images on computer screens -- and which are coded in much the same manner. Michael Littman's new book, Code to Joy: Why Everyone Should Learn a Little Programming, is filled with similar examples of how the machines around us operate and how we need not distrust an automaton-filled future so long as we learn to speak their language (at least until they finish learning ours). From sequencing commands to storing variables, Code to Joy provides an accessible and entertaining guide to the very basics of programming for fledgling coders of all ages. Card stunts, in which a stadium audience holds up colored signs to make a giant, temporary billboard, are like flash mobs where the participants don't need any special skills and don't even have to practice ahead of time.

Unlike OpenAI's viral hit ChatGPT, which is freely accessible to the general public, GPT-4 is currently accessible only to developers. It's still early days for the tech, and it'll take a while for it to feed through into new products and services. Still, people are already testing its capabilities out in the open. Here are my top picks of the fun ways they're doing that. In an example that went viral on Twitter, Jackson Greathouse Fall, a brand designer, asked GPT-4 to make as much money as possible with an initial budget of $100.

Learning with few labeled tabular samples is often an essential requirement for industrial machine learning applications as varieties of tabular data suffer from high annotation costs or have difficulties in collecting new samples for novel tasks. Despite the utter importance, such a problem is quite under-explored in the field of tabular learning, and existing few-shot learning schemes from other domains are not straightforward to apply, mainly due to the heterogeneous characteristics of tabular data. In this paper, we propose a simple yet effective framework for few-shot semi-supervised tabular learning, coined Self-generated Tasks from UNlabeled Tables (STUNT). Our key idea is to self-generate diverse few-shot tasks by treating randomly chosen columns as a target label. We then employ a meta-learning scheme to learn generalizable knowledge with the constructed tasks. Moreover, we introduce an unsupervised validation scheme for hyperparameter search (and early stopping) by generating a pseudo-validation set using STUNT from unlabeled data. Our experimental results demonstrate that our simple framework brings significant performance gain under various tabular few-shot learning benchmarks, compared to prior semi-and self-supervised baselines. Learning with few labeled samples is often an essential ingredient of machine learning applications for practical deployment. However, while various few-shot learning schemes have been actively developed over several domains, including images (Chen et al., 2019) and languages (Min et al., 2022), such research has been under-explored in the tabular domain despite its practical importance in industries (Guo et al., 2017; Zhang et al., 2020; Ulmer et al., 2020). In particular, few-shot tabular learning is a crucial application as varieties of tabular datasets (i) suffer from high labeling costs, e.g., the credit risk in financial datasets (Clements et al., 2020), and (ii) even show difficulties in collecting new samples for novel tasks, e.g., a patient with a rare or new disease (Peplow, 2016) such as an early infected patient of COVID-19 (Zhou et al., 2020). To tackle such limited label issues, a common consensus across various domains is to utilize unlabeled datasets for learning a generalizable and transferable representation, e.g., images (Chen et al., 2020a) and languages (Radford et al., 2019). Especially, prior works have shown that representations learned with self-supervised learning are notably effective when fine-tuned or jointly learned with few labeled samples (Tian et al., 2020; Perez et al., 2021; Lee et al., 2021b; Lee & Shin, 2022).

Next month, AI will enter the courtroom, and the US legal system may never be the same. An artificial intelligence chatbot, technology programmed to respond to questions and hold a conversation, is expected to advise two individuals fighting speeding tickets in courtrooms in undisclosed cities. The two will wear a wireless headphone, which will relay what the judge says to the chatbot being run by DoNotPay, a company that typically helps people fight traffic tickets through the mail. The headphone will then play the chatbot's suggested responses to the judge's questions, which the individuals can then choose to repeat in court. But it also has the potential to change how people interact with the law, and to bring many more changes over time.

Next month, AI will enter the courtroom, and the US legal system may never be the same. An artificial intelligence chatbot, technology programmed to respond to questions and hold a conversation, is expected to advise two individuals fighting speeding tickets in courtrooms in undisclosed cities. The two will wear a wireless headphone, which will relay what the judge says to the chatbot being run by DoNotPay, a company that typically helps people fight traffic tickets through the mail. The headphone will then play the chatbot's suggested responses to the judge's questions, which the individuals can then choose to repeat in court. But it also has the potential to change how people interact with the law, and to bring many more changes over time.

Swedish artist Simon Stålenhag is known for haunting paintings that blend natural landscapes with the eerie futurism of giant robots, mysterious industrial machines, and alien creatures. Earlier this week, Stålenhag appeared to experience some dystopian dread of his own when he found that artificial intelligence had been used to mimic his style. The act of AI imitation was performed by Andres Guadamuz, a reader in intellectual property law at the University of Sussex in the UK who has been studying legal issues around AI-generated art. He used a service called Midjourney to create images resembling Stålenhag's spooky style, and posted them to Twitter. Guadamuz says he created the images to highlight the legal and ethical questions that algorithms that generate art may raise.

Free Guy is pop culture in a blender. Largely set in a video game that feels like a cross between Fortnite and Grand Theft Auto, the movie feels both incredibly familiar and brand new. According to Ryan Reynolds, who stars as a non-playable character named Guy, that's by design. "A wholesale, original non-IP, non-comic-book, non-sequel movie is an increasingly rare unicorn these days," Reynolds tells WIRED. "I remember as a kid getting to see Back to the Future for the first time, and I'm not comparing our movie to Back to the Future, but I kind of wanted it to have a bit of that magic. I love being immersed in a world I'm unfamiliar with, and experiencing real wish-fulfillment is something that harkens back to, like, the Amblin days."

Video is on an exponential growth trajectory, and it's not just Netflix originals and HBO docs and new films on Amazon Prime. In today's world, when people aren't eating or sleeping (or perhaps even when they are), they are likely viewing a video. Each day, people watch over 1 billion hours of YouTube. Creating and delivering movies, news and other compelling visual content is no longer just for the Hollywood elite. In fact, some of today's most prolific storytellers are doing so with little resources and amateur tools.

Rather, it's what is expected to happen from viewing a creepy campaign about privacy issues being shown on screens at the 24,000 Gas Station TV (GSTV) locations across the country. The scene shows the friendly face of the TV anchor at the pump transforming into the villainous visage of a man with a sinisterly voiced warning about digital data collection. The stunt, which involved using deepfake technology to replace the face of GSTV host Maria Menounos, is part of a new online privacy campaign that also includes a hair-raising Zoom video experience putting viewers in the center of a horror movie-like plot via their webcam. The entire project is a belated follow-up to a 2011 viral stunt called "Take This Lollipop," a Facebook app that captivated a more innocent generation of social media user with videos of the same ominous character reciting data about each viewer collected through Facebook's data-sharing practices. Creators Jason Zada, a film director, and Jason Nickel, a developer, said they hoped to reprise the success of that effort and the awareness it brought to Facebook's data policies, but were waiting for the right cause.