Anthropic introduced Claude 3.5 Sonnet, a new AI language model that can compose text, analyze data, and write code. It features a 200,000 token context window and is available on the Claude website and through an API. Independent AI researcher Simon Willison praised the model, calling it 'really, really good' and stating it's the new best overall model.

Source: https://arstechnica.com/author/benjedwards/

Runway announced a new AI video synthesis model called Gen-3 Alpha, which can generate high-definition video from text prompts. The model can create 10-second-long video segments of people, places, and things, surpassing its previous model Gen-2. The company is working with a shoestring budget of compute resources.

Source: https://arstechnica.com/author/benjedwards/

SoftBank, a Japanese telecommunications giant, has been developing an 'emotion-canceling' technology powered by AI that alters the voices of angry customers to sound calmer during phone calls with customer service representatives. The project aims to reduce the psychological burden on operators suffering from harassment. The technology relies on an AI model to alter the tone and pitch of a customer's voice in real-time during a phone call, trained using a dataset of over 10,000 voice samples. The project has been in development for three years and is set to launch by March 2026.

Source: https://arstechnica.com/author/benjedwards/

A retired software engineer, Martin C. Martin, discovered a 55-year-old bug in the original Lunar Lander game's physics code. The game was created by Jim Storer, a 17-year-old high school student, in 1969. The game allows players to control a lunar module's descent onto the Moon's surface, requiring careful fuel management.

Source: https://arstechnica.com/author/benjedwards/

Wells Fargo terminated over a dozen employees from its wealth and investment management division after an investigation into claims of faking work activity on their computers. The investigation was conducted following allegations of simulation of keyboard activity to create the impression of active work. The Financial Industry Regulatory Authority (FINRA) was involved in the investigation and confirmed the terminations.

Source: https://arstechnica.com/author/benjedwards/

Source: https://arstechnica.com/author/benjedwards/

Turkish police arrested and detained a university student accused of developing an elaborate scheme to use AI and hidden devices to help him cheat on the TYT exam in Isparta. The student used a camera disguised as a shirt button connected to AI software via a router hidden in the sole of their shoe. The TYT is a nationally held university aptitude exam that determines a person's eligibility to attend a university in Turkey.

Source: https://arstechnica.com/author/benjedwards/

Stability AI released weights for Stable Diffusion 3 Medium, an AI image-synthesis model that turns text prompts into AI-generated images. The model has been criticized for generating anatomically incorrect visual abominations, particularly when rendering human limbs and bodies.

Source: https://arstechnica.com/author/benjedwards/

Apple premiered 'Apple Intelligence' at its Worldwide Developers Conference in Cupertino, announcing the integration of ChatGPT into its device OSes. The company had developed its own large language models (LLMs) for a year. AI developer Benjamin De Kr expressed confusion about Apple not training a competing LLM.

Source: https://arstechnica.com/author/benjedwards/

Source: https://arstechnica.com/author/benjedwards/