the-decoder

Sora 2 Android app launches in more countries, introduces paid generations

OpenAI has launched the Sora app for Android in the US, Canada, Japan, South Korea, Taiwan, Thailand, and Vietnam. The app is still free, but OpenAI says the service is "completely unsustainable" and ...
the-decoder

A self-rewriting AI from KAUST revives Jürgen Schmidhuber’s vision of a Gödel Machine

A research group at King Abdullah University of Science and Technology (KAUST) has unveiled the Huxley-Gödel Machine (HGM), an AI agent that can evolve by rewriting and improving its own code.
the-decoder

Microsoft and OpenAI set their own rules for AGI

Microsoft and OpenAI have decided they'll define for themselves what counts as artificial general intelligence (AGI) and when they've supposedly achieved it. The two companies say they'll appoint a ...
the-decoder

Reuters: Deepseek emerges as key AI partner in China’s military research

A Reuters analysis shows that China's People's Liberation Army is systematically integrating artificial intelligence from domestic companies such as Deepseek and Alibaba into military systems.
the-decoder

ChatGPT's memory could turn personal details into ads OpenAI CEO Altman once called dystopian

About one in five OpenAI employees is a Meta veteran, a hiring trend that's now shaping the company's culture and strategy. According to The Information, roughly 630 of OpenAI's 3,000 staffers ...
the-decoder

Junk data from X makes large language models lose reasoning skills, researchers show

Researchers have found that large language models permanently lose performance when they are continually trained on trivial online content, especially short and popular social media posts. This type ...
the-decoder

Misleading claims that most web content is now AI-generated are spreading on social media

A study making the rounds on social media claims that more than half of all web content is "created by AI instead of humans." According to the study, a piece of writing is considered "AI-generated" if ...
the-decoder

Meta AI's Yann LeCun says he played only an indirect role in the development of Llama models

Meta AI researcher Yann LeCun is distancing himself from the Llama models. In a recent post on X, LeCun said he "has not been involved in any Llama," except for a "very indirect" role in Llama 1 and ...
the-decoder

ChatGPT Atlas carries significant security risks, OpenAI warns

OpenAI's new browser, ChatGPT Atlas, could pose security risks, according to the company’s head of security, Dane Stuckey. One of the biggest issues involves so-called prompt injections. In these ...
the-decoder

Leading OpenAI researcher announced a GPT-5 math breakthrough that never happened

OpenAI researchers recently claimed a major math breakthrough on X, but quickly walked it back after criticism from the community, including Deepmind CEO Demis Hassabis, who called out the sloppy ...
the-decoder

Google Deepmind's new Gemma model helped discover a new cancer treatment pathway.

Google Deepmind and Yale University have introduced a new AI model called C2S-Scale 27B, built on the open Gemma model family. The system analyzes single cells and, according to Google Deepmind, ...
the-decoder

Japan warns OpenAI over Sora 2 after AI-generated anime videos spark copyright concerns

Japan's government has warned OpenAI not to violate copyright laws with its new video model, Sora 2, after a flood of AI-generated clips mimicking famous anime spread online. Officials said some of ...