Artificial Intelligence: News, Business, Science

AI research

Dec 13, 2025Dec 13, 2025

More AI agents isn't always better, new Google and MIT study finds

AI in practice

Dec 11, 2025Dec 11, 2025

GPT-5.2 lands to top Google's Gemini 3 in the AI benchmark game just four weeks after GPT-5.1

AI research

Dec 7, 2025

Corporate AI agents use simple workflows with human oversight instead of chasing full autonomy

AI research AI and society AI in practice

AI in practice

Dec 16, 2025Dec 16, 2025

Short

Nvidia is taking over software provider SchedMD to expand its presence in open-source technology. On Monday, the company confirmed it will continue to distribute SchedMD's "Slurm" software as an open-source product. The platform helps plan large-scale computing tasks in data centers, ensuring server capacity is used efficiently.

Nvidia views the technology as critical infrastructure for generative AI, noting that developers rely on it to train models. Financial terms of the deal were not disclosed. Founded in California in 2010, SchedMD employs around 40 people and serves clients like cloud provider CoreWeave and the Barcelona Supercomputing Center.

Nvidia

AI in practice

Dec 16, 2025Dec 16, 2025

Short

Google is linking its NotebookLM research tool directly to the Gemini chatbot. This integration lets users select specific notebooks as context for their Gemini queries, effectively expanding the chatbot's knowledge base beyond its initial training data and standard web results. While NotebookLM already includes a built-in chat function powered by a Gemini model, it remains quite limited—it doesn't even save chat histories. The new feature addresses this by allowing users to leverage multiple notebooks simultaneously within the main Gemini interface. It also supports integration with "Gems," the personalized versions of the chatbot. The rollout appears to be gradual, starting with browser users, though app support will likely follow soon.

NotebookLM started as an experimental tool in 2023. It has since established itself as a software with exemplary AI integration, particularly in the education sector. The tool makes it easy to set up RAG environments and thus make large document collections analyzable and searchable. Google regularly adds new functions to NotebookLM, most recently including one for deep research.

via X

AI and society

Dec 16, 2025Dec 16, 2025

Nearly half of US workers now use AI on the job, but most aren't using it daily

AI research

Dec 15, 2025Dec 15, 2025

The case against predicting tokens to build AGI

AI in practice

Dec 15, 2025Dec 15, 2025

Short

Robot vacuum pioneer iRobot has filed for bankruptcy and plans to hand control to its main Chinese supplier, Shenzhen PICEA Robotics. According to Bloomberg, shares of the Roomba maker will be wiped out under the bankruptcy plan. While the company will be delisted, it intends to continue operations as a going concern.

To set up the deal, Shenzhen PICEA acquired $191 million of iRobot's debt from the Carlyle Group. iRobot attributed the filing to a post-pandemic sales slump, supply chain issues, and stiffer competition from cheaper rivals. The move comes after a planned acquisition by Amazon fell apart in 2022 following opposition from EU regulators.

The company listed assets and liabilities between $100 million and $500 million. In a statement, iRobot confirmed it would continue paying employees and suppliers throughout the court proceedings.

Bloomberg

AI and society

Dec 15, 2025Dec 15, 2025

Short

Trump attempts to block state AI laws by withholding broadband billions, but faces shaky legal ground.

"I think the administration has a 30 to 35% chance of this working legally," says Dean Ball, a former White House official who contributed to the administration's AI Action Plan.

The executive order directs the Commerce Department to block states with onerous AI regulations from the $42 billion Broadband Equity, Access, and Deployment program (BEAD), reports Reuters in an analysis of the new order. However, experts doubt whether Congress intended to give the administration authority over state AI regulation when it authorized broadband funding. Furthermore, the move risks political blowback from within the party: Republican governors like Ron DeSantis have previously spoken against federal interference, and withholding funds would impact rural voters—a key demographic that supported Trump by wide margins.

Reuters

AI research

Dec 15, 2025Dec 14, 2025

AI models score off the charts on psychiatric tests when researchers treat them as therapy patients

AI research

Dec 14, 2025Dec 14, 2025

Reasoning models now ace all three CFA exam levels

AI and society

Dec 14, 2025Dec 14, 2025

Short

The Center for Humane Technology (CHT), a nonprofit organization advocating for ethical technology, has criticized a new executive order from the Trump administration that aims to undermine state AI laws.

According to the CHT, the regulation puts public safety at risk by preventing states from meaningfully regulating AI. At the same time, it offers no national replacement framework, creating what the organization calls a vacuum in accountability.

Americans understand the potential benefits and dangers of this technology. They believe government should help regulate AI, not provide a regulatory shield to an industry that prioritizes growth at any cost. (CHT)

The CHT points to documented AI harms, including deepfakes, fraud, and chatbot-related suicides among young people. Social media already showed what happens when technology goes unregulated, the organization argues. The government should protect the public instead of caving to the tech industry.

Trump argues that varying state regulations are slowing down the industry. AI companies like Anthropic, OpenAI, and Google support national regulation.

CHT

AI and society

Dec 14, 2025Dec 14, 2025

Short

Deepmind co-founder Shane Legg puts the odds of achieving "minimal AGI" at 50 percent by 2028. In an interview with Hannah Fry, Legg lays out his framework for thinking about artificial general intelligence. He describes a scale running from minimal AGI through full AGI to artificial superintelligence (ASI). Minimal AGI means an artificial agent that can handle the cognitive tasks most humans typically perform. Full AGI covers the entire range of human cognition, including exceptional achievements like developing new scientific theories or composing symphonies.

Legg believes minimal AGI could arrive in roughly two years. Full AGI would follow three to six years later. To measure progress, he proposes a comprehensive test suite: if an AI system passes all typical human cognitive tasks, and human teams can't find any weak points even after months of searching with full access to every detail of the system, the goal has been reached.

YouTube

Google News

Bank details