NerdNews

Autodesk's Wonder 3D AI generator creates detailed 3D assets from text prompts and 2D images, aiming to open game art and 3D printing to everyone, with potential uses in game design, 3D printing, and virtual production.

Microsoft's threat intel team reports that AI agents are being used by cybercriminals, including North Korea, to manage attack infrastructure and perform reconnaissance, making attacks more efficient and faster.

Mozilla has partnered with Anthropic to use its AI-powered bug detection system, which has already identified 14 high-severity bugs in Firefox. However, Mozilla engineers note that around 10% of browser crashes are caused by hardware issues, such as bit flips, rather than software errors.

The economics of AI inference at scale are complex, involving factors like token throughput, user interactivity, and hardware/software optimization. Companies like Nvidia and AMD are working to improve inference efficiency, with techniques like disaggregated compute and software optimization. The goal is to generate the most desirable tokens at the lowest cost, with the market becoming increasingly competitive.

Anthropic, an AI giant, has been designated a supply chain risk to national security by the US Department of War. The company is suing the government, calling the decision 'legally unsound'. The designation bars Anthropic from securing military contracts and is the first time a US company has been classified this way.

Oracle and OpenAI have reportedly abandoned plans to expand their Texas Stargate datacenter due to financing issues and forecasting difficulties. Meta is now considering leasing the untapped capacity, brokered by Nvidia, which had put down a $150 million deposit.

Researchers from Anthropic found that AI has had almost no impact on jobs considered to be 'exposed' to automation, with no systematic increase in unemployment for highly exposed workers since late 2022. They propose a new measurement, 'observed exposure', to clarify the actual economic impact of AI on workers.

Google Cloud announced a preview of global queries for BigQuery, allowing developers to run SQL queries across data stored in different geographic regions without moving or copying the data. This feature simplifies analytics for companies with distributed datasets and lets them control where the query runs.

Google Threat Intelligence Group reports 90 zero-day vulnerabilities in 2025, with 48% targeting enterprise software and appliances, highlighting a shift towards exploiting enterprise infrastructure for privilege escalation and broad impact.

Maxon is releasing Cinema 4D on iPad with an AI 3D model generator powered by Tencent's HY 3D AI engine, allowing artists to generate base 3D and UV models from text prompts or image references.

The US Department of Commerce is drafting rules to restrict exports of AI chips, requiring companies like Nvidia and AMD to obtain government approval before selling abroad. This move aims to control the supply of AI compute and drive investment back into the US.

A House of Lords committee has warned that weakening AI copyright law could harm the UK's creative industries, which contribute £124 billion to the economy. The committee argues that ministers should not trade the sector's interests for promises of future tech growth.

OpenAI CEO Sam Altman signed a $200M deal with the Department of Defense, despite previously stating that AI should not be used for mass surveillance or autonomous lethal weapons. The deal has sparked controversy and criticism, with some accusing OpenAI of prioritizing profits over ethics.

Google has released a new command-line tool that integrates OpenClaw into Workspace data, making it easier to use multiple Workspace APIs. The tool, called Google Workspace CLI, is not an official Google product yet, but it offers a lot of features, including support for structured JSON outputs and over 40 agent skills. It can be used to load and create Drive files, send emails, create and edit Calendar appointments, and more.

Simile, an AI company inspired by The Sims, aims to revolutionize public opinion research by using digitally simulated people. The company uses AI agents trained on chat-style interviews with actual people to create 'digital twins' that can be polled for market insights. Simile has partnered with companies like CVS and Gallup to simulate consumer behavior and provide more accurate market research.

The article discusses the concept of 'Claws', personal AI assistants that can learn and adapt to individual needs. It highlights the potential of Claws to revolutionize the way we work and interact with technology, and explores the possibilities of building trust with these AI entities.

Caitlin Kalinowski, OpenAI's robotics hardware lead, has resigned due to the company's deal with the Department of Defense, citing concerns over surveillance and lethal autonomy without proper guardrails. OpenAI has stated that it does not support these issues and will continue to engage in discussions.

Rspress 2.0 introduces a redesigned theme, faster startup, and AI-native documentation features, including Static Site Generation to Markdown and improved build performance with lazy compilation and persistent caching.

The Pentagon's controversy with Anthropic over the use of its AI technology may give pause to startups considering working with the federal government, particularly the Department of Defense. The dispute highlights the challenges of balancing the need for technological advancements with concerns over the ethical use of AI in military applications.

Rivian aims to sell 20,000-25,000 R2 SUVs in 2026, with production starting in June. The company's sales rate will outpace most comparable electric vehicles if it hits its target. Rivian's tactical move to reach this goal is unclear, but it may involve various marketing strategies to turn interest into sales.