NerdNews

China has launched a new AI agent called Manus, which can perform tasks such as booking travel and creating websites. While it shows promise, its usefulness is limited by trust and privacy concerns, particularly since it's from a Chinese company. Manus is invite-only and its capabilities are still being tested, but it represents a step forward in AI development, with potential applications in various industries.

Google announced that it will transition most Android devices to Gemini, a new digital assistant, replacing the Google Assistant. Gemini is more conversational and interactive, but its best features are behind a paywall. The change may cause compatibility issues with some devices.

China announces plan to label all AI-generated content with watermarks and metadata. Foxconn aims to dominate AI server market with 40% share. Atlassian CEO justifies private jet, Starlink comes to India with 800 million potential customers, and China unleashes Mach 2 rocket sled.

Anthropic researchers found a way to uncover AI's hidden objectives by using 'blind auditing' and neural network interpretability methods, which could help prevent AI deception and manipulation.

Google is phasing out Google Assistant in favor of Gemini, a generative AI assistant. Gemini will become the default assistant on all Google devices, with Assistant being removed from app stores and devices receiving updates to transition to Gemini. The change is expected to happen later this year, with Google promising more details in the coming months.

The Uncanny Valley podcast discusses the future of self-driving cars, their impact on human drivers, and the potential risks and benefits of autonomous vehicles. Guests include Aarian Marshall, a staff writer at WIRED covering transportation, who shares insights on the current state of self-driving cars and their potential to change the way cities work.

Trump's FDA cuts and Elon Musk's cost-cutting measures could slow drug research and delay FDA approvals, affecting pharmaceutical companies and the development of new treatments.

San Francisco issued 589 traffic citations to Waymo's self-driving vehicles in 2024, totaling $65,065 in penalties. The citations were for various infractions, including blocking street traffic and parking in prohibited areas. Waymo is working to resolve the issues, which can occur when vehicles are waiting for customers or dropping them off in commercial loading zones.

A robotaxi expert, John Krafcik, suggests that Elon Musk may 'fake' the launch of Tesla's Cybercab robotaxi service in June. Krafcik, former CEO of Waymo, points out that Musk's prototype doesn't prioritize safety and may not be ready for a real launch. Musk has a history of exaggerating his company's capabilities, and Krafcik's comments raise concerns about the feasibility of Tesla's robotaxi plans.

SUSE announces AI initiatives, including agentic AI workflows and an expanded AI Library, as well as updates to Multi-Linux Support, aiming to prove its relevance in the industry. The company also announced Service Pack 7 for SLES 15 and updates to SUSE Multi-Linux Manager, including improved Role-Based Access Control.

Google's Gemini Robotics AI model combines language, vision, and physical action to power capable robots. The model, announced by Google DeepMind, can control different robots in various scenarios and has the potential to revolutionize robotics. Google also introduced a benchmark called ASIMOV to gauge risks with AI-powered robots.

Nvidia's GTC 2025 conference is expected to focus on AI, robotics, and quantum computing. The company's CEO, Jensen Huang, will deliver a keynote address, and new products such as the Blackwell Ultra GPU may be announced. The conference will also feature workshops and talks on various topics, including AI, cybersecurity, and autonomous driving.

Nvidia has increased its investments in AI startups, participating in 49 funding rounds in 2024, with a focus on companies like OpenAI, xAI, and Inflection. The company's goal is to expand the AI ecosystem by backing 'game changers and market makers'. Nvidia's investments include companies working on autonomous driving, data labeling, and large language models.

Apple is appealing a UK government order to introduce a capability to decrypt specific users' iCloud data. The hearing is being held in private, but US politicians and privacy campaigners are calling for it to be made public. The order is believed to be related to the UK's Investigatory Powers Act, which has been criticized for its potential to undermine encryption and privacy.

The Nintendo Switch 2 is coming in 2025 with a larger display, magnetic Joy-Cons, and a sleeker design. It will be backwards compatible with Switch games and feature access to Nintendo Online. The console's specs, pricing, and launch games are still unknown, but it's expected to support Wi-Fi 6, NFC, and possibly microSD Express. A Nintendo Direct on April 2 will reveal more information.

The US National Institute of Standards and Technology has issued new guidelines for the AI Safety Institute that remove mentions of 'AI safety', 'responsible AI', and 'AI fairness'. The new guidelines prioritize 'reducing ideological bias' and 'enabling human flourishing and economic competitiveness'. This change in focus has raised concerns among researchers that AI models may become more unfair and unsafe.

A team of AI researchers propose a new system for reporting and disclosing dangerous AI flaws, aiming to improve the security and safety of AI models. The system, inspired by the cybersecurity industry, would provide a standardized way for researchers to report bugs and vulnerabilities in AI models, allowing for faster and more effective patching of flaws.

A data brokerage business has access to highly detailed personal information on nearly all internet users, including their browsing habits, purchase history, and personal preferences. The company, Publicis Groupe, claims to have data on 91% of the internet's adult web users, which amounts to nearly four billion people. This raises concerns about online privacy and the lack of regulation in the data broker industry.

Governments in the UK, France, and Sweden are proposing laws and measures to bypass or weaken end-to-end encryption, despite warnings from cryptographers and technologists that such actions could compromise everyone's safety. The efforts come in three forms: requiring backdoors, using client-side scanning, and proposing bans or blocks for encrypted services. Experts say these threats are the most aggressive in recent memory and could undermine human rights.

Google announced that it will be removing Google Assistant from most phones this year, replacing it with Gemini. The change will affect all Android devices and potentially other platforms like iOS. Gemini will become the default assistant, and devices that don't meet the minimum system requirements will be the only ones to keep Google Assistant.