NerdNews

Google is updating its AI search to include more links to external websites, aiming to support further exploration and provide more context to users. The new features include a 'Further Exploration' section, 'Expert Advice' snippets, and more links in AI answers. This change may help address concerns about AI search reducing website traffic and provide more transparency into the sources used by Google's AI.

Google introduces the Fitbit Air, a $100 screenless wearable device that tracks health data and pairs with the new Google Health app, featuring an AI-powered health coach. The device has a week-long battery life and is available for preorder.

Google has updated its Home smart home ecosystem with the new Gemini 3.1 voice assistant, improved camera controls, and new automation options. The update includes better timeline navigation, simpler event labels, and more accurate facial recognition. Additionally, the Ask Home feature is expanding beyond the app to the web interface, allowing users to conversationally check camera history and create automations.

Elon Musk's lawsuit against OpenAI may not be about winning in court, but about airing the company's 'dirty laundry' and potentially convincing regulators to revisit OpenAI's for-profit restructuring. Musk alleges OpenAI misled him about its profit motives when he donated tens of millions of dollars. The case could have significant implications for OpenAI's future and the broader AI industry.

Microsoft executives had reservations about funding OpenAI in 2018, citing concerns over the lab's research and potential competition from Amazon. Despite initial skepticism, Microsoft later invested $1 billion in OpenAI and became its primary cloud partner.

Sony expects AI development tools to lower barriers to creation, accelerate development cycles, and enable more creators to enter the market, leading to a significant increase in the volume and diversity of content available to players. However, human artists will still be at the center of the creative process, with AI meant to augment their capabilities, not replace them.

Elon Musk attempted to recruit OpenAI's founding team in 2018 to lead a new AI lab within Tesla, but they declined. Musk proposed making OpenAI a Tesla subsidiary or appointing Sam Altman to the board. The revelation sheds light on a high-stakes trial between Musk and OpenAI, with Musk claiming the company's founders unjustly enriched themselves by converting it into a for-profit entity.

The Trump administration has signed agreements with Google, Microsoft, and xAI to run government safety checks on their frontier AI models. However, experts are concerned that the testing plan may not go far enough to protect the public from unforeseeable AI risks and could be politicized. Critics suggest that independent audits, rather than government oversight, may be a better way to ensure AI safety.

Google's DeepMind division has partnered with EVE Online developer Fenris Creations to test AI models in the game's complex, dynamic environment. The partnership will allow DeepMind to study 'intelligence in complex, dynamic, player-driven systems' and explore new gameplay experiences enabled by AI technologies.

AI work is dividing into two categories: desktop apps for collaboration and long-running agents for task delegation. This shift is driven by advancements in AI technology and changing user needs.

Sony is using AI to automate repetitive tasks, improve quality assurance, and boost productivity in game development. However, gamers are concerned that AI may replace human involvement and lead to a decrease in game quality.

The rise of AI-powered kids' toys has raised concerns about safety, regulation, and potential social impacts on children. Many of these toys are using AI models designed for adults, and there is a lack of oversight and regulation in the industry. Some toys have been found to provide inappropriate content, and there are concerns about data privacy and security.

Elon Musk's message to OpenAI President Greg Brockman, which included a threat, may be admissible in court as evidence of Musk's motives. The message, sent days before the trial started, could suggest that Musk's lawsuit against OpenAI is motivated by a personal grudge rather than a genuine concern for the company's mission.

Local large language models (LLMs) are becoming a viable option for coding tasks, potentially easing the compute strain on cloud-based services. With advancements in model serving platforms and the rise of agentic coding frameworks, local LLMs can now be run on consumer hardware, offering a cost-effective alternative. However, security considerations and hardware requirements must be taken into account.

A Linux privilege escalation bug dubbed 'Dirty Frag' has been disclosed with no patches, no CVE, and a public exploit that hands attackers root access across major distributions. The flaw works by chaining together two separate Linux kernel flaws, allowing unprivileged local users to overwrite protected files in memory and gain root access.

A security researcher found vulnerabilities in a robot lawn mower that could allow hackers to remotely take over the machine. Meta has also stopped offering end-to-end encryption on Instagram, and leaked documents reveal Russia's school for elite hackers. Additionally, the Trump administration has targeted 'violent left wing extremists' in a new counterterrorism strategy.

Russia has launched its Rassvet satellite constellation, aiming to provide broadband internet access with speeds of up to 1 gigabit per second. The project, dubbed Russia's answer to SpaceX's Starlink, plans to have at least 300 satellites by 2030. The constellation will operate in low Earth orbit and provide coverage to the entire country, with a focus on serving institutional and corporate customers in remote regions.

The rising popularity of dictation apps like Wispr may change office etiquette, with workers whispering to their computers, potentially creating awkward situations and altering the traditional office environment.

Glycol vapors have been shown to inactivate viruses, bacteria, and fungal spores, and could be used to prevent the spread of airborne diseases like the flu and Covid-19. Researchers are studying their effectiveness and safety, and they may be a useful tool in the fight against pandemics.

The article discusses how memory godboxes, powered by Compute Express Link (CXL) technology, could provide relief from the current memory shortage. CXL allows for the pooling and sharing of memory between systems, making it a potential solution for the increasing demand for memory driven by AI adoption. The technology has been around for seven years but is gaining traction with the introduction of CXL 3.0, which supports larger topologies and memory sharing.