NerdNews

Elon Musk's lawsuit against Sam Altman will go to trial, with Musk accusing OpenAI of straying from its founding nonprofit mission to ensure AGI benefits humanity. The trial may reveal more about OpenAI's core people and could impact the company's plans to file for an IPO.

Salesforce has announced a new AI initiative called Headless 360, which aims to integrate AI into every aspect of its product, allowing users to automate businesses' relationships with customers and potentially never log in again. The initiative includes over 100 new tools and skills for developers and marks a significant shift in the company's approach to enterprise software.

The article discusses how AI executives, such as OpenAI's Sam Altman, have been warning about the dangers of AI, including existential risks, while also promoting its benefits and lobbying for regulation. However, their rhetoric has been criticized for being inconsistent and potentially inciting violence, as seen in recent attacks on Altman's home. The article argues that the AI executives have 'painted themselves into a corner' by highlighting the risks of AI and now trying to downplay them.

OpenAI has been making headlines with its recent acquisitions, including a personal finance startup and a business talk show. These moves may be attempts to address existential problems, such as creating a sustainable business model and improving its public image. The company is also facing competition from Anthropic, particularly in the enterprise sector.

The Verge's Installer newsletter discusses new AI apps, including OpenAI Codex, an all-in-one AI superapp, and Gemini for Mac, a new way to interact with Google's AI. Other topics include new action cameras, Android weather apps, and a sci-fi action game.

Anthropic has launched Claude Design, a new AI-powered design tool that allows users to create polished visuals like slide decks, app prototypes, and marketing one-pagers using simple text prompts. The tool is powered by Claude Opus 4.7 and is rolling out as a research preview to Claude Pro, Max, Team, and Enterprise subscribers. Wall Street seems to think it could be a serious threat to Figma and other software, as Figma's stock fell about 7% following the announcement.

OpenAI's investors are uncertain about Sam Altman's fitness to lead the company as it prepares to go public, citing his divided attention among various side projects, including Helion Energy and Stoke Space. The company's board is reportedly considering alternative leadership, with Bret Taylor, the current board chair and former Salesforce co-CEO, as a potential candidate.

AI vendors are passing the buck on security vulnerabilities, saying they are 'working as intended' or 'expected behavior'. This lack of accountability is a concern, especially with the increasing use of AI in corporate IT environments. Recent examples include Anthropic, Google, and Microsoft, who have all been notified of security flaws in their AI systems but have not taken adequate action to fix them.

Cloudflare introduces Agent Memory, a managed service that stores AI conversation data, allowing AI models to recall important information and forget irrelevant details, making the most of available context space and improving overall performance.

Commercial AI models have made significant progress in vulnerability research, with some models able to find and exploit vulnerabilities without complex prompts. Forescout's testing found that the most capable models, such as Claude Opus 4.6 and Kimi K2.5, can now perform tasks that were previously impossible for AI models, making them accessible to inexperienced attackers.

Software platforms must be treated as products, not just code, to deliver value and balance competing priorities such as engineering, design, usability, security, and value for internal customers and the organisation. A product mindset, clear ownership, and continuous investment can prevent bottlenecks, platform decay, and wasted effort.

The use of AI to write stories is becoming more prevalent in newsrooms, with some reporters using AI to generate drafts. However, this trend is raising concerns about the role of human writers and the potential loss of the human voice in journalism. While AI can assist with research and editing, it is unclear whether it can truly replace the creativity and nuance of human writing.

Nearly 40% of US data center projects may miss completion dates due to chronic shortages of labor, power, and equipment, with major tech companies like Microsoft and OpenAI affected, and growing local resistance to data center development

Meta's massive AI infrastructure investments, totaling $115-135 billion, are driving up prices for RAM and other components, resulting in a $50-100 price increase for Quest headsets, a 12-20% rise. The company's shift in focus from metaverse to AI is contributing to the price surge.

Uber has committed over $10 billion to buying autonomous vehicles and taking equity stakes in companies developing the technology. This marks a new era for the company, focusing on owning physical assets rather than developing technology in-house. Uber's investments include companies like WeRide, Lucid, and Nuro, with $2.5 billion in direct investments and $7.5 billion to be spent on buying robotaxis over the next few years.

NodeWeaver is positioning itself as a cost-effective alternative to VMware for edge computing workloads, with a perpetual licensing model and the ability to run on any x86 hardware. The company's platform can reduce costs by 60-80% and is being used by large enterprises such as Carnival Corporation.

A ransomware attack on Cookeville Regional Medical Center in July 2025 compromised the personal and medical data of over 337,000 patients. The attack, claimed by the Rhysida ransomware-as-a-service operation, accessed or acquired files between July 11 and July 14, 2025. The hospital began mailing breach notification letters in April 2026 and is offering 12 months of free identity theft protection.

World, a company founded by Sam Altman, has partnered with Tinder to allow users to verify their humanity by scanning their iris with an Orb device. This verification process creates a unique cryptographic key, or World ID, which can be used to prove a user's identity on the platform. The partnership aims to address the issue of bots and fake accounts on dating apps, and World has also announced other consumer and enterprise partnerships, including with Zoom and Docusign.

Vercel, a cloud development platform, has confirmed a security breach after hackers claimed to have stolen data. The breach occurred due to the compromise of a third-party AI tool's Google Workspace OAuth application, allowing hackers to access internal Vercel systems and steal environment variables. Vercel is advising customers to review their environment variables and enable sensitive variable features to ensure encryption at rest.

Google and Cloudflare have accelerated their deadlines for post-quantum cryptography (PQC) readiness to 2029, citing recent research that suggests quantum computers may be able to break current encryption algorithms sooner than expected. Other companies, such as Amazon and Microsoft, have longer timelines, but experts warn that denial and forgotten software dependencies could delay the transition and put the world at risk of repeating past security lapses.