-
Volantis develops photonic links to expand AI chip memory capacity
Volantis, a Sam Altman-backed AI chip startup, is developing an accelerator using photonic interconnects to overcome a fundamental constraint in AI chip design. By using light-based connections instead of electrical wires to link memory to processors, the company claims it can dramatically increase the amount of high-bandwidth memory available to large language models. This tackles…
-
Pentagon ends use of Anthropic’s Claude tools after safety dispute
The US Pentagon has officially ceased using Anthropic’s Claude AI tools, an official confirmed to the BBC, following months of contention with the company over AI safety features. Defence Secretary Pete Hegseth designated Anthropic a national security risk in February after the company refused to remove safety guardrails from its systems, citing concerns about mass…
-
F1 in Bahrain… in Malaysia: The race needed a software update to start
The 2026 Bahrain Grand Prix was relocated to Malaysia’s Sepang circuit due to the ongoing conflict with Iran, where torrential rain caused significant delays before the race start. A previously undetected software bug in the cars’ hybrid power systems subsequently malfunctioned during formation laps behind the safety car, triggering automatic 60-second shutdowns of the electric…
-
OpenAI to test image generation adverts in ChatGPT in October
OpenAI plans to test visual adverts in ChatGPT, initially while users generate images. The move expands the company’s advertising in the chatbot and could give brands a new way to promote products, while raising questions about how the adverts will appear and whether they will interrupt the experience. The test is due to begin in…
-
Nvidia raises Shield TV Pro price by $100 amid memory shortage
Nvidia has raised the price of its Shield TV Pro streaming box by $100, from $199.99 to $299.99, citing higher component costs linked to a memory shortage. The increase, which took effect on 2 October, makes the nearly seven-year-old device substantially more expensive for buyers. Nvidia has also stopped producing the standard Shield TV, which…
-
Debian kernel advisory lists 1,313 CVEs, but tally does not gauge risk
Debian’s latest Linux kernel security update lists 1,313 CVE identifiers, highlighting the growing volume of reported and fixed issues that maintainers must process. The article suggests AI-assisted bug hunting may be contributing to the workload, while stressing that the large tally alone does not show how severe or exploitable the issues are. The advisory, DSA-6528-1,…
-
Poland accuses Google of withholding data in publisher payment talks
Poland’s Office of Competition and Consumer Protection (UOKiK) has accused Google of unfair market practices in how it handles articles from local media publishers. The regulator claims Google failed to provide necessary data during compensation negotiations, improperly leveraging its dominant position. This case highlights growing antitrust concerns across Europe regarding how tech giants compensate content…
-
Here’s why it’s important to keep your phone apps updated
App updates receive constant complaints and are frequently ignored, but they serve critical functions beyond cosmetic changes. Users should keep apps current not just for new features and interface improvements, but primarily because updates patch security vulnerabilities that could expose devices and personal information to attack. Updates also deliver performance enhancements, bug fixes, and ensure…
-
HPE and NVIDIA executives to discuss AI data sovereignty at Washington dinner
The Register is hosting an off-the-record dinner in Washington on 29 October bringing together senior executives from HPE and NVIDIA with industry leaders to discuss AI data sovereignty. The event uses the Chatham House Rule, which means discussions can be shared without attributing statements to speakers, allowing frank conversation about the practical realities of AI…
-
Lola Vision develops software to run AI models on specialised chips
Lola Vision Systems, founded by Tayo Adesanya in 2024, is developing software to make it easier to run AI models on specific chips, as well as designing its own chips. The company aims to shorten the lengthy setup process and help device makers run models more reliably and with less power, particularly in fields where…