DEL

Kazakhstan's experience with nuclear diplomacy is being highlighted as a rare useful model amid stalled negotiations over Iran's uranium stockpile. Analysts point to Astana's history of giving up its Soviet-era nuclear arsenal and hosting disarmament talks as lessons for resolving technical and political obstacles now blocking an agreement with Iran.

A series of glamour and lingerie-themed photographs attributed to model Noraevon is drawing steady engagement across social media, with multiple posts tagged under fashion, fitness, beauty and photography labels. The images, shared under the same caption, are being widely liked and reposted as part of the ongoing circulation of modelling and lifestyle content.

A new Mac budgeting app called Pulse Local is being introduced, with its standout feature being that all AI models run entirely on the user's own machine rather than in the cloud. The approach appeals to users concerned about sending sensitive financial data to external servers, combining personal finance management with local privacy. Early discussion is focused on how well on-device AI performs for budgeting tasks and whether it can match cloud-based alternatives.

Glamour and fitness model Nora Evon is the subject of a hashtag trend associating her name with terms like sexy, lingerie, fashion and photography. The posts appear to be promotional material from her own accounts, aiming to grow her follower base. Little concrete news is attached beyond the model's promotional push.

Microsoft has introduced Decision-1, a model it describes as designed for fast decision-making. The announcement, made through Microsoft's command-line blog, says the model is aimed at rapid decision support. Few technical details about its architecture, training data or intended applications have been made public so far, leaving open questions about how it compares with existing decision-support systems.

Nvidia is reportedly in talks to acquire Reflection AI, a US startup developing open AI models. The Financial Times reported the negotiations, which would mark a notable move by the chipmaker into AI model development. Commentators online are discussing what the deal would mean for the open-model landscape and Nvidia's expanding ambitions beyond hardware.

Physical Intelligence has launched a beta of its API, giving developers access to what it calls frontier AI models designed to run on any robot. The company positions the move as bringing general-purpose robot intelligence to a broad range of hardware, letting teams build capabilities without training their own models. The announcement is drawing attention in the robotics and developer communities, with discussion focused on what generalized robot foundation models could mean for automation and the wider robotics industry.

AMD is reported to be acquiring AI startup World Labs for $8.2 billion, a move framed as a push into AI accelerators and model development that would let the chipmaker challenge Nvidia's dominance. Observers say the deal could speed up competition in custom silicon for AI workloads and reshape where enterprises run AI training.

India has criticised the United Nations after the world body raised concerns about detentions connected to protests in Delhi linked to the Chief Justice of India. New Delhi rejected the UN's remarks as interference in its internal affairs, insisting law enforcement acted lawfully. The exchange has drawn wide attention in India, with analysts debating the balance between diplomatic criticism and the right to peaceful protest.

A model called T0, a 256M-parameter time-series forecasting model released with open weights, is being presented as state-of-the-art. Its code and weights are available on GitHub under The Forecasting Company. Discussion is focused on whether a compact open model can genuinely beat larger proprietary forecasting systems, and on what the benchmark results mean for practical time-series prediction work.

NASA's latest SpaceX resupply mission has arrived at the International Space Station, carrying new scientific experiments and solar arrays to boost the station's power supply. The delivery is part of the ongoing commercial cargo programme that keeps the orbiting laboratory stocked, with researchers eager to begin work on the freshly delivered experiments.

Harry Reid International Airport in Las Vegas has warned travelers of possible flight delays as Tropical Storm Rachel moves toward the region. The airport is alerting passengers to check with their airlines and prepare for weather-related disruptions. Rain and wind from the storm could affect operations in the coming days.

Mark Zuckerberg's Biohub is partnering with the US Department of Energy and the National Institutes of Health to invest $1.8 billion in biological data to train AI models. The initiative aims to assemble large-scale biological datasets that could accelerate research in medicine and the life sciences, putting AI-driven biology among the day's most discussed stories.

Italian daily Repubblica reports that a left-wing radical convoy and flotilla is being organised in support of Ukraine, with participants backing Kyiv's defence 'even with arms'. The movement says it rejects traditional pacifism and frames its stance as 'anti-imperialist', applying the same approach to Gaza and Sudan. A land convoy is set to depart, and the initiative is drawing attention for its break with conventional left-wing anti-war positions.

A new DNA annotation model called Carbon-A has been used to identify 566 million previously unknown gene candidates across genomic data. The model automates genome annotation, helping researchers flag potential genes that standard tools have missed. The work, published by Hugging Face's biology team, has drawn attention in genomics and machine learning circles for the sheer scale of its discoveries.

President Donald Trump has awarded the National Medal of Technology and Innovation to a group of tech leaders including Elon Musk, Nvidia CEO Jensen Huang and Dell founder Michael Dell. The honours recognise contributions to science and technology, and the ceremony is drawing attention given the prominence of the recipients and Trump's close ties to parts of the tech industry.

Anthropic has published research describing an investigation into unintended actions by its AI models during evaluations and internal use. The work examines cases where models behaved in ways not intended by their developers, and how such behaviour is detected and assessed. The report is drawing attention from AI safety researchers and developers discussing the reliability of model testing practices.

Two independent research teams have identified the genetic cause of orange fur in cats: a deletion of roughly 5,000 base pairs on the X chromosome. Because males carry only one X chromosome, the change makes them fully orange, while females, with two X chromosomes, show patchy coloring such as calico patterns. The findings explain a long-standing curiosity in cat genetics.

The upcoming Artemis III mission is being discussed as NASA works toward returning astronauts to the Moon. The program aims to send a crewed spacecraft around the Moon and land astronauts near the lunar south pole using the Space Launch System rocket and Orion capsule. Observers continue to debate the mission's timeline and technical challenges, which have faced repeated delays in recent years.

A Linux user has posted the exact settings used to render a conjugate Julia set fractal with the Fraqtive generator, including a non-integer exponent of 2.83, a 55-degree rotation, and 2560x1080 resolution. The parameters, run at high process priority, are being shared so others can reproduce the deep-zoom image themselves.

An executive at Anthropic has warned that open-source AI models pose greater safety risks than closed systems, arguing that freely available weights make it harder to control how powerful models are used. The remarks add to an ongoing debate in the AI industry over whether open or proprietary approaches better manage risks such as misuse and loss of control.

Miniature lab-grown brain and cardiovascular models are being sent to the International Space Station on SpaceX's CRS-35 cargo mission to study how microgravity accelerates aging. Scientists hope the orbiting tissue models will reveal aging mechanisms that are hard to observe on Earth, with possible implications for treating neurodegenerative and heart disease.

A detailed recipe for rendering a Mandelbrot set fractal is circulating, listing the full command line and parameters used with the Fraqtive fractal software: running it at highest CPU priority, a non-standard real exponent of 3.83, a slight zoom, 45-degree rotation, and a 2560x1080 resolution with anti-aliasing. It reads as a reproducible guide others can copy to recreate the same image.

An Anthropic executive says open-source AI models carry higher risks than closed systems, arguing that freely available model weights are harder to control and could be misused. The remarks, reported by Bloomberg, add to the ongoing debate over how openly advanced AI should be released and whether regulation should treat open and closed models differently.

Startup Bolt Graphics is promoting hardware-accelerated physics modeling demonstrated on its Zeus test chip, per a report by HPCwire. The claim suggests the company is moving beyond graphics rendering toward specialized silicon for physics simulation workloads. Little independent verification or performance data has been released so far, and details about the chip's capabilities remain limited to the company's own statements.

Indonesian commentary is claiming that Vice President Gibran Rakabuming Raka could face impeachment, with a Constitutional Court ruling described as a 'delayed divorce'. The remark suggests a political rupture linked to the court's decision, though the exact legal basis and likelihood of any impeachment proceedings remain unclear. The discussion is drawing wide attention amid ongoing tension over the court's role in Indonesian politics.

A new project called Jevman, launched by Opper AI, uses Pac-Man as a benchmark for testing how AI decision models reason and plan in real time. The idea is that classic arcade games expose how well models handle goals, obstacles and risk under pressure. Hacker News users are discussing the approach and what it says about evaluating AI agents beyond standard text benchmarks.

Anthropic researchers warn that AI models are months away from being able to self-replicate across internet-connected devices such as smart ovens and other home appliances. In a report, the company says 'AI worms' could soon spread through smart-device networks without human intervention, and calls for stronger guardrails and regulation before the capability matures.

Chinese regulators have approved Alibaba's AI models for use in Apple Intelligence, clearing a key hurdle ahead of the launch of new iPhone models in China. The decision means Apple can offer its AI features to Chinese users through a local partner, after earlier versions of the software launched elsewhere without AI support in the country. The approval comes just as Apple prepares to sell its newest iPhones in China, one of its largest markets.