I know a lot of people want to celebrate any stumble OpenAI makes, and rightfully so, but the imminent death of its Atlas browser isn't a sign of a company retreating from a competitive market. If you didn't catch the news yesterday, OpenAI announced, as part of the release of ChatGPT Work, that it would deprecate Atlas on August 9 . Coverage of the news treated it like OpenAI was giving up on the browser space entirely, with headlines like The ChatGPT browser is already dead and OpenAI is shutting down the ChatGPT Atlas browser only months after its release making the rounds. Naturally, people on social media followed suit. For instance, one Bluesky user Pavel took it as an opportunity to round up a number of other dead OpenAI initiatives.

Mixture-of-experts networks (MoEs) have demonstrated remarkable efficiency in modern deep learning. Despite their empirical success, the theoretical foundations underlying their ability to model complex tasks remain poorly understood. In this work, we conduct a systematic study of the expressive power of MoEs in modeling complex tasks with two common structural priors: low-dimensionality and sparsity. For shallow MoEs, we prove that they can efficiently approximate functions supported on low-dimensional manifolds, overcoming the curse of dimensionality. For deep MoEs, we show that O(L)-layer MoEs with E experts per layer can approximate piecewise functions comprising EL pieces with compositional sparsity, i.e., they can exhibit an exponential number of structured tasks. Our analysis reveals the roles of critical architectural components and hyperparameters in MoEs, including the gating mechanism, expert networks, the number of experts, and the number of layers, and offers natural suggestions for MoE variants.

Public model repositories now contain millions of models, yet most remain undocumented and effectively lost: their capabilities, provenance, and constraints cannot be reliably determined. As a result, the field wastes training time and compute, propagates hidden biases, faces intellectual-property risks, and misses opportunities for model reuse and transfer. In this position paper, we advocate charting the world's model population in a unified structure we call the Model Atlas: a graph that captures models, their attributes, and the weight transformations connecting them. The Model Atlas enables applications in model forensics, meta-ML research, and model discovery, challenging tasks given today's unstructured model repositories. However, because most models lack documentation, large atlas regions remain uncharted. Addressing this gap motivates new machine learning methods that treat models themselves as data and infer properties such as functionality, performance, and lineage directly from their weights. We argue that a scalable path forward is to bypass the unique parameter symmetries that plague model weights. Charting all the world's models will require a community effort, and we hope its broad utility will rally researchers toward this goal.

Score-based generative models are trained in high-dimensional ambient spaces, yet many data distributions are supported on low-dimensional nonlinear structures. We prove that, for compact $d$-dimensional smooth manifolds $\mathcal{M} \subset [0,1]^D$ with $d > 2$ and $β$-Hölder densities strictly positive on $\mathcal{M}$, a variance-preserving SGM estimator attains the intrinsic Wasserstein--1 sample exponent $\tilde{\mathcal{O}}(D^{\mathcal{O}_β(d)}n^{-(β+1)/(d+2β)})$, up to logarithmic factors and explicit geometry and density factors. The full nonasymptotic bound explicitly isolates the finite-order geometry envelope, Hölder radius, density lower bound, ambient dependence, and finite-order correction terms. The analysis separates score approximation into a large-noise tangent-cell regime and a small-noise projection-centered, de-Gaussianized Laplace regime. The key technical ingredient is a ReLU implementation of nearest-projection coordinates via finite intrinsic anchors and Gauss--Newton iterations, rather than approximating the manifold projection as a black-box high-dimensional smooth map. Consequently, for families with polynomially controlled geometry and density lower bounds, the constructed score-network parameters have polynomial ambient dependence.

Science Space Deep Space Fastest comet ever recorded spewed 70 Olympic pools' worth of water daily More information Adding us as a Preferred Source in Google by using this link indicates that you would like to see more of our content in Google News results. A new study of the interstellar comet 3I/ATLAS led by the University of Michigan shows that its water has a remarkably high content of deuterium. This form of hydrogen is comparatively less abundant in our solar system, enabling researchers to glean new insights about other planetary processes at work in our galaxy. Breakthroughs, discoveries, and DIY tips sent six days a week. Astronomers knew 3I/ATLAS wasn't a local comet not long after first spotting it in July 2025 .

An artist's impression of 3I/ATLAS is shown as it passes near the Sun, illuminating one side of the comet. On the side of the comet closer to the sun, the methanol gas is shown in blue, with icy dust grains still present in the gas. On the dark side of the comet, the hydrogen cyanide is shown in orange. Breakthroughs, discoveries, and DIY tips sent six days a week. The comet 3I/ATLAS is well on its way back into deep space, but the famous cosmic visitor continues to fascinate astronomers.

Listed order was determined by a coin toss.

The interstellar space rock shows off the illuminating effects of its brush with the sun. Breakthroughs, discoveries, and DIY tips sent six days a week. After months of unprecedented observations, astronomers are bidding goodbye to the beloved comet 3I/ATLAS . First spotted in July 2025, the frigid, dusty space rock is only the third known interstellar object to pass through the solar system, offering researchers the rare opportunity to examine a visitor from deep space. Among other discoveries, scientists have since confirmed that the interstellar comet is the fastest ever recorded as well as covered in ice volcanoes --and definitely not extraterrestrial tourists .

We demonstrate the effectiveness of our method on both simulated 2D images andreal3Dbrain magnetic resonance (MR) images.

From surveys of the pre-Sputnik skies to analysis of interstellar visitors, scientists are rethinking how and where to look for physical traces of alien technology. Science fiction is awash in the material remnants of extraterrestrial civilizations, which surface in everything from the classic books of Arthur C. Clarke to game franchises like and . The discovery of the first interstellar objects in the solar system within the past decade has sparked speculation that they could be alien artifacts or spaceships, though the scientific consensus remains that all three of these visitors have natural explanations. That said, scientists have been anticipating the possibility of encountering alien artifacts since the dawn of the space age. "In the history of technosignatures, the possibility that there could be artifacts in the solar system has been around for a long time," says Adam Frank, a professor of astrophysics at the University of Rochester.