They expect you to die! Interactive takes on MI6's globetrotting spy have been around almost as long as the films, but that doesn't mean all of them were a success. 'The enormity of the idea helped me': how Patrick Gibson became gaming's new James Bond Bond finally arrived in an official video game capacity in 1984, courtesy of Parker Brothers. The game grouped several 007 adventures (Diamonds Are Forever, The Spy Who Loved Me, Moonraker and For Your Eyes Only) together. Yet despite including elements from each movie, it was essentially the same game throughout: an unsatisfying and tricky mashup of the arcade games Moon Patrol and Scramble, with the player controlling Bond's amphibious Lotus from The Spy Who Loved Me. Obscure pub trivia fact: due to the dispute between Bond producers Eon and screenwriter Kevin McClory, the Diamonds Are Forever segment replaced Blofeld with a villain named Seraffino.

Follow this section to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW? Smart Alerts: Get notified about major news as it happens. Follow this tag to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW?

PCWorld reveals that robot vacuums require regular maintenance and human intervention despite being marketed as fully autonomous cleaning devices. While these devices excel at surface dust and daily cleaning tasks, they struggle with deep-seated pet hair, larger debris, and need consistent emptying and care. Advanced models with LiDAR mapping and scheduling features offer better navigation than cheaper alternatives, but even premium options cannot replace traditional vacuums for thorough cleaning. Years ago, I owned a robot vacuum named Gerald. After a catastrophic tumble down the stairs, I retired him early and have stuck with traditional vacuuming ever since. But after spending more time with them lately and chatting with the experts that make them, I realized I've been holding onto a lot of misconceptions, especially about how "smart" these things actually are in real homes with clutter and pets. I used to think suction power was everything.

Two former staffers have created a new, perplexing company. And DOGE alumni make splashy announcements about entering complex industries with scant qualifications while promising to "root out waste." This, at least, is the premise of Special, a newly announced start-up co-founded by Justin Fox and Nate Cavanaugh, two former Department of Government Efficiency staffers who left the federal government "motivated to extend the ethos of our work at DOGE back into the private sector," as they wrote on Special's website. The company officially launched last week with funding from the Elon Musk-friendly contingent of Silicon Valley, including the venture groups Andreessen Horowitz and Human Capital. Special is also backed by investments from numerous Musk associates, including Steve Davis, Musk's top lieutenant at DOGE.

A Waymo nearly hit me, but I'm still optimistic about driverless cars A near miss with a Waymo while cycling through London hasn't changed my optimistic stance on driverless cars, but we can't ever let our guard down, says Matthew Sparkes Waymo driverless cars are in London, but is this a positive move for road safety? Waymo's driverless cars have been rolling through London for months, although they aren't taking passengers yet and a human sits ready to seize control if needed. Every time I've encountered them, they have seemed cautious and predictable. But recently, I had a near miss. I was circling a roundabout as I cycled home from work and a Waymo was about to pull onto it in front of me.

Behavior cloning has shown promise for robot manipulation, but real-world demonstrations are costly to acquire at scale. While simulated data offers a scalable alternative, particularly with advances in automated demonstration generation, transferring policies to the real world is hampered by various simulation and real domain gaps. In this work, we propose a unified sim-and-real co-training framework for learning generalizable manipulation policies that primarily leverages simulation and only requires a few real-world demonstrations. Central to our approach is learning a domain-invariant, task-relevant feature space. Our key insight is that aligning the joint distributions of observations and their corresponding actions across domains provides a richer signal than aligning observations (marginals) alone. We achieve this by embedding an Optimal Transport (OT)-inspired loss within the co-training framework, and extend this to an Unbalanced OT framework to handle the imbalance between abundant simulation data and limited real-world examples. We validate our method on challenging manipulation tasks, showing it can leverage abundant simulation data to achieve up to a 30\% improvement in the real-world success rate and even generalize to scenarios seen only in simulation.

Hallucinations present a significant challenge in the development and evaluation of large language models (LLMs), directly affecting their reliability and accuracy. While notable advancements have been made in research on textual and visual hallucinations, there is still a lack of a comprehensive benchmark for evaluating auditory hallucinations in large audio language models (LALMs).

The success of Graph Neural Networks (GNNs) leverages the homophily principle, where connected nodes share similar features and labels. However, this assumption breaks down in heterophilic graphs, where same-class nodes are often distributed across distant neighborhoods rather than immediate connections. Recent attempts expand the receptive field through multi-hop aggregation schemes that explicitly preserve intermediate representations from each hop distance. While effective at capturing heterophilic patterns, these methods require separate weight matrices per hop and feature concatenation, causing parameters to scale linearly with hop count. This leads to high computational complexity and GPU memory consumption. We propose Gated Multi-hop Message Passing (GAMMA), where nodes assess how relevant the aggregated information is from their k-hop neighbors. This assessment occurs through multiple refinement steps where the node compares each hop's embedding with its current representation, allowing it to focus on the most informative hops. During the forward pass, GAMMA finds the optimal mix of multi-hop information local to each node using a single feature vector without needing separate representations for each hop, thereby maintaining dimensionality comparable to single hop GNNs. In addition, we propose a weight sharing scheme that leverages a unified transformation for aggregated features from multiple hops so the global heterophilic patterns specific to each hop are learned during training.

Diffusion models have emerged as a powerful paradigm for modern generative modeling, demonstrating strong potential for large language models (LLMs). Unlike conventional autoregressive (AR) models that generate tokens sequentially, diffusion models allow for parallel sampling, offering a promising path to accelerate generation and eliminate the left-to-right generation constraints. Despite their empirical success, theoretical understandings of diffusion language models remain underdeveloped. In this work, we develop convergence guarantees for diffusion language models from an information-theoretic perspective. Our analysis demonstrates that the sampling error, measured by the Kullback-Leibler (KL) divergence, decays inversely with the number of iterations $T$ and scales linearly with the mutual information between tokens in the target text sequence.

Sparse Autoencoders (SAEs) are widely used to interpret neural networks by identifying meaningful concepts from their representations. However, do SAEs truly uncover all concepts a model relies on, or are they inherently biased toward certain kinds of concepts? We introduce a unified framework that recasts SAEs as solutions to a bilevel optimization problem, revealing a fundamental challenge: each SAE imposes structural assumptions about how concepts are encoded in model representations, which in turn shapes what it can and cannot detect. This means different SAEs are not interchangeable -- switching architectures can expose entirely new concepts or obscure existing ones. To systematically probe this effect, we evaluate SAEs across a spectrum of settings: from controlled toy models that isolate key variables, to semi-synthetic experiments on real model activations and finally to large-scale, naturalistic datasets. Across this progression, we examine two fundamental properties that real-world concepts often exhibit: heterogeneity in intrinsic dimensionality (some concepts are inherently low-dimensional, others are not) and nonlinear separability. We show that SAEs fail to recover concepts when these properties are ignored, and we design a new SAE that explicitly incorporates both, enabling the discovery of previously hidden concepts and reinforcing our theoretical insights. Our findings challenge the idea of a universal SAE and underscores the need for architecture-specific choices in model interpretability. Overall, we argue an SAE does not just reveal concepts -- it determines what can be seen at all.