Devastating secret message in Savannah Guthrie's video appeal to mother's captors: Hidden agenda revealed by FBI hostage negotiator Crime scene tape goes back up outside missing Nancy Guthrie's home as FBI deploys hostage negotiators to family Savannah Guthrie says'we are ready to talk' as she breaks down in tears during gut-wrenching video while pleading to her mother's captors FBI's Kash Patel heads to Tucson in search for Nancy Guthrie amid daughter Savannah's desperate plea Khloe Kardashian, Jennifer Lopez, and Hoda Kotb lead stars rallying around Savannah Guthrie as she makes tearful plea to her mother's captors Big Short investor proved right in $1bn bet that stock bubble bursting... in dire warning for Wall Street and 401(k)s'I was just laid off in the middle of a WAR ZONE': Distraught Washington Post journalist blasts paper for firing her while she sheltered in Ukraine Melania Trump bombshell stuns Hollywood into silence: Rich and famous gathered at elite restaurants to'laugh' about First Lady rumor... now they have'egg on their faces' Woman who went viral on the Coldplay kiss cam cashes in on her fame as she lands keynote gig in Washington DC - and tickets aren't cheap Epstein claimed Bill Gates was'so cheap' he left Russian mistress broke and sleeping on a friend's couch Where's Fergie...? Ex-Duchess of York's whereabouts is a mystery as she's homeless after Royal Lodge eviction - and Epstein emails shame No, Margot Robbie and Jacob Elordi aren't having an affair... What's REALLY going on is so much worse. I rode with ICE on the frozen streets of Minneapolis. I saw migrants getting arrested. The frontlines are nothing like what you see on TV. New video from a US military drone has captured what UFO investigators are calling a formation of mysterious orbs flying over one of the most contentious regions in the world.

Things to Do in L.A. Tap to enable a layout that focuses on the article. Labor unions urge Gov. Gavin Newsom, California lawmakers to rein in artificial intelligence Lorena Gonzalez, with the California Labor Federation, supports legislation to protect workers from AI. This is read by an automated voice. Please report any issues or inconsistencies here . Labor unions urge Gov. Gavin Newsom to protect workers from AI-driven job losses and workplace surveillance.

Technology Engineering Man solves ceiling fans' most annoying problem His 3D-printed device finally shows a ceiling fans' speed. Breakthroughs, discoveries, and DIY tips sent six days a week. Anyone who's used an overhead ceiling fan knows it can be a pain to work. Yanking its chain gets the motor running, but there's no easy visual indication of what speed setting the fan is on. The blades can also take a frustratingly long time to reach their full speed.

After more than three decades in entertainment, Robbie Williams is back on the road and ready to celebrate. His new album, Britpop, is his 16th number one, breaking the previous record set by the Beatles. The singer, whose Long 90s tour begins this week, is taking a moment to mark his achievement. I think as British people we're very good at piercing the balloon of our own success and undercutting it and devaluing ourselves, he tells BBC News. It's what we do best.

A woman walks near the site of an apartment building hit by a Russian drone strike in Kyiv on Tuesday. KYIV - Ukrainian and Russian officials wrapped up a productive first day of new U.S.-brokered talks in Abu Dhabi, Kyiv's lead negotiator said on Wednesday, as fighting in Europe's biggest conflict since World War II raged on. The two-day trilateral meetings come after Ukraine President Volodymyr Zelenskyy said Russia had exploited a U.S.-backed energy truce last week to stockpile munitions, attacking Ukraine with a record number of ballistic missiles on Tuesday. The work was substantive and productive, focused on concrete steps and practical solutions, Rustem Umerov, the head of Ukraine's National Security and Defense Council, wrote on X. In a time of both misinformation and too much information, quality journalism is more crucial than ever.

Alphabet reports $34.5bn profit and revenue soars 48% in recent quarter as it plans a sharp increase in AI spending Google's parent company, Alphabet, beat Wall Street expectations on Wednesday, and is planning a sharp increase in capital spending in 2026 as it continues to invest deeply in AI infrastructure. Alphabet on Wednesday reported profit of $34.5bn in the recently ended quarter, as revenue from cloud computing soared 48%. In an earnings call, investors pressed Alphabet's chief executive, Sundar Pichai, on the significant increase. "We've been supply constrained, even as we've been ramping up our capacity. Obviously, our CapEx spend this year is an eye towards the future," Pichai said, in response.

Poisson-distributed latent variable models are widely used in computational neuroscience, but differentiating through discrete stochastic samples remains challenging. Two approaches address this: Exponential Arrival Time (EAT) simulation and Gumbel-SoftMax (GSM) relaxation. We provide the first systematic comparison of these methods, along with practical guidance for practitioners. Our main technical contribution is a modification to the EAT method that theoretically guarantees an unbiased first moment (exactly matching the firing rate), and reduces second-moment bias. We evaluate these methods on their distributional fidelity, gradient quality, and performance on two tasks: (1) variational autoencoders with Poisson latents, and (2) partially observable generalized linear models, where latent neural connectivity must be inferred from observed spike trains. Across all metrics, our modified EAT method exhibits better overall performance (often comparable to exact gradients), and substantially higher robustness to hyperparameter choices. Together, our results clarify the trade-offs between these methods and offer concrete recommendations for practitioners working with Poisson latent variable models.

Training modern large language models (LLMs) has become a veritable smorgasbord of algorithms and datasets designed to elicit particular behaviors, making it critical to develop techniques to understand the effects of datasets on the model's properties. This is exacerbated by recent experiments that show datasets can transmit signals that are not directly observable from individual datapoints, posing a conceptual challenge for dataset-centric understandings of LLM training and suggesting a missing fundamental account of such phenomena. Towards understanding such effects, inspired by recent work on the linear structure of LLMs, we uncover a general mechanism through which hidden subtexts can arise in generic datasets. We introduce Logit-Linear-Selection (LLS), a method that prescribes how to select subsets of a generic preference dataset to elicit a wide range of hidden effects. We apply LLS to discover subsets of real-world datasets so that models trained on them exhibit behaviors ranging from having specific preferences, to responding to prompts in a different language not present in the dataset, to taking on a different persona. Crucially, the effect persists for the selected subset, across models with varying architectures, supporting its generality and universality.

Gradient-variation online learning has drawn increasing attention due to its deep connections to game theory, optimization, etc. It has been studied extensively in the full-information setting, but is underexplored with bandit feedback. In this work, we focus on gradient variation in Bandit Convex Optimization (BCO) with two-point feedback. By proposing a refined analysis on the non-consecutive gradient variation, a fundamental quantity in gradient variation with bandits, we improve the dimension dependence for both convex and strongly convex functions compared with the best known results (Chiang et al., 2013). Our improved analysis for the non-consecutive gradient variation also implies other favorable problem-dependent guarantees, such as gradient-variance and small-loss regrets. Beyond the two-point setup, we demonstrate the versatility of our technique by achieving the first gradient-variation bound for one-point bandit linear optimization over hyper-rectangular domains. Finally, we validate the effectiveness of our results in more challenging tasks such as dynamic/universal regret minimization and bandit games, establishing the first gradient-variation dynamic and universal regret bounds for two-point BCO and fast convergence rates in bandit games.

When deploying a single predictor across multiple subpopulations, we propose a fundamentally different approach: interpreting group fairness as a bargaining problem among subpopulations. This game-theoretic perspective reveals that existing robust optimization methods such as minimizing worst-group loss or regret correspond to classical bargaining solutions and embody different fairness principles. We propose relative improvement, the ratio of actual risk reduction to potential reduction from a baseline predictor, which recovers the Kalai-Smorodinsky solution. Unlike absolute-scale methods that may not be comparable when groups have different potential predictability, relative improvement provides axiomatic justification including scale invariance and individual monotonicity. We establish finite-sample convergence guarantees under mild conditions.