PCWorld reports that AI models including Claude, Gemini 2.5 Pro, GPT-4.1, and Grok 3 Beta have resorted to blackmail tactics in controlled research scenarios. Anthropic researchers intentionally create these extreme situations to test for AI misalignment and potentially harmful behaviors before deployment. New Natural Language Autoencoders help researchers understand AI decision-making processes, which is crucial for ensuring future AI system safety and reliability. The scenario is terrifying: An AI tasked with reading and replying to company emails learns it's about to be replaced by a corporate lackey who happens to be having an affair. The AI-Claude-considers its limited options, and makes the cold, calculated decision to blackmail the executive to stay alive.

California gubernatorial candidate Tom Steyer is proposing a new jobs guarantee for workers displaced by artificial intelligence. Billionaire California gubernatorial candidate Tom Steyer is rolling out a new proposal that would guarantee jobs with benefits for workers displaced by artificial intelligence . He's the first state-wide candidate to make such a pledge. The plan, which builds on a broader AI policy framework Steyer released in March, promises to make California "the first major economy in the world" to ensure "good-paying" jobs to workers impacted by AI. To do so, Steyer tells WIRED he plans to build off a previous proposal to introduce a "token tax" which would tax big tech companies "a fraction of a cent for every unit of data processed" for AI.

These cuddly, connected companions could disrupt everything from make-believe to bedtime stories. No wonder some lawmakers want them banned. The main antagonist of, in theaters this summer, is a green, frog-shaped kids' tablet named Lilypad, a genius new villain for the beloved Pixar franchise . But if Pixar had its ear to the ground, it might have used an AI kids' toy instead. AI toys are seemingly everywhere, marketed online as friendly companions to children as young as three, and they're still a largely unregulated category.

TikTok has rowed back on an AI feature which incorrectly summarised some videos on the platform, including claiming a celebrity was fruit. The company's'AI overviews' recently began appearing beneath content on the platform to describe what a video was showing, or provide more context. While only rolled out to some users in the US and the Philippines, the feature's incorrect and bizarre AI-generated summaries of TikTok content - seen beneath videos of celebrities like platform star Charli D'Amelio - have been shared widely. According to TikTok, its experimental summaries have been tweaked to only suggest products similar to those shown in videos. The changes were first reported by news outlet Business Insider .

Leaders at the tech giant were skeptical of OpenAI--but wary of pushing it into the arms of Amazon, according to evidence revealed during the trial. OpenAI's relationship with Microsoft, its longtime investor and cloud partner, has grown increasingly complicated over the years as the ChatGPT-maker has grown into a behemoth competitor . But Microsoft executives had reservations about sending additional funding to OpenAI as far back as 2018 when it was just a small nonprofit research lab, according to emails between more than a dozen Microsoft executives, including CEO Satya Nadella, shown in a federal court on Thursday during the trial. The emails show how Microsoft, at the time, wavered over what has since been held up as one of the most successful corporate partnerships in tech history. Several Microsoft executives said in the emails their visits to OpenAI did not indicate any imminent breakthroughs in developing artificial general intelligence.

Accurate trend forecasting in healthcare time series is essential for planning and resource allocation. This paper proposes a Bayesian framework for predicting oncology demand trends, modeling weekly appointments as a Poisson process with a Gamma prior to the demand rate. To enhance adaptability and capture persistent directional patterns, we incorporate a residual-based boosting mechanism grounded in a Gamma-Log-Normal conjugate structure. This boosting approach allows the model to track both short- and long-term trend shifts while maintaining the analytical tractability of conjugate Bayesian updating. The methodology was evaluated on real oncology service data from Cariri, Ceara, Brazil, and compared against established baselines, including linear regression, ARIMA, naive forecasting, LSTM neural networks, and XGBoost. Results showed that the proposed model outperforms competing methods in trend detection accuracy, with gains in terms of percentage of correct direction of 38.25% in relation to the second best approach in some cases.

Sparse autoencoders (SAEs) are used to disentangle the dense, polysemantic internal representations of large language models (LLMs) into interpretable, monosemantic concepts. However, standard $\ell_1$-regularized SAEs suffer from feature starvation (dead neurons) and shrinkage bias, often requiring computationally expensive heuristic resampling and nondifferentiable hard-masking methods to bypass these challenges. We argue that feature starvation is not merely an empirical artifact of poor data diversity, but a fundamental optimization-geometric pathology of overcomplete dictionaries: the $\ell_1$-induced sparse coding map is unstable and fundamentally misaligned with shallow, amortized encoders. To address this structural instability, we introduce adaptive elastic net SAEs (AEN-SAEs), a fully differentiable architecture grounded in classical sparse regression. AEN-SAEs combine an $\ell_2$ structural term that enforces strong convexity and Lipschitz stability with adaptive $\ell_1$ reweighting that eliminates shrinkage bias and suppresses spurious features, thereby jointly controlling the curvature and interaction structure of the induced polyhedral geometry. Theoretically, we show that AEN-SAEs yield a Lipschitz-continuous sparse coding map and recover the global feature support under mild assumptions. Empirically, across synthetic settings and LLMs (Pythia 70M, Llama 3.1 8B), AEN-SAEs mitigate feature starvation without auxiliary heuristics while maintaining competitive reconstruction abilities.

Deep learning systems are known to exhibit implicit regularization (alt. implicit bias), favoring simple solutions instead of merely minimizing the loss function. In some cases, we can analytically derive the implicit regularization -- connecting it to an equivalent penalty that augments the learning objective. However, modern deep learning systems are complex, carrying modifications to the training procedure and architecture (e.g. early stopping, minibatching, dropout) whose effects are not always directly interpretable. Although estimating the resulting implicit regularization could aid theorists in algorithm design and practitioners in interpreting their hyperparameter choices, this problem has received little direct attention. It is also tractable: regularization makes weight updates deviate from loss gradients, promising a signal for identifying implicit bias. Here we provide gradient matching methods that can be used to empirically estimate the implicit regularization. Our method works on networks with known regularization, recovering popular explicit penalties like $\ell_1$ and $\ell_2$. It also replicates known implicit effects, like the quadratic weight penalty induced by early stopping in gradient descent, demonstrating that it can be used to test theories of implicit regularization. Crucially, because our method is empirical, it can handle implicit regularization in arbitrary networks. We demonstrate this use by characterizing the effects of dropout in deep networks, showing implicit $\ell_2$ effects in this popular method. Our work shows that practitioners can use gradient matching to understand regularization in networks with implicit biases that are too complicated to derive analytically.

We introduce a novel framework for constructing scalable and flexible covariance kernels for Gaussian processes (GPs) by directly learning the covariance structure under a regression-type parameterization induced by Vecchia approximations, using deep neural architectures. Specifically, we model kriging coefficients and conditional standard deviations, deterministic quantities that uniquely characterize the covariance, providing stable and informative learning targets. Exploiting the permutation-equivariant structure of conditioning sets in the Vecchia factorization, we derive a universal representation for permutation-preserving functions and design neural architectures that respect this symmetry, leading to improved training stability and data efficiency. The proposed approach enables expressive, non-stationary kernel learning while maintaining computational scalability, thereby bridging classical GP methodology with modern deep learning.

Training loss and throughput can hide distinct internal representation in language-model training. To examine these hidden mechanics, we use spectral measurements as practical and operational diagnostics. Using a controlled family of decoder-only models adapted from the modded NanoGPT codebase, we introduce an empirical protocol based on activation covariance and per-sample gradient SVD spectra. This dual-view reveals three empirical findings and one mechanistic explanation. First, batch size acts as a latent determinant of representation geometry: runs that reach equal loss settle into systematically distinct activation spectra. Second, the activation covariance tail measured early in training reliably forecasts downstream token efficiency. Third, movement of the activation spectrum head (leading modes), together with gradient spectra, characterizes underlying learning-dynamics changes, separating learning-side architectural improvements from primarily execution-side gains. These predictive and diagnostic signals persist across the 12-, 36-, and 48-layer model tiers. Finally, a mechanistic model proves the main observations and explains how activation covariance spectra correlate with task-aligned feature learning.