We'll link all the stories we spoke about today in the show notes. Adriana Tapia produced this episode, Amar Lal at Macro Sound mixed this episode, Matt Giles and Daniel Roman fact-checked this episode, Mark Leyda was our San Francisco studio engineer, Kate Osborn is our executive producer, and Katie Drummond is WIRED's global editorial director.

Microsoft stock has slumped 12 percent as part of a software industry sell-off, stoking fears of whether hefty investments in artificial intelligence will pay off across the sector. The Redmond, Washington-based tech giant is on track Thursday to finish at its worst day since March 2020 and has seen approximately $400bn in valuation wiped out. Capital expenditures grew by 66 percent in the second quarter compared with the same period the year before, reaching a record $37.5bn for the quarter. Meanwhile, Microsoft predicted Azure growth to stay stable in the period from January to March at 37 percent to 38 percent, after slowing in the last three months of 2025, partially due to AI chip capacity constraints. "[Wall Street] wanted to see less cap-ex spending and faster cloud/AI monetisation and coming out of the gates, it's the opposite. We have said this is a multi-year journey, and Redmond needs to focus on its data center buildout with more customers heading down the AI path. It's a balancing act with 2026 the inflection year for AI and MSFT [Microsoft]," Dan Ives, analyst at Wedbush Securities, said in a note provided to Al Jazeera.

Meta wowed Wall Street with improvements in ad targeting fueled by AI alongside huge investment. Big tech earnings so far this week have sent a clear warning: investors are willing to overlook soaring spending on artificial intelligence if it fuels strong growth, but are quick to punish companies that fall short. The contrast was clear in Thursday's stock market reaction to earnings from Microsoft and Meta, highlighting how dramatically the stakes have changed since the launch of ChatGPT started the AI boom more than three years ago. Shares of the Instagram parent surged more than 9% on strong sales, while those of Microsoft slumped 10% after its cloud business failed to impress. "The market appears to be questioning whether these massive capital expenditure hikes will generate sufficient returns," said Jesse Cohen, senior analyst at Investing.com.

Google's Project Genie lets you generate your own interactive worlds You'll need a Google AI Ultra subscription to try the showcase. Project Genie allows people outside of Google to try the company's Genie 3 world model. This past summer, Google DeepMind debuted Genie 3 . It's what's known as a world world, an AI system capable of generating images and reacting as the user moves through the environment the software is simulating. At the time, DeepMind positioned Genie 3 as a tool for training AI agents.

AI chat toy company Bondu left its web console almost entirely unprotected. Researchers who accessed it found nearly all the conversations children had with the company's stuffed animals. Earlier this month, Joseph Thacker's neighbor mentioned to him that she'd preordered a couple of stuffed dinosaur toys for her children. She'd chosen the toys, called Bondus, because they offered an AI chat feature that lets children talk to the toy like a kind of machine-learning-enabled imaginary friend. But she knew Thacker, a security researcher, had done work on AI risks for kids, and she was curious about his thoughts.

Gestala is the latest company to emerge from China's burgeoning brain-computer interface industry. It plans to access the brain with noninvasive ultrasound technology. China's brain-computer interface industry is growing fast, and the newest company to emerge from the country is aiming to access the brain without the use of invasive implants . Gestala, newly founded in Chengdu with offices in Shanghai and Hong Kong, plans to use ultrasound technology to stimulate--and eventually read from--the brain, according to CEO and cofounder Phoenix Peng. It's the second company to launch in recent weeks with the aim of tapping into the brain with ultrasound.

Please enable Javascript in your browser to view Slate interactives. Slate Crossword: Get Too Old for This S--, Perhaps? (Six Letters) Which President Was Largely Blamed by for the Financial Panic of 1837? Slate is published by The Slate Group, a Graham Holdings Company. Slate relies on advertising to support our journalism. If you value our work, please disable your ad blocker.

Nvidia helped DeepSeek hone AI models later used by China's military China's DeepSeek received extensive technical assistance from Nvidia as a legitimate commercial partner hone artificial intelligence models that were later used by the Chinese military, it has been revealed. SAN FRANCISCO - U.S. chipmaker Nvidia helped China's DeepSeek hone artificial intelligence models that were later used by the Chinese military, the chairman of a U.S. House of Representatives committee said in a letter on Wednesday. DeepSeek shook markets early last year with a set of AI models that rivaled some of the best offerings from the United States but were developed with far less computing power, fueling concerns in Washington that China could catch up with the U.S. in AI despite U.S. restrictions on the sale of high-powered computing chips to China. In a letter to U.S. Commerce Secretary Howard Lutnick, Rep. John Moolenaar, a Michigan Republican who chairs the House Select Committee on China, said documents obtained by the committee from Nvidia showed the achievement came after extensive technical assistance from Nvidia. In a time of both misinformation and too much information, quality journalism is more crucial than ever. By subscribing, you can help us get the story right.

Deep neural networks (DNNs) achieve remarkable performance on a wide range of tasks, yet their mathematical analysis remains fragmented: stability and generalization are typically studied in disparate frameworks and on a case-by-case basis. Architecturally, DNNs rely on the recursive application of parametrized functions, a mechanism that can be unstable and difficult to train, making stability a primary concern. Even when training succeeds, there are few rigorous results on how well such models generalize beyond the observed data, especially in the generative setting. In this work, we leverage the theory of stochastic Iterated Function Systems (IFS) and show that two important deep architectures can be viewed as, or canonically associated with, place-dependent IFS. This connection allows us to import results from random dynamical systems to (i) establish the existence and uniqueness of invariant measures under suitable contractivity assumptions, and (ii) derive a Wasserstein generalization bound for generative modeling. The bound naturally leads to a new training objective that directly controls the collage-type approximation error between the data distribution and its image under the learned transfer operator. We illustrate the theory on a controlled 2D example and empirically evaluate the proposed objective on standard image datasets (MNIST, CelebA, CIFAR-10).

We study model-agnostic post-hoc calibration methods intended to improve probabilistic predictions in supervised binary classification on real i.i.d. tabular data, with particular emphasis on conformal and Venn-based approaches that provide distribution-free validity guarantees under exchangeability. We benchmark 21 widely used classifiers, including linear models, SVMs, tree ensembles (CatBoost, XGBoost, LightGBM), and modern tabular neural and foundation models, on binary tasks from the TabArena-v0.1 suite using randomized, stratified five-fold cross-validation with a held-out test fold. Five calibrators; Isotonic regression, Platt scaling, Beta calibration, Venn-Abers predictors, and Pearsonify are trained on a separate calibration split and applied to test predictions. Calibration is evaluated using proper scoring rules (log-loss and Brier score) and diagnostic measures (Spiegelhalter's Z, ECE, and ECI), alongside discrimination (AUC-ROC) and standard classification metrics. Across tasks and architectures, Venn-Abers predictors achieve the largest average reductions in log-loss, followed closely by Beta calibration, while Platt scaling exhibits weaker and less consistent effects. Beta calibration improves log-loss most frequently across tasks, whereas Venn-Abers displays fewer instances of extreme degradation and slightly more instances of extreme improvement. Importantly, we find that commonly used calibration procedures, most notably Platt scaling and isotonic regression, can systematically degrade proper scoring performance for strong modern tabular models. Overall classification performance is often preserved, but calibration effects vary substantially across datasets and architectures, and no method dominates uniformly. In expectation, all methods except Pearsonify slightly increase accuracy, but the effect is marginal, with the largest expected gain about 0.008%.