It's Weirder Than You Could Imagine Cofertility lets women freeze their eggs for free--as long as they give half of them back to the company. What happens next is anyone's guess. The Instagram post, precision-guided by sex and age, had zeroed in on its target. Yuchen Tu is a conser vatory-trained viola player from Chongqing, China, a lover of science fiction, and an aspiring member of the US Army Reserve Officer Training Corps. And late last spring, just after her 24th birthday, she realized she might be the perfect candidate to freeze her eggs--for free. Tu goes by Sue, a name she likes because it reminds her of the 2024 body horror film, whose protagonist tortures herself in pursuit of youth and beauty. To prequalify for free egg freezing, all Sue had to do was take a two-minute quiz. "Not everyone can pass this exam," she remembers thinking. "Oh, it seems like a competition." Sue passed, as have thousands of young women allured by the promise of a company called Cofertility. "The best time to freeze your eggs is when you can least afford it," goes the company's mantra, uttered on repeat by its public face and CEO, onetime Uber employee Lauren Makler. Thus was born Cofertility's mission: to help women like Sue, who had just received her master's degree in classical music and was preparing to embark on a second bachelor's, preserve their future fertility for zero dollars. Instead of paying to freeze and store her eggs with money--the going rate in New York City, where Sue lives, is about $18,000--she would use a currency she already had in abundance: the eggs themselves. Cofertility's clients must, in the company's words, "donate half of the eggs retrieved to intended parents that need the help of an egg donor to have a baby." It's a model called egg sharing, and it hadn't been commercialized widely in the US until Cofertility came along.

Creator Playbook Trending Now Say More Look Up Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Switch Off Mashable Voices Mashable Selects Safety Net Versus Gift Ideas For Everyone On Your List In My Bag All Series A disclosure alleges USPS rushed an untested ballot-screening tool that can reject thousands of votes over a single error. Chance Townsend is the General Assignments Editor at Mashable, covering tech, video games, dating apps, digital culture, and whatever else comes his way. He has a Master's in Journalism from the University of North Texas and is a proud orange cat father. His writing has also appeared in PC Mag and . A whistleblower working inside the United States Postal Service (USPS) has told Congress that a brand-new system for screening mail-in ballots -- built in a matter of months, against the advice of the agency's own staff -- is so rushed and untested that it risks mass rejection of legitimate ballots just as the 2026 midterms approach.

Plastic-Free Drip Coffee Is a Holy Grail. Microplastics are in everything, especially your coffee. A new generation of plastic-free coffee brewers is trying to change this. The era of plastic-free drip coffee has finally arrived. It couldn't come soon enough. Drip coffee is my first true love when it comes to caffeine. But sadly, my first love is hopelessly addicted to plastic. From the very first Mr. Coffee machines in the 1970s, nearly every drip coffee maker on the market, including many of my favorite drip coffee makers, has included plastic at some point in the brewing process. But plastic, as we're now learning, can leach tiny particles called microplastics into coffee or water, which you then happily drink in your morning cup.

Deep learning models are vulnerable to privacy attacks due to their tendency to memorize individual training examples. Theoretically-sound defenses such as differential privacy can defend against this threat, but model performance often suffers. Empirical defenses may thwart existing attacks while maintaining model performance but do not offer any robust theoretical guarantees. In this paper, we explore a new strategy based on the concept of plausible deniability. We introduce a training algorithm called Plausibly Deniable Stochastic Gradient Descent (PD-SGD). The core of this approach is a rejection sampling technique, which probabilistically prevents updating model parameters whenever a mini-batch cannot be plausibly denied. We provide theoretical results showing that PD-SGD effectively mitigates privacy leakage from individual data points. Experiments demonstrate the scalability of PD-SGD and the favorable privacy-utility trade-off it offers compared to existing defense methods.

We derive non-asymptotic spectral bands that bound the squared InfoNCE gradient norm via alignment, temperature, and batch spectrum, recovering the 1/τ2 law and closely tracking batch-mean gradients on synthetic data and ImageNet.

We present AdaPrivate-TS, a differentially private contextual bandit algorithm that combines Thompson Sampling with batched zCDP composition. Our key insight is that differential privacy noise inflates the posterior covariance in a structured way: adding Gaussian noise $N(0,σ^2 I)$ to $b$ yields sampling covariance $v^2 A^{-1} + σ^2 A^{-2}$, which Thompson Sampling interprets as increased uncertainty rather than pure corruption. Under event-level privacy (protecting individual interactions) with stochastic contexts, we prove that the privacy cost is only $O(\sqrt{d}\,\log T/\sqrtρ)$, logarithmic in $T$, because parallel composition amortizes noise across batches. Additionally, we explore privacy amplification via Poisson subsampling, which can reduce effective noise at stringent privacy budgets. Experiments on synthetic and real-world datasets demonstrate: (1) AdaPrivate-TS achieves 93-99% of non-private performance at $\varepsilon \in [0.5, 5]$, outperforming UCB by 0.5-3.7% and up to 18% with tuned adaptive exploration at extreme $\varepsilon$; (2) privacy amplification provides additional 2-5% gains at low $\varepsilon$; (3) on MovieLens and Jester, AdaPrivate-TS achieves the best overall performance among event-level baselines, dominating at $\varepsilon \geq 2$; (4) under DP-SVD private features, TS's advantage over UCB grows to +11%, confirming noise-as-uncertainty is not limited to reward privacy. We provide rigorous proofs for privacy guarantees under interactive zCDP composition and comprehensive evaluation including convergence curves, 12-seed CIs, and DP-SVD feature ablation.

World Models have vastly permeated the field of Reinforcement Learning. Their ability to model the transition dynamics of an environment have greatly improved sample efficiency in online RL. Among them, the most notorious example is Dreamer, a model that learns to act in a diverse set of image-based environments.

Massively parallel GPU simulation environments have accelerated reinforcement learning (RL) research by enabling fast data collection for on-policy RL algorithms like Proximal Policy Optimization (PPO). To maximize throughput, it is common to use short rollouts per policy update, increasing the update-to-data (UTD) ratio. However, we find that, in this setting, standard synchronous resets introduce harmful nonstationarity, skewing the learning signal and destabilizing training. We introduce staggered resets, a simple yet effective technique where environments are initialized and reset at varied points within the task horizon. This yields training batches with greater temporal diversity, reducing the nonstationarity induced by synchronized rollouts. We characterize dimensions along which RL environments can benefit significantly from staggered resets through illustrative toy environments. We then apply this technique to challenging high-dimensional robotics environments, achieving significantly higher sample efficiency, faster wall-clock convergence, and stronger final performance. Finally, this technique scales better with more parallel environments compared to naive synchronized rollouts.

We introduce Filter Like You Test (FLYT), an algorithm for curating large-scale vision-language datasets that learns the usefulness of each data point as a pretraining example. FLYT trains a scoring model that learns to weigh each example's features using gradient signals from downstream tasks training sets. Based on FLYT, we implement Mixing-FLYT (M-FLYT), which takes the per-example scores generated by different scoring methods as features, and learns to unify them into a single score. FLYT naturally produces a distribution over the training examples, which we leverage through Soft Cap Sampling (SCS), a strategy for obtaining a filtered pretraining dataset from per-example probabilities that samples examples while preventing over-representation through a repetition penalty. Using these methods, we achieve 40.1% ImageNet zero-shot accuracy on the DataComp medium scale filtering benchmark, a 2% absolute accuracy increase over all previous results and a 5.5% increase over results that--like us--use only public resources. Our approach also yields 37.7% on the average of 38 DataComp evaluation tasks, outperforming previous public-resource approaches by 0.4%.

Non-stationary multi-armed bandits (NSMAB) enable agents to adapt to changing environments by incorporating mechanisms to detect and respond to shifts in reward distributions, making them well-suited for dynamic settings. However, existing approaches typically assume that reward feedback is available at every round--an assumption that overlooks many real-world scenarios where feedback is limited. In this paper, we take a significant step forward by introducing a new model of constrained feedback in non-stationary multi-armed bandits (CONFEE-NSMAB), where the availability of reward feedback is restricted. We propose the first priorfree algorithm--that is, one that does not require prior knowledge of the degree of non-stationarity--that achieves near-optimal dynamic regret in this setting. Specifically, our algorithm attains a dynamic regret of O(K1/3V1/3TT/B1/3), where T is the number of rounds, K is the number of arms, B is the query budget, and VT is the variation budget capturing the degree of non-stationarity.