The Signal Foundation president recalls where she was when she heard Trump cabinet officials had added a journalist to a highly sensitive group chat. In March of this year, Meredith Whittaker was at her kitchen table in Paris when Signal, the encrypted messaging service she runs, suddenly became an international headline . A colleague sent their group chat the story ricocheting across the globe: "The Trump Administration Accidentally Texted Me Its War Plans." Of course, you know the rest: In the piece, The Atlantic's editor in chief, Jeffrey Goldberg, detailed how he'd been added to a Signal chat about an upcoming military operation in Yemen. Over the following days and weeks, the incident would become known as " SignalGate "--and created a legitimate risk that the fallout would cause people to question Signal's security, instead of pointing their fingers at the profoundly dubious op-sec of senior-level Trump officials. In fact, Signal's user numbers grew by leaps and bounds, both in the US and around the world. It's growth that, Whittaker thinks, is coming at a time when "people are feeling in a much deeper, much more personal way why privacy might be important." On this week's episode of, I talked to Whittaker, who also cofounded the AI Now Institute, about the aftermath of SignalGate, the trajectory of artificial intelligence, and the tech industry's current relationship with politics. Nice to see you, Katie. Nice to see you, too. Brace yourself, we always start these conversations with a little warmup, so I'm going to ask you some very fast questions. I knew you were gonna say that. What's the weirdest AI application you've ever seen? A chatbot that pretends to be your friend.

What is autism and what are Trump's unproven claims about a Tylenol link? US President Donald Trump has claimed there is a link between the use of painkiller Tylenol by pregnant women and an increased risk of autism in some children. Going against current scientific advice and medical opinion, he said the drug, known as paracetamol in many countries, is no good and women should fight like hell to only take it in extreme cases, such as for high fevers. Medical bodies say the drug is safe and that it remains the best treatment for pain and fever during pregnancy. What is autism and how is it diagnosed?

Why is Israel still in southern Lebanon? A war to shape Lebanon's future Lebanon is pushing to get more support from the United States after another deadly Israeli drone attack on southern Lebanon, which this time killed five people, including three children, the latest in a series of near-daily violations by Israel of the US-brokered November 2024 ceasefire. President Joseph Aoun and other officials met with a delegation led by US Secretary of State Marco Rubio, the Lebanese presidency said in a statement on Tuesday. The Lebanese president said he wants Israel to stop occupying parts of his country, is looking to gear its army with "equipment and supplies" from the US, and needs Washington's support to hold a conference dedicated to reconstruction in Lebanon. Amid ongoing efforts to disarm Hezbollah, Aoun emphasised that the Lebanese army's mandate includes "all Lebanese regions" as the country tries to seize an opportunity "to achieve just, comprehensive, and lasting peace in the Middle East region". He is also scheduled to address the United Nations General Assembly on Tuesday, where he is expected to denounce Israeli attacks across the region, including in Gaza and Lebanon.

A deadly drone attack in an impoverished area of Haiti's capital, Port-au-Prince, which killed at least 11 people, including eight children, is being blamed on the government, as the country's use of the UAVs in its war on gangs comes under increasing scrutiny. The incident happened on Saturday night in Cite Soleil, one of Port-au-Prince's most dangerous neighbourhoods, in the city's west along the coast, as Albert Steevenson, known as Djouma or "King Jouma", who is a suspected gang leader, was celebrating his birthday. One of the group's leaders and most notorious figures, Jimmy Cherizier, known as Barbecue, promised to avenge the attack. Claudia Bobrun, 30, whose daughter was killed in the attack, showed The Associated Press news agency a video of the eight-year-old in a pool of blood, as she burst into tears. Merika, another four-year-old victim of the attack, was playing with other children at 8pm in the Simon Pele neighbourhood, in Cite Soleil, where the suspected kamikaze drone exploded.

Porsche's stock tumbled by more than 7% on Monday after warning last week that delays in its electric vehicle (EV) rollout will dent the carmaker's 2025 earnings. Caught between electrification and its iconic petrol-powered sports cars, the German firm said it will slow its push for EVs as demand weakens. Shares of its parent Volkswagen also fell by more than 7% on the same day after saying it will spend billions to overhaul Porsche's line-up of vehicles. The companies' struggles reflect the challenges for European manufacturers, who are faced with intense competition from Chinese rivals and a slowing economy that's dampening demand for luxury cars. Porsche said in a statement on Friday that it has reduced its projected profit margin from up to 7% to 2% or less.

Language models obtain extensive capabilities through pre-training. However, the pre-training process remains a black box. In this work, we track linear interpretable feature evolution across pre-training snapshots using a sparse dictionary learning method called crosscoders. We find that most features begin to form around a specific point, while more complex patterns emerge in later training stages. Feature attribution analyses reveal causal connections between feature evolution and downstream performance. Our feature-level observations are highly consistent with previous findings on Transformer's two-stage learning process, which we term a statistical learning phase and a feature learning phase. Our work opens up the possibility to track fine-grained representation progress during language model learning dynamics.

Recent advances have significantly improved our understanding of the sample complexity of learning in average-reward Markov decision processes (AMDPs) under the generative model. However, much less is known about the constrained average-reward MDP (CAMDP), where policies must satisfy long-run average constraints. In this work, we address this gap by studying the sample complexity of learning an $ε$-optimal policy in CAMDPs under a generative model. We propose a model-based algorithm that operates under two settings: (i) relaxed feasibility, which allows small constraint violations, and (ii) strict feasibility, where the output policy satisfies the constraint. We show that our algorithm achieves sample complexities of $\tilde{O}\left(\frac{S A (B+H)}{ ε^2}\right)$ and $\tilde{O} \left(\frac{S A (B+H)}{ε^2 ζ^2} \right)$ under the relaxed and strict feasibility settings, respectively. Here, $ζ$ is the Slater constant indicating the size of the feasible region, $H$ is the span bound of the bias function, and $B$ is the transient time bound. Moreover, a matching lower bound of $\tildeΩ\left(\frac{S A (B+H)}{ ε^2ζ^2}\right)$ for the strict feasibility case is established, thus providing the first minimax-optimal bounds for CAMDPs. Our results close the theoretical gap in understanding the complexity of constrained average-reward MDPs.

Translating cultural content poses challenges for machine translation systems due to the differences in conceptualizations between cultures, where language alone may fail to convey sufficient context to capture region-specific meanings. In this work, we investigate whether images can act as cultural context in multimodal translation. We introduce CaMMT, a human-curated benchmark of over 5,800 triples of images along with parallel captions in English and regional languages. Using this dataset, we evaluate five Vision Language Models (VLMs) in text-only and text+image settings. Through automatic and human evaluations, we find that visual context generally improves translation quality, especially in handling Culturally-Specific Items (CSIs), disambiguation, and correct gender marking. By releasing CaMMT, our objective is to support broader efforts to build and evaluate multimodal translation systems that are better aligned with cultural nuance and regional variations.

This paper presents DeDisCo, Georgetown University's entry in the DISRPT 2025 shared task on discourse relation classification. We test two approaches, using an mt5-based encoder and a decoder based approach using the openly available Qwen model. We also experiment on training with augmented dataset for low-resource languages using matched data translated automatically from English, as well as using some additional linguistic features inspired by entries in previous editions of the Shared Task. Our system achieves a macro-accuracy score of 71.28, and we provide some interpretation and error analysis for our results.

In recent years, great progress has been made in the field of forecasting meteorological variables. Recently, deep learning architectures have made a major breakthrough in forecasting the daily average temperature over a ten-day horizon. However, advances in forecasting events related to the maximum temperature over short horizons remain a challenge for the community. A problem that is even more complex consists in making predictions of the maximum daily temperatures in the short, medium, and long term. In this work, we focus on forecasting events related to the maximum daily temperature over medium-term periods (90 days). Therefore, instead of addressing the problem from a meteorological point of view, this article tackles it from a climatological point of view. Due to the complexity of this problem, a common approach is to frame the study as a temporal classification problem with the classes: maximum temperature "above normal", "normal" or "below normal". From a practical point of view, we created a large historical dataset (from 1981 to 2018) collecting information from weather stations located in South America. In addition, we also integrated exogenous information from the Pacific, Atlantic, and Indian Ocean basins. We applied the AutoGluonTS platform to solve the above-mentioned problem. This AutoML tool shows competitive forecasting performance with respect to large operational platforms dedicated to tackling this climatological problem; but with a "relatively" low computational cost in terms of time and resources.