Surgical scene understanding is critical for surgical training and robotic decision-making in robot-assisted surgery. Recent advances in Multimodal Large Language Models (MLLMs) have demonstrated great potential for advancing scene perception in the medical domain, facilitating surgeons to understand surgical scenes and procedures. However, these methods are primarily oriented towards image-based analysis or global video understanding, overlooking the fine-grained video reasoning that is crucial for analyzing specific processes and capturing detailed task execution within a surgical procedure. To bridge this gap, we propose SurgVidLM, the first video language model designed to address both full and fine-grained surgical video comprehension. To train our SurgVidLM, we construct the SVU-31K that is a large-scale dataset with over 31K video-instruction pairs, enabling both holistic understanding and detailed analysis of surgical procedures. Building on this resource, SurgVidLM incorporates a two-stage StageFocus mechanism: the first stage extracts global procedural context, while the second stage performs high-frequency local analysis guided by temporal cues. We also develop the Multi-frequency Fusion Attention to effectively integrate low- and high-frequency visual tokens, ensuring the preservation of critical task-specific details. Experimental results demonstrate that SurgVidLM significantly outperforms state-of-the-art Vid-LLMs of comparable parameter scale in both full and fine-grained video understanding tasks, showcasing its superior capability in capturing the context of complex robot-assisted surgeries. Our code and dataset will be publicly accessible soon.

Recent work shows that post-training datasets for LLMs can be substantially downsampled without noticeably deteriorating performance. However, data selection often incurs high computational costs or is limited to narrow domains. In this paper, we demonstrate that data selection can be both -- efficient and universal -- by using a multi-step pipeline in which we efficiently bin data points into groups, estimate quality using specialized models, and score difficulty with a robust, lightweight method. Task-based categorization allows us to control the composition of our final data -- crucial for finetuning multi-purpose models. To guarantee diversity, we improve upon previous work using embedding models and a clustering algorithm. This integrated strategy enables high-performance fine-tuning with minimal overhead.

Interaction between learner and feedback-giver has come into focus recently for post-training of Large Language Models (LLMs), through the use of reward models that judge the appropriateness of a model's response. In this paper, we investigate whether Dialogue Games -- goal-directed and rule-governed activities driven predominantly by verbal actions -- can also serve as a source of feedback signals for learning. We introduce Playpen, an environment for off- and online learning through Dialogue Game self-play, and investigate a representative set of post-training methods: supervised fine-tuning; direct alignment (DPO); and reinforcement learning with GRPO. We experiment with post-training a small LLM (Llama-3.1-8B-Instruct), evaluating performance on unseen instances of training games as well as unseen games, and on standard benchmarks. We find that imitation learning through SFT improves performance on unseen instances, but negatively impacts other skills, while interactive learning with GRPO shows balanced improvements without loss of skills. We release the framework and the baseline training setups to foster research in the promising new direction of learning in (synthetic) interaction.

Conversations around birth control usually focus on women, but Kevin Eisenfrats, one of the MIT Technology Review 2025 Innovators Under 35, is working to change that. His company, Contraline, is working toward testing new birth control options for men . Exclusive: A record-breaking baby has been born from an embryo that's over 30 years old Jessica Hamzelou Therapists are secretly using ChatGPT. Exclusive: A record-breaking baby has been born from an embryo that's over 30 years old The embryos were created in 1994, while the expectant father was still a toddler, and donated via a Christian "embryo adoption" agency. Therapists are secretly using ChatGPT. Some therapists are using AI during therapy sessions.

A single AI model trained to control numerous robotic bodies can operate unfamiliar hardware and adapt eerily well to serious injuries. A four-legged robot that keeps crawling even after all four of its legs have been hacked off with a chainsaw is the stuff of nightmares for most people. For Deepak Pathak, cofounder and CEO of the startup Skild AI, the dystopian feat of adaptation is an encouraging sign of a new, more general kind of robotic intelligence. "This is something we call an omni-bodied brain," Pathak tells me. His startup developed the generalist artificial intelligence algorithm to address a key challenge with advancing robotics: "Any robot, any task, one brain.

Demand for AI chips is booming--and so is the need for software to run them. Chris Lattner's startup Modular just raised $250 million to build the best developer tools for AI hardware. At a certain point between building Apple's developer tools, leading a core part of Google's AI infrastructure team, and clashing with Elon Musk during a stint as Tesla's Autopilot chief, Chris Lattner's vision for his life's work started to come into focus. AI was taking over the world, and demand was growing for the chips that powered it. But the software stack for those chips was dominated by just a few big companies.

The A.I. Bubble Is Coming for Your Browser Artificial-intelligence startups, like the makers of the "smart" web browser Dia, are being acquired for vast sums. There's an old business maxim dating to the California gold rush: it's easier to make money selling picks and shovels to aspiring miners than to strike it rich finding gold. Artificial intelligence is in a picks-and-shovels phase right now. If gold, in this metaphor, is artificial general intelligence--a machine smarter than a human--or some version of a digital god, then tech companies are snapping up the tools to create one, including graphics-processing units, data centers, and trained A.I. models. That scramble is why Mark Zuckerberg is paying a twenty-four-year-old A.I. researcher two hundred and fifty million dollars to work at Meta, and why Sam Altman, the C.E.O. of OpenAI, recently said that the company would spend "trillions of dollars" building infrastructure.

That's why we've created the AI Hype Index--a simple, at-a-glance summary of everything you need to know about the state of the industry. Millions of us use chatbots every day, even though we don't really know how they work or how using them affects us. In a bid to address this, the FTC recently launched an inquiry into how chatbots affect children and teenagers. Elsewhere, OpenAI has started to shed more light on what people are actually using ChatGPT for, and why it thinks its LLMs are so prone to making stuff up. There's still plenty we don't know--but that isn't stopping governments from forging ahead with AI projects. In the US, RFK Jr. is pushing his staffers to use ChatGPT, while Albania is using a chatbot for public contract procurement.

Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning, positioning them as promising tools for supporting human problem-solving. However, what happens when their performance is affected by misinformation, i.e., incorrect inputs introduced by users due to oversights or gaps in knowledge? Such misinformation is prevalent in real-world interactions with LLMs, yet how it propagates within LLMs' reasoning process remains underexplored. Focusing on mathematical reasoning, we present a comprehensive analysis of how misinformation affects intermediate reasoning steps and final answers. We also examine how effectively LLMs can correct misinformation when explicitly instructed to do so. Even with explicit instructions, LLMs succeed less than half the time in rectifying misinformation, despite possessing correct internal knowledge, leading to significant accuracy drops (10.02% - 72.20%), and the degradation holds with thinking models (4.30% - 19.97%). Further analysis shows that applying factual corrections early in the reasoning process most effectively reduces misinformation propagation, and fine-tuning on synthesized data with early-stage corrections significantly improves reasoning factuality. Our work offers a practical approach to mitigating misinformation propagation.