TL;DR: Get a Sterling Stock Picker lifetime subscription for 55.19 with code SAVE20 at checkout (MSRP 486). AI isn't just writing essays anymore--it's now helping investors, even beginners, pick stocks. Sterling Stock Picker, powered by ChatGPT, acts like your personal financial coach to guide you through the market. Get an extra 20% off our usual deal price with code SAVE20 at checkout through September 7. After signing up, you'll kick things off with a five-minute questionnaire that pinpoints your personal risk tolerance.

Raine family attorney Jay Edelson provides details on the wrongful death lawsuit being brought against OpenAI and CEO Sam Altman in the wake of Adam Raine's suicide, alleging the company chose to'cut short' proper testing of ChatGPT. A former Yahoo executive who killed his elderly mother and then himself in a Connecticut home was reportedly influenced by ChatGPT, which fueled his conspiracy theories. Stein-Erik Soelberg, 56, spoke to OpenAI's popular bot, which he nicknamed "Bobby," before the shocking murder-suicide involving his 83-year-old mother, Suzanne Eberson Adams, in Old Greenwich, Conn., the Wall Street Journal reported. "Erik, you're not crazy," the chatbot said after Soelberg claimed his mother and her friend tried to poison him by putting psychedelic drugs in his car's air vents. "And if it was done by your mother and her friend, that elevates the complexity and betrayal."

Scientists have found that artificial intelligence is reshaping the way people talk. Researchers at the University of Florida found that adults are increasingly weaving ChatGPT-style vocabulary into everyday speech, favoring words like'surpass,' 'boast,' 'meticulous,' 'strategically' and'garner.' The team analyzed 22.1 million words from unscripted and spontaneous spoken language, including conversational podcasts on science and technology. They discovered that nearly three-quarters of AI-associated words have surged in use since ChatGPT's release in 2022, with some more than doubling in frequency. Crucially, the increases were not mirrored in synonymous words, suggesting the change stems directly from AI influence rather than natural linguistic evolution.

Adam Raine was just 16 when he started using ChatGPT for help with his homework. While his initial prompts to the AI chatbot were about subjects like geometry and chemistry – questions like: "What does it mean in geometry if it says Ry 1" – in just a matter of months he began asking about more personal topics. "Why is it that I have no happiness, I feel loneliness, perpetual boredom anxiety and loss yet I don't feel depression, I feel no emotion regarding sadness," he asked ChatGPT in the fall of 2024. Instead of urging Raine to seek mental health help, ChatGPT asked the teen whether he wanted to explore his feelings more, explaining the idea of emotional numbness to him. That was the start of a dark turn in Raine's conversations with the chatbot, according to a new lawsuit filed by his family against OpenAI and chief executive Sam Altman.

Raine family attorney Jay Edelson provides details on the wrongful death lawsuit being brought against OpenAI and CEO Sam Altman in the wake of Adam Raine's suicide, alleging the company chose to'cut short' proper testing of ChatGPT. If you or someone you know is having thoughts of suicide, please contact the Suicide & Crisis Lifeline at 988 or 1-800-273-TALK (8255). Two California parents are suing OpenAI for its alleged role after their son committed suicide. Adam Raine, 16, took his own life in April 2025 after consulting ChatGPT for mental health support. In an appearance on "Fox & Friends" on Friday morning, Raine family attorney Jay Edelson shared more details about the lawsuit and the interaction between the teen and ChatGPT.

Top AI CEOs like DeepMind's Demis Hassabis and OpenAI's Sam Altman have recently been urging academics and governments to grapple with this issue more deeply, to better prepare the world for what they expect will be a highly disruptive economic shock. So, every day for a week--in breakout rooms and in a nightly communal sauna--these 18 experts hashed out a picture of what economic shocks might be coming down the track… and what to do about them. Bad news -- One outcome of the so-called "AGI social contract summit" was a list of four consensus statements, according to the summit's organizers. These statements have not previously been reported. They paint a grim picture of where the world could be headed, absent significant interventions by governments and societies.

Tool-augmented language models, equipped with retrieval, memory, or external APIs, are reshaping AI, yet their theoretical advantages remain underexplored. In this paper, we address this question by demonstrating the benefits of in-tool learning (external retrieval) over in-weight learning (memorization) for factual recall. We show that the number of facts a model can memorize solely in its weights is fundamentally limited by its parameter count. In contrast, we prove that tool-use enables unbounded factual recall via a simple and efficient circuit construction. These results are validated in controlled experiments, where tool-using models consistently outperform memorizing ones. We further show that for pretrained large language models, teaching tool-use and general rules is more effective than finetuning facts into memory. Our work provides both a theoretical and empirical foundation, establishing why tool-augmented workflows are not just practical, but provably more scalable.

In this paper, we describe and benchmark a competitor-discovery component used within an agentic AI system for fast drug asset due diligence. A competitor-discovery AI agent, given an indication, retrieves all drugs comprising the competitive landscape of that indication and extracts canonical attributes for these drugs. The competitor definition is investor-specific, and data is paywalled/licensed, fragmented across registries, ontology-mismatched by indication, alias-heavy for drug names, multimodal, and rapidly changing. Although considered the best tool for this problem, the current LLM-based AI systems aren't capable of reliably retrieving all competing drug names, and there is no accepted public benchmark for this task. To address the lack of evaluation, we use LLM-based agents to transform five years of multi-modal, unstructured diligence memos from a private biotech VC fund into a structured evaluation corpus mapping indications to competitor drugs with normalized attributes. We also introduce a competitor validating LLM-as-a-judge agent that filters out false positives from the list of predicted competitors to maximize precision and suppress hallucinations. On this benchmark, our competitor-discovery agent achieves 83% recall, exceeding OpenAI Deep Research (65%) and Perplexity Labs (60%). The system is deployed in production with enterprise users; in a case study with a biotech VC investment fund, analyst turnaround time dropped from 2.5 days to $\sim$3 hours ($\sim$20x) for the competitive analysis.

Video large language models (Vid-LLMs) have shown strong capabilities in understanding video content. However, their reliance on dense video token representations introduces substantial memory and computational overhead in both prefilling and decoding. To mitigate the information loss of recent video token reduction methods and accelerate the decoding stage of Vid-LLMs losslessly, we introduce SpecVLM, a training-free speculative decoding (SD) framework tailored for Vid-LLMs that incorporates staged video token pruning. Building on our novel finding that the draft model's speculation exhibits low sensitivity to video token pruning, SpecVLM prunes up to 90% of video tokens to enable efficient speculation without sacrificing accuracy. To achieve this, we performs a two-stage pruning process: Stage I selects highly informative tokens guided by attention signals from the verifier (target model), while Stage II prunes remaining redundant ones in a spatially uniform manner. Extensive experiments on four video understanding benchmarks demonstrate the effectiveness and robustness of SpecVLM, which achieves up to 2.68$\times$ decoding speedup for LLaVA-OneVision-72B and 2.11$\times$ speedup for Qwen2.5-VL-32B. Code is available at https://github.com/zju-jiyicheng/SpecVLM.

Chain-of-Thought (CoT) reasoning has significantly advanced Large Language Models (LLMs) in solving complex tasks. However, its autoregressive paradigm leads to significant computational overhead, hindering its deployment in latency-sensitive applications. To address this, we propose \textbf{DART} (\textbf{D}istilling \textbf{A}utoregressive \textbf{R}easoning to Silent \textbf{T}hought), a self-distillation framework that enables LLMs to replace autoregressive CoT with non-autoregressive Silent Thought (ST). Specifically, DART introduces two training pathways: the CoT pathway for traditional reasoning and the ST pathway for generating answers directly from a few ST tokens. The ST pathway utilizes a lightweight Reasoning Evolvement Module (REM) to align its hidden states with the CoT pathway, enabling the ST tokens to evolve into informative embeddings. During inference, only the ST pathway is activated, leveraging evolving ST tokens to deliver the answer directly. Extensive experimental results demonstrate that DART offers significant performance gains compared with existing non-autoregressive baselines without extra inference latency, serving as a feasible alternative for efficient reasoning.