Chinese open models are spreading fast, from Hugging Face to Silicon Valley. In this photo illustration, the DeepSeek apps is seen on a phone in front of a flag of China on January 28, 2025 in Hong Kong, China. The past year has marked a turning point for Chinese AI. Since DeepSeek released its R1 reasoning model in January 2025, Chinese companies have repeatedly delivered AI models that match the performance of leading Western models at a fraction of the cost. Just last week the Chinese firm Moonshot AI released its latest open-weight model, Kimi K2.5, which came close to top proprietary systems such as Anthropic's Claude Opus on some early benchmarks. The difference: K2.5 is roughly one-seventh Opus's price.

Which chatbot we choose could become an extension and reflection of our personalities, like the clothes we wear or car we drive. Which chatbot we choose could become an extension and reflection of our personalities, like the clothes we wear or car we drive. From'nerdy' Gemini to'edgy' Grok: how developers are shaping AI behaviours Do you want an AI assistant that gushes about how it "loves humanity" or one that spews sarcasm? How about a political propagandist ready to lie? If so, ChatGPT, Grok and Qwen are at your disposal. Companies that create AI assistants, from the US to China, are increasingly wrestling with how to mould their characters, and it is no abstract debate.

WIRED analyzed more than 5,000 papers from NeurIPS using OpenAI's Codex to understand the areas where the US and China actually work together on AI research. The US and China are, by many measures, archrivals in the field of artificial intelligence, with companies racing to outdo each other on algorithms, models, and specialized silicon . And yet, the world's AI superpowers still collaborate to a surprising degree when it comes to cutting-edge research. A WIRED analysis of more than 5,000 AI research papers presented last month at the industry's premier conference, Neural Information Processing Systems ( NeurIPS), reveals a significant amount of collaboration between US and Chinese labs. The analysis found that 141 out of the 5,290 total papers (roughly 3 percent) involve collaboration between authors affiliated with US institutions and those affiliated with Chinese ones.

An AI model that learns without human input--by posing interesting queries for itself--might point the way to superintelligence. Even the smartest artificial intelligence models are essentially copycats. They learn either by consuming examples of human work or by trying to solve problems that have been set for them by human instructors. But perhaps AI can, in fact, learn in a more human way--by figuring out interesting questions to ask itself and attempting to find the right answer. A project from Tsinghua University, the Beijing Institute for General Artificial Intelligence (BIGAI), and Pennsylvania State University shows that AI can learn to reason in this way by playing with computer code.

After 10 years of journaling, my only regret is not starting sooner. Keeping a daily diary doesn't come easily to most people, but it takes less effort than you might imagine. It could also become a meaningful way to reflect and grow as a person. For more than 10 years, I've written a few words every morning, and what I've learned from this practice has changed my life. My only regret is not starting sooner.

Google's AI is now even smarter, and more versatile. Gemini Live is the more conversational, natural language way of interacting with the Google Gemini AI bot using your voice. The idea is you chat with it like you would chat with a friend, interruptions and all, even if the actual answers are the same as you'd get from typing your queries into Gemini as normal. Now, about a year and a half after its debut, Gemini Live has been given what Google is describing as its "biggest update ever." The update makes the Gemini Live mode even more natural and even more conversational than before, with a better understanding of tone, nuance, pronunciation, and rhythm.

In the AI boom, chatbots and GPTs come and go quickly. On a drizzly and windswept afternoon this summer, I visited the headquarters of Rokid, a startup developing smart glasses in Hangzhou, China. As I chatted with engineers, their words were swiftly translated from Mandarin to English, and then transcribed onto a tiny translucent screen just above my right eye using one of the company's new prototype devices. Rokid's high-tech spectacles use Qwen, an open-weight large language model developed by the Chinese ecommerce giant Alibaba. OpenAI's GPT-5, Google's Gemini 3, and Anthropic's Claude often score higher on benchmarks designed to gauge different dimensions of machine cleverness.

We investigate how LLMs encode sociodemographic attributes of human conversational partners inferred from indirect cues such as names and occupations. We show that LLMs develop linear representations of user demographics within activation space, wherein stereotypically associated attributes are encoded along interpretable geometric directions. We first probe residual streams across layers of four open transformer-based LLMs (Magistral 24B, Qwen3 14B, GPT-OSS 20B, OLMo2-1B) prompted with explicit demographic disclosure. We show that the same probes predict demographics from implicit cues: names activate census-aligned gender and race representations, while occupations trigger representations correlated with real-world workforce statistics. These linear representations allow us to explain demographic inferences implicitly formed by LLMs during conversation. We demonstrate that these implicit demographic representations actively shape downstream behavior, such as career recommendations. Our study further highlights that models that pass bias benchmark tests may still harbor and leverage implicit biases, with implications for fairness when applied at scale.

With the rapid advancement of low-altitude remote sensing and Vision-Language Models (VLMs), Embodied Agents based on Unmanned Aerial Vehicles (UAVs) have shown significant potential in autonomous tasks. However, current evaluation methods for UAV-Embodied Agents (UAV-EAs) remain constrained by the lack of standardized benchmarks, diverse testing scenarios and open system interfaces. To address these challenges, we propose BEDI (Benchmark for Embodied Drone Intelligence), a systematic and standardized benchmark designed for evaluating UAV-EAs. Specifically, we introduce a novel Dynamic Chain-of-Embodied-Task paradigm based on the perception-decision-action loop, which decomposes complex UAV tasks into standardized, measurable subtasks. Building on this paradigm, we design a unified evaluation framework encompassing six core sub-skills: semantic perception, spatial perception, motion control, tool utilization, task planning and action generation. Furthermore, we develop a hybrid testing platform that incorporates a wide range of both virtual and real-world scenarios, enabling a comprehensive evaluation of UAV-EAs across diverse contexts. The platform also offers open and standardized interfaces, allowing researchers to customize tasks and extend scenarios, thereby enhancing flexibility and scalability in the evaluation process. Finally, through empirical evaluations of several state-of-the-art (SOTA) VLMs, we reveal their limitations in embodied UAV tasks, underscoring the critical role of the BEDI benchmark in advancing embodied intelligence research and model optimization. By filling the gap in systematic and standardized evaluation within this field, BEDI facilitates objective model comparison and lays a robust foundation for future development in this field. Our benchmark is now publicly available at https://github.com/lostwolves/BEDI.

Large language models are increasingly used to evaluate other models, yet these judgments typically lack any representation of confidence. This pilot study tests whether framing an evaluation task as a betting game (a fictional prediction market with its own LLM currency) improves forecasting accuracy and surfaces calibrated confidence signals. We generated 100 math and logic questions with verifiable answers. Six Baseline models (three current-generation, three prior-generation) answered all items. Three Predictor models then forecasted, for each question-baseline pair, if the baseline would answer correctly. Each predictor completed matched runs in two conditions: Control (simple correct/incorrect predictions) and Incentive (predictions plus wagers of 1-100,000 LLMCoin under even odds, starting from a 1,000,000 LLMCoin bankroll). Across 5,400 predictions per condition, Incentive runs showed modestly higher accuracy (81.5% vs. 79.1%, p = .089, d = 0.86) and significantly faster learning across rounds (12.0 vs. 2.9 percentage-point improvement from Round 1 to Round 4, p = .011). Most notably, stake size tracked confidence. "Whale" bets of 40,000+ coins were correct ~99% of the time, while small bets (<1,000 coins) showed only ~74% accuracy. The key finding is not that fictional money makes models smarter; accuracy gains were modest and did not reach statistical significance (p = .089) in this pilot. Rather, the betting mechanic created a legible confidence signal absent from binary yes/no outputs. This suggests that simple financial framing may help transform LLMs into risk-aware forecasters, making their internal beliefs visible and usable. The protocol offers a foundation for future work for meta-evaluation systems and what may become LLM-to-LLM prediction markets.