Discover the AI Tools

Insta360 is rolling out an AI voice assistant for its GO Ultra thumb camera, powered by Alibaba's Qwen in mainland China and Google's Gemini elsewhere. The assistant, activated by saying "Hey Kira" or long-pressing the capture button, offers face-to-face translation, photo-based Q&A, and voice Q&A. Founder Liu Jingkang believes this could redefine thumb cameras, signaling AI's push into consumer hardware.

Google's Creative Lab has unveiled Gemma Translator, a Raspberry Pi-powered device that translates conversations in real time without an internet connection. Running a 5.1-billion-parameter AI model locally, it handles speech recognition, translation, and synthesis on-device. This isn't just a tech demo—it's a glimpse into a future where language barriers dissolve, even in the most remote corners of the world.

Stanford and Arc Institute researchers used the AI model Evo to generate 700,000 viral genome sequences, then synthesized 285 and found 16 that could infect and kill E. coli. This marks the first time AI has designed entire functional viruses, but it also raises biosecurity concerns. The study, published in Science, shows both the promise and the risks of generative biology.

Insta360 has rolled out an AI voice assistant for its GO Ultra thumb camera, tapping Alibaba's Qwen in mainland China and Google's Gemini elsewhere. The feature brings face-to-face translation, photo-based Q&A, and voice commands to a tiny 53-gram device. Founder Liu Jingkang believes this could redefine the thumb camera and even become a top-selling personal AI assistant. Here's how it works and what it means for the wearable AI race.

Neon and Castform have trained a 4B-parameter open-source model that outperforms GPT-5.6Sol in document search accuracy, while costing just 1% per inference. This breakthrough challenges the assumption that bigger is always better, and signals a shift toward agent-based search paradigms. The model's success, achieved through reinforcement learning, could make advanced AI search more accessible and affordable for everyone.

小米智能摄像机4 Max AI变焦版于8月7日正式开售，售价799元，京东到手价739元。这款摄像机最大的亮点是内置了小米自研的AI看护大模型，能够识别儿童学习、老人跌倒、宠物活动等场景，并支持自然语言搜索。硬件上配备双摄像头，支持4K画质和12倍混合变焦。AI功能提供三个月免费试用，之后需订阅，每月32.8元或每年349元。

Volcano Engine has rolled out Seedance 2.5, an upgraded AI video generation API that promises better instruction following, longer narratives, and enhanced realism. It can now produce 30-second clips natively and handle up to 50 multimodal references. The update also brings more precise video editing and support for over a dozen languages. To sweeten the deal, they're offering limited-time discounts on older versions, making AI video creation more accessible than ever.

Unitree's founder, Wang Xingxing, recently shared his vision for the future of embodied intelligence. In an online roadshow, he emphasized that going public is just a new beginning. The company plans to double down on core tech like embodied large models and high-performance actuators, while exploring new product forms—from humanoid robots to mechs. The goal? To let robots take over dangerous and repetitive jobs, freeing humans for more creative pursuits. As the industry heats up, Unitree is positioning itself as a full-stack player in the robot ecosystem.

In the high-stakes world of AI, money talks—and it's talking loudly. Alphabet is planning to borrow up to $25 billion through bond sales, while SoftBank is leveraging its OpenAI shares for a $10 billion loan. These moves highlight the immense financial appetite of tech giants racing to dominate artificial intelligence. But with such heavy debts, the question looms: will the AI boom pay off, or are we witnessing a bubble in the making?

ByteDance is reportedly in early talks to train a massive AI model with over 5 trillion parameters, which would surpass Alibaba's and Moonshot AI's largest models. The project, led by Seed Foundation's Xiang Liang, is still unconfirmed but signals a shift toward foundational model competition in China.

OpenAI 宣布取消 ChatGPT 的文本聊天限制，所有用户均可无限制使用。新推出的 GPT-5.6 系列模型在事实错误率上大幅降低，并引入了“思考”按钮和“思考级别”滑块，让用户能更灵活地控制 AI 的推理深度。这一举措标志着 ChatGPT 正从简单的问答工具，向高频、智能的交互平台迈进。

OpenAI is reportedly developing its first AI hardware device—a doughnut-shaped smart speaker without a screen, priced between $300 and $400. Designed with premium materials and unique moving parts, it aims to bring ChatGPT into everyday spaces. But with the smart speaker market's profit challenges and a looming Apple lawsuit, will this device succeed? Here's what we know so far.

OpenAI is set to release its new flagship model, Astra, next week. Leaks suggest it's the largest pre-training since GPT-4.5, with a focus on multi-agent collaboration. Astra has reportedly solved 10 open math problems, hinting at its advanced reasoning. This launch comes amid fierce competition from Google and Anthropic, positioning Astra as a potential game-changer for complex tasks and enterprise AI.

OpenAI's debut hardware is shaping up to be a screenless smart speaker, roughly the size of a hockey puck, priced between $300 and $400. Designed with Jony Ive's LoveFrom, it packs a camera, battery, and a self-moving component, aiming to be an always-on AI companion. But it arrives amid Apple's trade secret lawsuit, adding a layer of intrigue.

AI music platform Suno is taking steps to curb misuse of its generated tracks. The company will add watermarks, restrict downloads, and update community guidelines to prevent counterfeit songs. These moves come amid lawsuits from record labels and a recent court ruling against Suno in Germany. Suno also partnered with Musixmatch for copyright detection, aiming to balance creativity with responsibility.

Yushu Technology's IPO strategic placement revealed DeepSeek's 933,399-share allocation, worth about 141 million yuan, with a 36-month lock-up. Tencent also participated. The company's valuation hits 60.9 billion yuan, but high valuation ratios raise concerns. This marks a milestone for embodied AI investments.

OpenAI has rolled out a major ChatGPT update, upgrading free users to the GPT-5.6 Luna model and lifting chat limits, while paid users get a new slider to control thinking depth. This move aims to attract more users and fend off competitors like Google Gemini and Anthropic Claude.

OpenAI has rolled out a major update for ChatGPT, giving free users unlimited text chats with the new GPT-5.6 Luna model. Plus and Pro users also benefit from an optimized GPT-5.6 Sol in chat, which promises better accuracy and fewer hallucinations. The update also simplifies model selection with a new slider for reasoning effort. This week, all free users will see the changes, marking a significant step in making advanced AI accessible to everyone.

ByteDance is reportedly developing a massive AI model with over 50 trillion parameters, surpassing competitors like Alibaba's Qwen and Moonshot's Kimi. Led by Seed Foundation's Xiang Liang, the project reflects a strategic shift toward long-term self-research, with CEO Zhang Yiming banning distillation of open-source models. This move signals ByteDance's commitment to foundational innovation over short-term gains, aiming to redefine its position in the AI race.

OpenAI has become Microsoft's AI cash cow, contributing $24.1 billion in revenue last fiscal year—about 70% of its total AI income. As Microsoft pushes toward a $37 billion annual AI target, it's also diversifying with self-developed models and open-source advocacy. But this partnership is getting complicated: they're allies and rivals, and the balance between collaboration and competition will shape the future of AI.