Researchers have developed a new method to verify the authenticity of large language models by analyzing their unique 'behavioral fingerprints,' achieving 59.5% accuracy in identifying a model's docum…
A developer's analysis of open-source AI APIs versus self-hosting reveals that self-hosting costs are often underestimated, with hidden expenses like DevOps engineer time and monitoring making API cal…
Tuyaai launches an AI-powered sports and health management platform featuring calorie tracking, sleep analysis, and a fitness coach, alongside a suite of AI tools including note transcription, transla…
Evaluation-Driven Development (EDD) is emerging as the critical discipline in AI engineering, according to a senior QA professional. With frontier models from labs like Anthropic, OpenAI, Google, and …
Al Jazeera's AJLabs asked nine leading AI models to predict the 2026 FIFA Men's World Cup champion, with France emerging as the favorite to lift the trophy after receiving five of the nine champion vo…
China's AI-native applications reached 499 million monthly active users (MAU) by May 2026, an 85.4% surge year-over-year, according to a QuestMobile report published July 14. ByteDance's Doubao leads …
A developer who initially believed self-hosting open-source AI models was the obvious cost-saving move discovered the opposite after running detailed numbers. The engineer found that API-based open-we…
China’s AI-native apps reached 499 million monthly active users by May 2026, up 85.4% year-over-year, according to QuestMobile’s 2026 first-half report released July 14. Doubao led with 382 million us…
A developer benchmarked 15 AI models on speed and cost through Global API's infrastructure, finding that Qwen3-8B delivers 70 tokens per second at $0.01 per million output tokens—99% cheaper than prem…
An indie hacker spent a month testing Chinese AI APIs from DeepSeek, Qwen, Kimi, and GLM, finding that DeepSeek V4 Flash offers 95% of GPT-4o quality at 2.5% of the price. The developer recommends Dee…
A 30-month study of over 26,000 Chinese students found that using generative AI tools like Alibaba's Qwen boosted homework scores by 18% and cut completion time from 64 to 45 minutes, but monthly exam…
A study of more than 26,000 middle and high school students in central China found that using AI homework tools boosted homework scores by 18% but cut exam scores by 20%, with the negative effect on h…
AI has been used in every stage of a cyberattack for the first time, according to research released Monday by cybersecurity firm Check Point. The technology generated commands, tested vulnerabilities …
Anthropic accused Alibaba's Qwen AI lab of orchestrating the largest known AI model theft, using roughly 25,000 fake accounts to generate over 28.8 million interactions with Claude between April 22 an…
A coding bootcamp graduate discovered that open-source AI models like Qwen3 and DeepSeek offer comparable quality to GPT-4o at a fraction of the cost, with some models priced as low as $0.01 per milli…
ZETIC has released a curated catalog of open-source on-device AI apps that run entirely on the phone's NPU with no cloud dependency, eliminating compliance costs and per-token fees. The collection inc…
MindsDB has launched MindsHub Cowork, an open-source unified workspace that lets users delegate entire projects—apps, websites, research, analysis, and scheduled operations—to AI agents while swapping…
A developer benchmarked 10 large language models on five coding tasks, finding that DeepSeek V4 Flash offers the best value-to-quality ratio at $0.25 per million output tokens, while Qwen3-Coder-30B e…
Together AI closed an $800 million Series C at an $8.3 billion valuation, reporting $1.15 billion in annual bookings and an inference engine that hits 500 tokens per second on DeepSeek-V3.1. The compa…
A researcher found that linear probes on model internals add little value for detecting reward hacking in GRPO training when the hack is already verifiable from the model's output. Training Qwen2.5-0.…