Discover quality AI tools
Bring together the world’s leading artificial intelligence tools to help improve work efficiency
Articles published by GaryBrown
Nubia NaviX Ultra, the world's first AI agent smartphone, secured MIIT network access approval and is set for launch in September 2026. Equipped with Doubao, it features a system-level AI engine enabling cross-app task automation via natural language commands. Unlike traditional AI phones, it processes data locally for enhanced privacy. This launch signals the transition of AI agent smartphones from concept verification to large-scale commercialization, reshaping mobile human-computer interaction.
Nubia NaviX Ultra, the world's first AI agent smartphone, secured MIIT network access approval and is set for launch in September 2026. Equipped with Doubao, it features a system-level AI engine enabling cross-app task automation via natural language commands. Unlike traditional AI phones, it processes data locally for enhanced privacy. This launch signals the transition of AI agent smartphones from concept verification to large-scale commercialization, reshaping mobile human-computer interaction.
Zhipu reported H1 2026 revenue of 954 million yuan, up nearly 400% year-on-year, driven by its MaaS platform which generated 825 million yuan, representing 86.5% of total revenue. Token usage on the platform surged over 40 times, while paid daily active users grew by 603%. The average API price also rose by approximately 101%, reflecting strong market recognition of its commercial value and service premium. Cloud-based large model services have solidified their position as Zhipu's core commercial pillar, demonstrating significant acceleration in the company's overall commercialization process.
Zhipu reported H1 2026 revenue of 954 million yuan, up nearly 400% year-on-year, driven by its MaaS platform which generated 825 million yuan, representing 86.5% of total revenue. Token usage on the platform surged over 40 times, while paid daily active users grew by 603%. The average API price also rose by approximately 101%, reflecting strong market recognition of its commercial value and service premium. Cloud-based large model services have solidified their position as Zhipu's core commercial pillar, demonstrating significant acceleration in the company's overall commercialization process.
Google plans to deploy 12-15 million TPU v9 AI chips by 2028, potentially surpassing NVIDIA's estimated 12.4 million AI GPU shipments. The in-house ASIC chips, never sold publicly, help reduce computing costs and supply chain reliance. TPU v9's 4 Compute Die design may force multi-supplier strategy, opening opportunities for Intel and Samsung, challenging NVIDIA's AI chip dominance.
Google plans to deploy 12-15 million TPU v9 AI chips by 2028, potentially surpassing NVIDIA's estimated 12.4 million AI GPU shipments. The in-house ASIC chips, never sold publicly, help reduce computing costs and supply chain reliance. TPU v9's 4 Compute Die design may force multi-supplier strategy, opening opportunities for Intel and Samsung, challenging NVIDIA's AI chip dominance.
The US imposed strict export controls on Anthropic's AI models, limiting non-US access and raising monopoly concerns. In response, China's Zhipu open-sourced GLM-5.2 with 1M context and code generation, released under MIT license. Other domestic models like MiniMax and Kimi also embrace openness, fostering an inclusive AI ecosystem.
The US imposed strict export controls on Anthropic's AI models, limiting non-US access and raising monopoly concerns. In response, China's Zhipu open-sourced GLM-5.2 with 1M context and code generation, released under MIT license. Other domestic models like MiniMax and Kimi also embrace openness, fostering an inclusive AI ecosystem.
Anthropic's AI security model Mythos was tested on the widely-used curl codebase. After scanning 176,000 lines of code, it initially flagged 5 potential vulnerabilities. Manual review by the curl security team determined most were false positives or regular bugs, leaving only one low-severity issue confirmed. Curl's founder noted the codebase had already been refined by many tools and audits, making new, significant finds difficult. While AI tools offer advantages in understanding protocols and code context, they currently excel at finding known error types, not novel logic flaws. The test highlights that rigorous security engineering remains foundational for core software safety.
Anthropic's AI security model Mythos was tested on the widely-used curl codebase. After scanning 176,000 lines of code, it initially flagged 5 potential vulnerabilities. Manual review by the curl security team determined most were false positives or regular bugs, leaving only one low-severity issue confirmed. Curl's founder noted the codebase had already been refined by many tools and audits, making new, significant finds difficult. While AI tools offer advantages in understanding protocols and code context, they currently excel at finding known error types, not novel logic flaws. The test highlights that rigorous security engineering remains foundational for core software safety.
Nvidia expands open-source AI efforts with two key moves. It acquired SchedMD, the developer behind the open-source Slurm workload manager critical for HPC and AI infrastructure. Simultaneously, it launched the Nemotron 3 family of open AI models, designed for building efficient AI agents across various task complexities.
Nvidia expands open-source AI efforts with two key moves. It acquired SchedMD, the developer behind the open-source Slurm workload manager critical for HPC and AI infrastructure. Simultaneously, it launched the Nemotron 3 family of open AI models, designed for building efficient AI agents across various task complexities.
OpenAI launches ChatGPT Atlas an AI browser for Mac with plans for all major platforms The browser aims to make ChatGPT the primary surface for search and answers targeting its 800 million weekly users It currently lacks traditional features like ad blockers focusing instead on integrating AI features such as memory to collect more user context and strengthen ChatGPTS central role
OpenAI launches ChatGPT Atlas an AI browser for Mac with plans for all major platforms The browser aims to make ChatGPT the primary surface for search and answers targeting its 800 million weekly users It currently lacks traditional features like ad blockers focusing instead on integrating AI features such as memory to collect more user context and strengthen ChatGPTS central role
OpenAI launched Sora 2, an advanced AI video generator with improved realism, alongside a social app featuring TikTok-style feeds and friend-inclusive "cameos." Available initially in the U.S. and Canada for ChatGPT Pro users, the invite-only platform raises concerns over misuse of likeness and safety.
OpenAI launched Sora 2, an advanced AI video generator with improved realism, alongside a social app featuring TikTok-style feeds and friend-inclusive "cameos." Available initially in the U.S. and Canada for ChatGPT Pro users, the invite-only platform raises concerns over misuse of likeness and safety.
