Technology

Remember the days when shooting video required dedicated software like Filmic, a tool I once bought to unlock decent camera controls and settings reminiscent of classic video cameras (yes, old-timers might recall the Canon XL1)? Those were the days of real gear. Now, Blackmagic Design has stepped up the game with Blackmagic Camera, a free app for iPhones that brings nearly all the professional features of their high-end cameras to your pocket.

AMD has confirmed plans to resume shipments of its MI308 accelerators to China, hot on the heels of Nvidia’s announcement just hours earlier about gaining approval to sell its H20 chips. The U.S. Department of Commerce has informed AMD that its export license applications will be reviewed with a strong likelihood of approval, marking a significant policy shift in Washington.

In a surprising move, Chinese tech giant Baidu has announced plans to open-source its generative AI model, Ernie, marking a significant departure from its long-standing opposition to open-source policies. This decision is poised to reshape the competitive landscape of AI development, potentially boosting the adoption of open-source models and influencing pricing dynamics in the AI industry.

The U.S. Department of Defense (DoD) has awarded contracts worth up to $200 million each to OpenAI, Alphabet's Google, Anthropic, and Elon Musk's xAI, marking a significant push to scale advanced AI capabilities within the military. Announced on July 14, 2025, these contracts aim to develop "agentic AI workflows" to address critical national security challenges, according to the DoD's Chief Digital and Artificial Intelligence Office (CDAO). The initiative reflects a strategic effort to leverage cutting-edge commercial AI technologies to maintain a strategic advantage over adversaries.

Perplexity AI has officially unveiled its much-anticipated web browser, Comet, marking a significant step forward in the integration of artificial intelligence into everyday internet use. Announced earlier, Comet is now accessible, bringing to life the concept of "agent AI"—an intelligent system capable of not just searching for information but also thinking, acting, and making decisions on behalf of users.

Google has initiated the rollout of its AI assistant, Gemini, on smartwatches, effectively replacing Google Assistant on wearable devices. The update is now available for Pixel Watch users and will extend to models from Samsung, OPPO, OnePlus, Xiaomi, and other manufacturers running Wear OS 4 or later in the coming weeks.

Hugging Face, renowned for its contributions to open-source AI through its vast ecosystem of machine learning models, datasets, and tools, has taken a bold step into the physical world with the launch of Reachy Mini. This expressive, compact, and fully open-source robot is designed to democratize robotics and AI experimentation, making it accessible to developers, educators, hobbyists, and even kids coding at home. Priced starting at just $299, Reachy Mini combines affordability with powerful AI integration, offering a unique platform for human-robot interaction, creative coding, and real-world AI application development.

Google DeepMind’s latest innovation, Veo 3, has taken AI video generation to new heights by enabling the transformation of a single photo into a fully animated video complete with synchronized audio and lifelike movement. Upload an image, and watch as the character comes to life—speaking, moving, and maintaining consistent appearance across every frame with impeccable lip-sync.

To the untrained eye, the current surge in venture capital flowing into tech startups—particularly in Silicon Valley—might look like a bubble, a frenzy, or even a fever dream. It’s an understandable reaction, but the uninitiated often miss the bigger picture. Here’s why this perception exists—and why it’s both right and wrong.

Adobe has just dropped a game-changer for iPhone photographers with Project Indigo, a free, professional-grade camera app that transforms your smartphone into a cinematic powerhouse. Available now on the App Store, this experimental app from Adobe Labs brings a level of control and quality typically reserved for dedicated cameras, making it a must-have for both casual shooters and serious shutterbugs.

Jensen Huang and the Nvidia team have worked their magic once more, unleashing a groundbreaking speech recognition model that promises to outshine all transcribers. This new tool tackles virtually any speech—whether from videos, phone calls, lectures, or even noisy auditoriums—delivering results that feel almost otherworldly.

Imagine waking up from a vivid dream—say, one where you’re sprinting from a horde of walking wardrobes or sitting in a classroom in your underwear—and being able to watch it play out like a short film. This is no longer the stuff of sci-fi. Modem Works, a forward-thinking startup, has introduced the Dream Recorder, a device that transforms your dream recollections into AI-generated video clips. And the best part? You can build it yourself with a Raspberry Pi, a 3D printer, and open-source code from GitHub