Sign Up | Advertise | Podcast | AI University | | | Welcome, AI enthusiasts. | Siri's about to get a new pair of digital eyes, and developers are currently first in line to test-drive them. | But after a sluggish Apple Intelligence rollout and rivals racing ahead with similar features, will the tech giant finally deliver on its AI promises? Let's get into it... | | In today's AI rundown: | Apple preps developers for Siri's AI upgrade Tencent unveils open-source Hunyuan-Large model Turn any UI design into working code Apple exploring smart glasses market 5 new AI tools & 5 new AI jobs More AI & tech news
| Read time: 4 minutes |
|
| | | | APPLE | |  | Image source: Apple |
| The Rundown: Apple just started rolling out new developer tools for upcoming Siri screen awareness features with Apple Intelligence, signaling a major enhancement to the digital assistant's contextual understanding capabilities. | The details: | New 'App Intent APIs' allow developers to make their apps' onscreen content accessible to Siri and Apple Intelligence. The system will enable direct interactions with visible content across browsers, documents, photos, and more — all without screenshot workarounds. Early ChatGPT integration testing is already available in the iOS 18.2 beta, though full-screen awareness features are expected in a future update. The feature will look to compete with recent releases from competitors like Claude's computer use feature and Copilot Vision.
| Why it matters: Apple Intelligence has underwhelmed so far, but evolving Siri beyond voice commands into a context-aware assistant will be a welcomed improvement. Given the lackluster rollouts, these upgrades may require a 'see it to believe it' mindset before adding Apple to the AI leaderboards. |
|
| | TOGETHER WITH SPEECHMATICS | | | The Rundown: Flow by Speechmatics is a brand new API that is redefining voice-driven applications — helping deliver effortless conversations powered by ultra-accurate speech technology across multiple languages. | Flow offers: | Real-time responsiveness to every word, embracing interruptions and crosstalk Industry-first support for group conversations Inclusive understanding of any language, accent, or dialect
| Experience the Flow demo and get started with the API today. |
|
| | TENCENT | |  | Image source: Tencent |
| The Rundown: Tencent just released Hunyuan-Large, a new open-source language model that combines scale with a Mixture-of-Experts (MoE) architecture to achieve performances on par with rivals like Llama-405B. | The details: | The model features 389B total parameters but activates only 52B for efficiency, using innovative routing strategies and learning rate techniques. Hunyuan-Large was trained on 7T tokens (including 1.5T of synthetic data), enabling SOTA performance across math, coding, and reasoning tasks. Tencent's model achieved 88.4% on the MMLU benchmark, surpassing LLama3.1-405B's 85.2% despite using fewer active parameters. Through specialized long-context training techniques, the model also supports context lengths up to 256K tokens, double that of similar rivals.
| Why it matters: Large open-source models are continuing to accelerate. Tencent's impressive results with fewer active parameters could reshape how we think about scaling systems — potentially offering a more efficient path forward instead of simply making models bigger. |
|
| | AI TRAINING | | | The Rundown: v0 can transform screenshots of your favorite interfaces into fully functional code with custom enhancements and interactive features. | Step-by-step: | Take a clear screenshot of the interface you want to replicate. Upload the image to v0 and write a prompt describing what you want to build. Enhance your interface by requesting additional features like animations or responsive layouts. Export the production-ready code and customize it as needed.
| Pro tip: You can paste the code provided into Cursor Composer, which adds it to your project. To learn more about it, check out this tutorial. |
|
| | PRESENTED BY SECTION | | | The Rundown: Join Section and Dan Hendrycks of the Center for AI Safety to explore how humanity can navigate the not-so-distant future of AGI, the most advanced (and potentially dangerous) technology ever created. | Join on Nov 21 for a discussion around questions including: | Can AGI systems really dominate human interests? What aspects of AGI are we not talking about enough? How can we ensure AGI doesn't lead us into a dystopian future?
| Secure your free spot and join the discussion. |
|
| | APPLE | |  | Image source: Midjourney / The Rundown |
| The Rundown: Apple is reportedly taking its first serious steps toward potential smart glasses development with a new internal research initiative called 'Atlas', according to a report from Bloomberg. | The details: | The internal 'Atlas' research program is reportedly currently gathering employee feedback on existing smart glasses products and use cases. The research follows Meta's growing success in the category with its Ray-Ban smart glasses and recent prototype demos of 'Orion.' Apple's Vision Pro headset has faced major adoption challenges since debuting in February, with recent reports of scaled-back production. While a product would be years away, entering the category could align with efforts to reduce the cost and bulkiness of the Vision Pro.
| Why it matters: While the Vision Pro had all the hype, Meta's glasses have had far more success—and this research may be recognition that the future of AR may be everyday glasses rather than bulky headsets. While just an idea for now, Apple glasses could be more appealing as an accessory rather than a complex new system to learn. |
|
| | | | | 💻 Overlay by Crisp - AI-powered website search engine that fights bounce rates while reducing support load 🗂️ Orbit AI by Nifty - Automate project management workflows with AI-powered tools ✍️ Alta 2.0 - AI writing tool offering a more human, personalized content creation experience 🎥 Melies - AI filmmaking software to help transform ideas into stunning movies ⚙️ Squire AI - Customize your code review with natural language
| | |
|
| | | | Former Meta AR lead Caitlin Kalinowski announced she is joining OpenAI to lead the company's robotics and consumer hardware efforts to 'bring AI into the physical world'. | T-Mobile will reportedly pay $100M to OpenAI over the next three years to develop an 'intent-driven' AI platform that can take actions for users and integrate with operations and transaction systems for customer service tasks. | Meta's plans for a nuclear-powered AI facility hit a setback after a rare species of bees were discovered at the proposed site, causing regulatory and environmental issues. | Apple's iOS 18.2 Beta 2 revealed that ChatGPT integration with Siri will include daily usage limits for free users and a $19.99 monthly Plus upgrade option offering expanded access to GPT-4o features and DALL-E image generation. | Amazon secured FAA approval to deploy its new MK30 delivery drones, enabling beyond-line-of-sight flights and moving the company closer to broader autonomous deliveries. | Unitree Robotics posted a new video showcasing demos of its Humanoid G1 and Go2 robots, including a more natural walking gait and enhanced balance and coordination. | Google announced plans for a new AI hub in Saudi Arabia focused on Arabic language models and regional applications, despite previous commitments to distance itself from fossil fuel industry development. |
|
| | | | That's it for today!Before you go we'd love to know what you thought of today's newsletter to help us improve The Rundown experience for you. | | See you soon, | Rowan, Joey, Zach, and Alvaro—aka The Rundown Team | |
|
|
|
Tidak ada komentar:
Posting Komentar