A free-to-try Sora competitor!

PLUS: Google taps AI for health wearable insights

Welcome, AI enthusiasts.

The AI video landscape is leveling up — with or without OpenAI’s long-awaited Sora model along for the ride.

Luma’s new ‘Dream Machine’ is the latest model to raise the bar, and its public availability is already leading to some mind-blowing generations. Let’s explore…

In today’s AI rundown:

  • Luma Labs unveils ‘Dream Machine’

  • Samsung’s new AI chip roadmap

  • Transform your photos with AI upscaling

  • Google AI gets personal with wearables

  • 5 new AI tools & 4 new AI jobs

  • More AI & tech news

Read time: 4 minutes

LATEST DEVELOPMENTS

LUMA LABS

Image source: Luma Labs

The Rundown: Luma Labs just launched Dream Machine, a new AI model that can generate high-quality, realistic 5-second video clips from text and image prompts — and is already available to the public, unlike OpenAI’s Sora.

The details:

  • The model generates 5-second clips from text prompts or a reference image with impressive smoothness, consistency, and camera movements.

  • Dream Machine is a transformer model trained on video content, allowing for more coherent outputs that better mimic real-world interactions and physics.

  • Luma says the model can produce 120 frames of video in 120 seconds, though wait times were longer at release with high demand.

  • The tool launched with a free plan offering 30 video generations, with a paid tier extending to 2,000 outputs per month.

Why it matters: While the public still awaits the launch of OpenAI’s Sora, other models like KLING and Dream Machine are already giving users a taste of powerful new capabilities. AI video is taking a massive leap — and each new competitor is raising the bar even higher for when Sora does open to the masses.

Check out some of the best generations from Luma’s launch here.

TOGETHER WITH LIVEKIT

The Rundown: LiveKit enables developers to build multimodal AI experiences with real-time voice and video.

With LiveKit, you can:

  • Build voice AI assistants, customer support agents, video game NPCs, and more

  • Stream voice and video between users and AI models with ultra-low latency

  • Take advantage of plugins for LLMs, turn detection, interruptions, text-to-speech, and more

Start building today and enjoy 100K minutes of free voice every month.

SAMSUNG

Image source: Bloomberg

The Rundown: Samsung just announced a series of AI chip developments, including a new manufacturing roadmap, the latest chipmaking tech, as well as a ‘turnkey’ platform for AI chip solutions.

The details:

  • The company touted new advanced performance in its chip manufacturing process, aiming to compete with TSMC for business from vendors like Nvidia.

  • Samsung said it predicts a 5x increase in its AI customer base and a 9x increase in revenue by 2028, attributed to production upgrades and new AI chips.

  • The ‘Samsung AI Solutions’ platform will aim to create a one-stop AI chipmaking service.

  • Samsung also hired ex-Apple exec Murat Akbacak to lead a new North America AI Center, who previously led Siri efforts.

Why it matters: With tech giants all spending massively in the race to acquire the most advanced custom AI chips, manufacturers like Samsung are looking to break through TSMC’s market dominance and get a bigger piece of the action.

AI TRAINING

The Rundown: Magnific AI is one of the leading AI tools that lets you upscale and enhance your blurry images into extremely high-resolution photos.

Step-by-Step:

  1. Head over to Magnific AI and log in. It’s a paid tool, but if you’re a student of The Rundown University, you’ll get free credits.

  2. On the main dashboard, click "Upload" to select your image.

  3. Choose your scale factor and optimization type. Add a prompt for specific enhancements if needed.

  4. Fine-tune your image with options like Creativity, HDR, Resemblance, and Fractality.

  5. Click "Upscale" to process your image and obtain your new upscaled image.

PRESENTED BY BRILLIANT

The Rundown: Brilliant's expertly crafted courses help you master the fundamentals of AI, math, and programming — helping you stay competitive in today's rapidly evolving landscape in just minutes a day.

Brilliant offers:

  • Bite-sized, interactive lessons that make complex AI concepts accessible and engaging

  • Personalized features and competitive elements to keep you motivated and on track

  • Hands-on problem-solving and real-world applications for impactful learning

P.S. — enjoy 20% off a premium annual subscription, exclusive to The Rundown readers.

GOOGLE

Image source: Google

The Rundown: Google just published two new research papers, introducing an AI model that can interpret personal health data from wearable devices and an AI agent that provides detailed insights, answer health questions, and more.

The details:

  • Google's PH-LLM is a version of Gemini fine-tuned to reason using data from wearables and generate coaching insights for sleep and fitness.

  • PH-LLM performs comparably to human experts on health insights for fitness and sleep, also achieving expert performance on certification exams.

  • Google also revealed PHIA, an AI agent that combines Gemini’s language skills with code and search capabilities to analyze wearable health data.

  • PHIA scored 84% on health insight questions, demonstrating strong reasoning and data analysis capabilities on complex queries.

Why it matters: Google’s research showcases AI’s potential to take health wearable tracking to the next level — going beyond generic insights to enable personalized support. Once ingrained into popular devices, these types of AI tools will put an expert fitness and nutrition coach on the wrists of every user.

NEW TOOLS & JOBS

QUICK HITS

Microsoft released a new research paper presenting DenseAV, an architecture that learns the meaning of words and sounds by watching unlabeled videos.

Midjourney unveiled model personalization, allowing users to better tune outputs to their own styles and preferences by rating images.

Stability AI released the open model weights for Stable Diffusion 3 Medium, with the 2B parameter text-to-image model offering advanced photorealism, prompt understanding, and typography capabilities.

Oracle announced new cloud infrastructure deals with OpenAI and Microsoft to ‘extend Azure AI’s capacity’ and scale deployment of AI models.

OpenAI has reportedly doubled its annual revenue to $3.4B over the last six months, according to The Information.

Pika Labs rolled out a new model update to its video generation platform, showcasing improved quality.

THAT’S A WRAP

SPONSOR US

Get your product in front of over 600k+ AI enthusiasts

Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Get in touch today.

FEEDBACK

How would you rate today's newsletter?

Vote below to help us improve the newsletter for you.

Login or Subscribe to participate in polls.

If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.

Join the conversation

or to participate.