New AI Voice freaks out the web, Ray2 is faster now, and new Hedra model

All the news from the AI world for creatives.

Erik Knöbl | AI Video Creator

Mar 09, 2025

Good day, creatives

In today’s Newsletter:

1️⃣ There is a new AI Voice, and everyone is buzzing about it

2️⃣ Ray2 Flash: Faster, more affordable AI Video creation

3️⃣ Hedra drops first-ever ‘omnimodal’ video model

4️⃣ Quick news you may have missed

5️⃣ Quiz: Can you identify AI Art?

6️⃣ Prompt Fight: Futuristic Soldier

7️⃣ Video AI of the Week

8️⃣ Good reads of the Week

Let’s go.

There is a new AI Voice, and everyone is buzzing about it

There’s a new voice AI that’s freaking out the internet: A voice assistant that doesn't just respond to your questions but actually sounds like a human—pausing thoughtfully, breathing, expressing excitement, or offering warm reassurance.

It’s Sesame, and its Conversational Speech Model (CSM) is available here via a demo of voice models Maya and Miles, and they are the most human-sounding AI voice I've ever heard. Give it a try.

Basically, the model understands:

The emotional context of the conversation.
Natural timing, pauses, and emphasis.
When to adjust tone to match the situation.
How to maintain a consistent personality.

The Verge's Sean Hollister described it as:

Why is this important?

Sesame could transform how we interact with technology, bringing us closer to speaking with our devices as naturally as we do with each other. And because the movie Her is no longer sci-fi.

Ray2 Flash: Faster, more affordable AI Video creation

LumaLabs has introduced Ray2 Flash, a new model that delivers 3x faster performance at a third of the cost. Flash enhances Ray’s text-to-video, image-to-video, and audio capabilities, ensuring high-quality, production-ready results. Designed for speed and efficiency, it offers all subscribers the ability to create more, faster, and without limits. Ray2 Flash is available now.

Hedra drops first-ever ‘omnimodal’ video model

The a16z-backed startup unveiled Character-3, which reasons across image, text, and audio to create clips that are smooth and realistic. You can also access third-party models through Hedra Studio, so you don’t have to constantly switch between different platforms.

With Hedra Studio, you can:

Text-to-video and audio-to-video
Character and general-purpose image and video creation
Dynamic backgrounds, text-to-emotion, and top AI model integration

Quick news you may have missed

ElevenLabs is now on the Google Cloud Marketplace to bring voice AI to Enterprise. You can combine text to speech and conversational voice agents with Google Cloud's infrastructure and Gemini 2.0 Flash to deploy human-like voices at scale across customer support, education, entertainment and media production.
Pika has upgraded Pikadditions and Pikaswaps to stunning 1080p resolution, delivering enhanced accuracy and precision for an ultra-realistic experience. These improvements ensure higher-quality results for creators. Now available at Pika.art for all Pika paid plans, with availability in the app coming soon.

Quiz: Can you identify AI Art?

The answer is at the end of the post.

Prompt Fight: Futuristic Soldier

Let’s see how the top AI Art generators compare to each other, with a prompt shared by @arkitek666

Prompt: high-resolution, stylized photograph featuring a futuristic soldier in a desert environment. The subject is wearing a heavily worn, sand-colored cloak with frayed edges, covering a high-tech helmet with a reflective visor displaying a digital interface with a blue circular pattern. The helmet is angular and metallic, suggesting advanced technology. The soldier's attire includes tactical gear with various pouches and a visible patch on the shoulder. The background is a sandy, arid landscape, enhancing the rugged and survivalist theme. The overall color palette is dominated by earthy tones, with the blue digital display providing a stark contrast. The composition focuses on the soldier's upper body, emphasizing the intricate details of the gear and helmet.