Mon, July 13 at 6:00 AM
You can now interrupt ChatGPT Voice mid-sentence 💬
🎧 ok so ChatGPT actually talks back and forth like a real person?

Yeah, and it’s really good. OpenAI just rolled out GPT-Live, a new set of voice models that power ChatGPT Voice and are built to sound natural and handle real back and forth without the awkward pauses.

Did you ever think AI voice chat would get THIS good?

Login or Subscribe to participate

🗣️ what makes it different from the old voice mode?

The new models can listen and speak at the same time, so you can cut in mid sentence and it keeps up. That full duplex trick also unlocks much smoother live translation.

.... and this ties into the big 5.6 thing everyone's posting about?

Yep. It landed the same week OpenAI opened its GPT-5.6 family, Sol, Terra and Luna, to everyone, so the voice upgrade rides in on a much bigger rollout. The mini version is quietly replacing the old Advanced Voice Mode by default.

Meta guns for the coding throne 👑
💻 wait, Meta has a serious coding model now too?

It does. After last week’s news about Mark Zuckerbergs disappointment in it’s AI agents, Meta just pushed Muse Spark 1.1. This is its strongest model yet for coding and agent style tasks, three months after the first version quietly showed up for a handful of partners.

🏷️ can normal developers actually get it?

Now yes, through a public preview on Meta's developer portal, with a waitlist for fresh signups. Meta's AI chief pitched the pricing as very aggressive and attractive next to the competition.

any catch?

It wouldn’t be Meta if there wasn’t one. For now, it only runs on Meta's own setup, not open marketplaces like OpenRouter.

Over the next few days and weeks, we’ll learn whether this is a true contender or another flop. But, for now, Zuckerberg’s betting that lower costs will lure coders to its platform.

weekly scoop 🍦
🧭 weekly challenge: turn old clips into something new
what's the challenge?

This week, use Google Photos' new Video Remix to turn a boring old clip into a share worthy moment in a few taps.

📱 Step 1: Open Google Photos. Update the app and pick any video sitting in your camera roll, ideally something with movement or a few people in it.
🎬 Step 2: Hit Video Remix. Open the clip, find the Remix option, and let it rework your footage into a brand new style automatically.
Step 3: Pick a vibe and share. Try a few looks, keep the one that makes you laugh, and send it to someone who was actually there.

Is GPT-Live the update that finally makes talking to an AI feel normal, or just Advanced Voice Mode with better manners? And what will you build with Muse Spark 1.1? We'd love to hear your thoughts!

Zoe from Overclocked 

Keep Reading