Hey there!

I’ll be honest, I was supposed to be productive this weekend. But, I got completely sucked into the latest AI news.

The future of tech is just too cool to ignore! Now, I’m here to share the good stuff with you.

Grab your favourite drink, relax for a bit, and let’s dive into the AI updates that are making waves this week.

What’s Inside Today’s Newsletter

🪰 Buzz Around AI

🪄 AI Creation

🔹 AI Learning

♨️ What’s Hot In AI

📰 AI News

Read time: 10 mins

🪰 Buzz Around AI

OpenAI’s Sora Video Model Gets Exposed

"Sora PR Puppets," a protest group, leaked OpenAI’s unreleased Sora video model on Hugging Face.

And honestly, I wasn’t expecting this!

We got an unexpected look at the new features, and while it’s exciting, it also brings up some serious questions.

It feels like a big turning point in the AI world leaks like this don't happen every day. Kinda makes you wonder what OpenAI is up to behind the scenes.

What We Know:

• The protest group claims OpenAI had hundreds of artists test the model for free—while controlling all the content they created.

• The leak stayed live for a few hours before it was taken down, but not before users saw OpenAI’s watermark on the generated clips.

• The leaked model generated 1080p clips in seconds—way faster than the expected 10-minute render times.

• There are rumours that OpenAI is working on a new version of Sora with features like in-painting and image generation to fix those slow render times.

Why This Matters: While other companies rush to release their AI video tools, Sora has stayed behind closed doors. This leak gives us a peek at OpenAI's progress, but it’s not a game-changer yet.

However, it does shed light on some tension with the artists who’ve been testing it could this leak shake things up?

AI Beats Experts at Predicting Scientific Results

A study from University College London shows that AI can predict scientific outcomes better than expert neuroscientists.

It’s also finding hidden patterns in research that could shape future studies.

The Details:

• The researchers used a tool called BrainBench to test 15 AI models and 171 neuroscience experts. They had to tell the difference between real and fake research results.

• AI scored 81% accuracy, while experts only hit 63%. The BrainGPT model, trained on neuroscience papers, did even better—86%.

• This suggests that scientific research follows more predictable patterns than we thought. AI can use these patterns to guide future experiments.

• The team is working on tools that help scientists check their research plans before they begin. This could save time and resources.

Why It Matters: AI’s ability to spot patterns isn’t new, but using it to predict scientific results could change research forever. Scientists could confirm their ideas before starting experiments, speeding up discoveries and avoiding mistakes.

🪄 AI Creation

Prompt:- A close-up photo of a woman in her 35 with green eyes and reddish-brown hair that flows in a Carrie Bradshaw style, wavy. She looks straight into the camera. The tone and atmosphere of the photo make it look like she is Eve in the garden. The background should be dark, but there should some be olive leaves in the corners.

🔹AI Learning

Make Videos in Minutes with Google Vids

Google Vids is a new AI tool that makes creating videos easy.

It helps you record, write, edit, and produce videos all in one place.

Step-by-Step:

• Access Google Vids via Google Drive. Click "New" > "Google Vids" (it’s not available for all accounts yet).

• Start by describing your video idea when prompted.

• Customize the AI-generated outline and choose a design style.

• Edit your video using text, media, and audio tools.

Pro Tip: Upload your media by clicking "Media" in the sidebar and selecting "Upload." This personal touch will make your video stand out from those using only stock content.

♨️ What’s Hot In AI

🖥️ Zoom Goes All In on AI with New Rebrand

Zoom has rebranded from Zoom Video Communications to Zoom Communications. It’s shifting focus from video calls to being an AI-first workplace platform.

The Details:

• Zoom’s new tagline is “AI-first work platform for human connection.” The goal? To help people work “happier, smarter, and faster” with AI tools.

• The AI Companion will be central to this, offering more context, web access, and the ability to take action across the platform.

• The rebrand follows new releases like AI Companion 2.0 and Zoom Docs, all aimed at competing with big tech.

• CEO Eric Yuan also shared his plan for AI digital twins, hoping to cut work weeks down to four days.

Why It Matters: Zoom, which became a key player during the pandemic, is now betting on AI to drive its future. This shift shows how AI is becoming essential in every aspect of work—and others will likely follow.

📰 AI News

• Anthropic introduced custom writing styles for Claude. Users can now pick preset tones or train the AI to match their style with sample text uploads.

• Inflection AI's new CEO, Sean White, announced a big change. The company will stop competing in AI model development and focus on enterprise tools through recent acquisitions.

• Perplexity CEO Aravind Srinivas teased a new voice assistant device. He plans to create a sub-$50 device that can “reliably answer questions” after his social media post went viral.

• xAI plans to launch a standalone app to compete with OpenAI’s ChatGPT by December. This will be the company’s first product outside of the X platform.