AI Can Watch Your Screen Now 👀

Welcome, night owls. 🌙

The AI marketing landscape never sleeps, and this week proved it. Google and OpenAI just dropped real-time AI that can watch your screen and help as you work, Gemini released crazy image editing capabilities, and Google's Veo 2 crushes Sora.

While everyone else is focused on the tech, I'm focused on what this means for us marketers.

What's brewing tonight:

• Google & OpenAI's AI can now watch your screen in real-time

• Gemini's image powers let you edit any photo with just words

• Google's Veo 2 is pushing AI video beyond what we thought possible

• 6 posts worth your late-night read

• 4 AI tools I tested (so you don't have to)

Let’s dive in.

🤖 Google & OpenAI can now watch your screen and help in real-time

The rundown: Both Google and OpenAI just launched AI assistants that can actually watch your screen while you work. If you want to try it right now, head to Google AI Studio, click on “Stream Realtime”, share your screen, and the AI can see everything you're doing and help instantly.

The details:

• OpenAI launched GPT-4o vision with instant responses (320 milliseconds) and live video analysis

• Google followed with Gemini Live - you can test it now in Google AI Studio through screen sharing

• Full Gemini release with additional features coming later this year with "Project Astra"

• Both priced around $20/month with free tiers available

• They can handle everything from coding to translation to live video analysis

Why marketers should care: I've been testing this in Google AI Studio, and it's incredible. Imagine working in Photoshop and asking, in real-time, how to make specific changes. Or editing a video and getting instant feedback on improvements. Or writing a newsletter with live proofreading. It's like having a co-creator watching what you do, giving feedback and new inspiration 24/7.

🎨 Gemini's new image powers: Edit any photo with just words

The Rundown: Google just announced something that'll change how we all work with images forever. Gemini 2.0 can now edit photos based on natural conversation - no Photoshop skills needed. Just describe what you want to change, and it happens. While it's in early beta now (limited access), this is about to make visual content creation accessible to everyone.

The Details:

• Generate images from text descriptions (think: "Eiffel tower with fireworks")

• Edit existing images through conversation ("make this more vibrant")

• Create mixed content like illustrated recipes in one go

• Combine multiple images (like putting your cat's photo on a pillow)

• Edit images across multiple chat turns ("make it blue" → "now add snow")

• All generated images include SynthID watermark for transparency

Why marketers should care: I'm terrible at design. Even if I knew Photoshop inside and out, I still wouldn't know what looks good. That's what designers are for - they have the vision. What used to take hours in Photoshop can now happen in seconds with a simple description. For marketing teams, this means faster iterations, more creative testing, and the ability to produce way more visual content without the technical barriers.

🎬 Google's new Veo 2 is pushing AI video beyond what we thought possible

The Rundown: Just days after OpenAI's Sora announcement, Google dropped Veo 2 - their latest AI video model. The examples show an AI that deeply understands cinematography, physics, and human movement. The model creates 4K videos with specific camera movements, lens effects, and extended lengths.

The Details:

• Creates high-quality videos up to 4K resolution and several minutes long

• Understands cinematography terms (try "18mm lens" or "shallow depth of field")

• Improved physics and human movement for more realistic results

• Fewer awkward mistakes like extra fingers that plague other AI models

• Available through VideoFX (Google Labs waitlist) and coming to YouTube Shorts

• All videos include invisible SynthID watermark

Why it matters for marketers: I post a lot about AI video generation because it will reshape marketing as we know it. In 5 years, most brands will have in-house generative AI experts. AI will eliminate some marketing jobs and create entirely new ones we haven't imagined yet.

For those of you interested in getting ahead of this curve (which you should be), I highly recommend checking out Curious Refuge's bootcamp. These tools are becoming more powerful every week, and the time to start learning is now.

📜 Interesting posts

• Midjourney launches Moodboard feature. Check it out here 

• 5 easy ways AI transforms content (that actually work)

• YouTube is cracking down on AI clones of celebrities and creators

• Want to hire a Gen AI creator? Runway launches talent network for brands to find Gen AI experts

• If you are using AI UGC, here is a guide on how to do it compliantly and avoid a 50k fine

• Here's how to use Sora to turn product images into product videos

🛠️ Tools I'm Testing

1 - Argil AI. I decided to use this one to create content with my AI clone. Their feature to add custom gestures is the "chefs kiss" to make it more realistic. Check this out.

2 - SwayID - you can upload AI UGC or influencer posts and make sure they are FTC compliant. $50k fines are no joke.

3 - Poppy AI - scaling my content on TikTok in 2025 is top of my priority list. Here's how I am using PoppyAI to do it.

4 - Kling AI - Here's how to take a product image, use AI to have a model wear it, and then use AI to make the model display what it looks like. Here’s the tutorial to do it.