😳Apple’s using Google’s AI in IOS 18

ALSO: Stability releases AI-powered video creation tool

Today’s all about Apple and video generation. Years ago we could say, “That photo looks too good! It’s gotta be Photoshop.” It was obvious. Clear. But video generation is right at our doorstep.

While Apple scrambles with pouring $1B into R&D only to not catch up, Google struggles to secure a way for us to know what’s real and fake.

• Apple’s partnering with Google to bring Gemini to IOS.

• Stability upgraded their AI video generation.

• Nvidia dominates GPU chips for AI.

• Youtube labels AI generated videos now.

🤔Apple’s working with Google to make a Gemini-powered IOS

What happened: Bloomberg has reported Apple is partnering with Google to integrate Gemini into IOS features. To be blunt, Apple is giving up the ghost on making their own AI for IOS 18. So they’ll use Google’s AI instead.

The details: 

• Apple is partnering with Google to infuse Gemini into IOS 18’s features.

• Apple has spent over $1 billion on AI development in an effort to catch up to competitors.

• This means Apple’s own efforts haven’t gone as far as they’d want yet.

• Apple also looked into using ChatGPT as well.

• Apple’s own models might power some functions of IOS 18, just not image or text related.

• The partnership could spark issues with a pending antitrust lawsuit against Google.

Why it matters: There’s a good chance we’ll be using Gemini in IOS 18 for images and text, which could pose a problem. Our smartphones are about to get a lot smarter 🧠. Most people already don’t know they’re interacting with AI in their daily lives. So how much is about to change?

This does not discount Apple making it’s own AI. It just means they’ll use something else for now. Still, $1B is a lot of money to drop the ball on delivering an AI for IOS 18.

Wow EVERYTHING went up. The recent AI announcements and the Nvidia conference (a lot of the announcements were made there) bolstered the stocks for sure.

As of 03.18.2024 market close.

Nvidia jumps into the robotics space with their AI platform GR00T. Yes, like Groot from Guardians of the Galaxy 🌳 (only the first was good btw) but with zeroes instead of O’s. Some of the companies Nvidia is partnering with are Agility Robotics, Boston Dynamics, and Tesla. GR00T is a foundation model that will allow developers to create new applications for humanoid robots. Everyone’s racing to be the first to provide useable robots for warehouse work (especially for Amazon, who’s definitely thirsty for these bots).

Did you know? Most robotics companies use virtual environments to test their robots. Basically they load the robots into a videogame, have hundreds of them walk or run through courses, and collect that data for real world use. This is common practice. Pretty cool right? Robots are literally training for the real world using 🎮 videogames.

A healthcare AI startup reaches a $500 million valuation. The company, Hippocratic, is creating AI “agents” to aid hospitals and clinics. It’s part of a new wave of AI investments to create solutions or specific industries. The AI agents have not been field tested yet. In their words, they’re waiting until it’s “safe”. Pretty understandable. This isn’t ChatGPT writing your report, they can’t afford AI hallucinations and wrong answers in a hospital environment. Btw the startup is only one year old. Wow.

🎥Stability released a new video generation model

What happened: Stable Diffusion has released Stable Video 3D (SV3D), an AI model that generates videos from scratch. The model is open source (yay!) and free. And it’s pretty damn impressive. Take that Sora!

The details:

• This is an upgrade from Stable Diffusion Video.

• You can download SV3D today. You’ll need a beefy rig to run it though.

• You can train the AI yourself and create any manner of content you want.

• SV3D comes in two models.

  • SV3D_u: Generates orbital videos from a single image (no camera control).

  • SV3D_p: Extends capability by taking both single images and orbital views, allowing 3D video creation along specific camera paths.

• Key features:

  • Novel-View Generation: Creates new, consistent views of an object from any angle, improving object representation compared to previous models.

  • 3D Generation: SV3D leverages its novel-view generation to optimize 3D NeRF models and meshes, resulting in higher quality 3D output. Uses techniques to reduce issues common in 3D generation like baked-in lighting.

What is Stable Diffusion?

In case you don’t know, Stable Diffusion is an open source (that also means free) image generation model. You can download it on your computer, train it up, then create any images you want. And you can control what it trains on.

That includes NSFW, violence, consistent characters, or anything really. You control the AI’s content limitations entirely, so it can create much more freely than Midjourney and DALLE. The problem is how hard Stable Diffusion is to use and setup. (This is what those fake social media influencers use.)

Now we have the video equivalent.

💰Nvidia secures their AI position with new GPU chips

What happened: Nvidia announced their new generation of GPU chips called the “Blackwell” GB200. Their GPU chips are designed specifically to power AI and cloud computing systems.

The details: 

• Nvidia is releasing their new Blackwell GPUs for AI and cloud computing architecture.

• Nvidia’s a major provider of GPU chips to AI and cloud companies. Nvidia is a vital part of a the AI chain. 

• Nvidia is also releasing new software called NIM that will make it easier to deploy AI models.

Why this matters: Last year Nvidia was worth $1 trillion. Today they're worth $2 trillion. All thanks to AI. Whatever Nvidia does will reflect upon the rest of the AI industry. And if you’re in the industry then your company probably uses Nvidia chips right now. Keep that Nvidia stock, they’re not going anywhere.

🏷️Youtube labels AI videos now

Google’s trying to fight the AI apocalypse ahead of the game. From now on Youtube videos that use realistic AI video generation will require a label so viewers know it’s AI generated. This is mostly a move to fight deepfakes, scams, and government regulations and snip them in the bud before they bloom.

What this means:

• Any video with realistic AI generated content must have a label to warn viewers.

• Labeled videos include realistic altered faces, realistic changes to events or places, and realistic fictional major events.

• Clearly unrealistic content (like animation) doesn’t need to be disclosed/labeled. Same goes for small changes (color changes) and captions.

Reasons for the move:

• Audiences distrust and resent AI. 

• Governments are beginning to regulate AI and fear deepfakes may affect elections.

• Google wants to stay ahead of the curve as Stable Diffusion and OpenAI advance their AI video generation (Sora will likely release this year).

Takeaway: Expect AI generated videos to pop their heads out this year. Google sure is. In two years tv shows and movies will regularly use AI videos.

Earlier this year Tyler Perry cancelled an $800 million studio expansion after seeing what OpenAI’s Sora can do. That’s how big this is. AI videos will be everywhere.

Productivity

🛠️AI Tools

Today we’re showing off AI-powered voice dubbing.

Imagine this: You sit down for an interview. You’re nervous but manage to group all the right words together into something that passes for an interview. Maybe it even sounds good.

There’s just one problem.

Your interview is stuck in one language. Even if you speak different languages, you can’t rerecord the interview.

And even hiring someone to dub over it (which is expensive) won’t look natural because the voice doesn’t match your lips.

Problems like these are solved by AI generated videos. The same technology that deepfakes and scams bystanders can also solve vital production problems.

Piplio dubs naturally over your voice (and even makes breathing sounds) in whatever language you want, all while keeping your voice, tone, pace the same. It even alters the image to make it seem like your mouth is saying the words of that language.

It’s actually pretty astonishing.

Influencers can literally 10x their reach with less effort than it took to edit the base video.

This is just a taste of what AI will do to video very soon. It’s a massive opportunity… if you actually seize it.

We’re at your beck and call!

Got feedback? Rate today’s newsletter by clicking below! 

Got more personal feedback? Fill in the box when you rate us and give us your criticism and feedback.

Thank you for reading! 

❤️Share the (ARTIFICIAL) Love!

Got a friend who wants to appear as smart as you? An AI fanatic or someone shaking in their boots about getting replaced? Share us by clicking the link below and you’ll get our love ;)