🤖 Labels Want AI Artists

PLUS: Nvidia Takes Another Step Up

What's up? You're reading Inclined AI. I’m headed to Vegas this weekend. So if someone else is writing this Monday, it means I used this newsletter as collateral.

Here’s some news to bet on:

  • Google is in talks with Universal Music Group

  • Nvidia announces their new chip (again)

  • Stability AI is struggling, y’all

  • We have a white paper contributed by a reader (I think that’s so amazing, tbh)

Who’s ready for an AI Taylor Swift?

It’s a possibility now that Universal Music is talking to Google about AI voice replication tools. Put another way, Google wants to license artists' songs and voices for AI purposes.

Funny enough, this is the same label that panicked over an AI Drake song that became a hit overnight. They used everything in their legal power to take it down.

So, why the turn of fate?

It’s money or the potential to profit off talent without their involvement. Grimes is an example, but she isn’t the archetype Universal Music wants to replicate.

Outside the US, there’s a growing number of virtual pop stars run by a combination of AI and fans. That’s what the music industry is moving towards.

Yep, the next famous musician may be an illusion.

And the track record overseas shows that it’s a viable model. AI doesn’t get tired or burned out. It doesn’t push back for the sake of artistic voice. Plus, with fans involved, the byproduct is guaranteed to find some sort of general audience.

Not sold on the idea? I get it. After all, can the industry do this without the approval of the artists they license? It’s hard to say.

It depends on the contract and ownership rights, but you can see why the writer's and actors’ strikes are happening. This situation is what they want to avoid.

So, while nothing is set in stone, expect the negotiations to move quickly between Universal Music and Google. An AI popstar is on the horizon.

The only question is if it can sell records like Taylor Swift and Drake.

Meet the AI company that’s built a brand new type of search

What if you took the latest in neural networks and NLP, then married it to a database of more than 5M photos and videos?

Well, you’d get technology that would let you engineer hyper-specific searches to dive through mass amounts of visual content. You’d get a brand-new type of search product.

You’d get, as the AI company Archive calls it, Super Search.

What is it — and why should you care?

At its core, Super Search is an AI search engine for text and video. Think of something, type it out, and Super Search will find it. Super Search works like you think, and it feels like magic.

You can search with images and video, not just text. And you can search for specific things within a video — for example, you could drag-and-drop a product image and Super Search could surface an unboxing video that features that product.

Right now, Super Search is being used by a small number of leading marketers and fast-growing brands, who leverage it to dive through visual context 10x faster than they used to. Because Super Search works for any post you’ve been tagged in on Instagram and TikTok, brands can find new influencers and winning ad creative in a matter of seconds.

Of course, the actual practical implications of AI tech like this are near-limitless.

Archive’s got a whole series of new AI developments coming down the pipeline — with incredibly broad applications. If you work in marketing, or you’re fascinated by the potential of this AI tech for a number of applications, then you’ll want to check Archive out.

Want to see the tech? Try Archive for free.

Computer chip city --ar 2:1 --s 600 --chaos 25 --v 5.1

Nvidia announces some old news but makes it new. 

That’s an oxymoron, but get this: Nvidia announced a fresh, more powerful AI chip called GH200 to replace their H100 offering. It’s coming out next year and is reported to have 282GB memory capacity.

But some of you know that the whole thing sounds fishy.

Why? Because Nvidia already announced the GH200 months ago, but with different details. You can’t make this up.

The assumption is that Nvidia is disguising a delay in production with some clever PR. Announcing the GH200 allows them to shift back the release date without making it look like they are missing deadlines.

Does it matter?

At the end of the day, Nvidia can’t even make enough H100 GPUs to keep up with demand. They are the leader in the space, and that gives them leeway.

We’re all excited for the GH200 anyway. There’s no need for the cloak-and-dagger work, guys.

It’ll be interesting to see if this new chip takes some pressure off the H100 supply and if more startups and researchers can get to building. It’s one of the two constraints in AI right now.

So make it happen, Nvidia!

⭐️ The fault in our Stability AI

😬 When creating headshots AI fixes things that aren’t an issue

📌 An AI DeepFake company can’t seem to stop their projects from falling into bad situations

🤦 AI gives a voice to dead and missing kids, but where’s the moral line?

🧬 Discovering new antibodies is a task that AI excels at

🐭 Disney makes a task force to examine potential use cases for AI as a way to cut costs

📢 Boost your web presence with AI by following this guideline

🧐 Reader contribution: after our piece on moderators and their troubles, one of our readers shared their proposal for a solution; take a look!

Check Out Our Sunday Edition

You can dive into more AI news and topics with us every week by subscribing to our premium edition.

We’ve written about the following:

If you’re not already subscribed, that’s okay. We’re offering a free 7-day trial so that you can read this one. That’s how excited I am to post it.

  • Recast - turn your read-later list into bite-size audio convos (link)

  • Threado - instant answers to queries on Slack, Discord & web (link)

  • Pixcap - mix-n-match 3D elements & style them (link)

  • Neverinstall - your first AI cloud computer (link)

  • Spinach - AI scrum master (link)

  • Wirestock - generate & sell AI art directly from Discord (link)

  • Raizer - where founders find investors (link)

  • CleverContent - generate a content calendar (link)

Reinvent the Gummy Bear

Why is it that when it’s not shaped like a bear, it feels a bit crueler to do? Is it just me?

- That’s it for today. I hope you enjoyed the latest edition of inclined.ai - Davis.