😺 Ultimate Image AI showdown

PLUS: WestWorld robots in real life?!

Welcome, humans.

Because we’re all AI enthusiasts here, we’re sure at least SOME of you have seen the HBO series Westworld (based on the Michael Crichton novel).

If so, this Torso robot from Clone Robotics might look a little familiar:

“I have come to think of so much of consciousness as a burden, a weight, and we have spared them that. Anxiety, self-loathing, guilt. The hosts are the ones who are free…” — Dr. Ford (shared by Dr. Jim Fan)

If you’ve never seen Westworld before, this looks dead on like the Drone Hosts (which seems intentional). Except, instead of making the costume out of rubber molds, Torso is made up of a bunch of complicated anthropomorphic robot bones.

This reminds us of the time the DIY inventions Youtuber Simone Giertz created her own Westworld robot, which is absolutely hilarious if you’ve never seen it:

“Have you ever seen anything so full of splendor?”

Here’s what you need to know about AI today:

We review all the new AI image models that just dropped.
Google open-sourced its generative text watermarking tool.
OpenAI unveiled 50x faster image generator.
Apple launched its new AI features in beta.

Advertise in The Neuron here.

Canva, Stability, Midjourney, and Ideogram walk into a bar…

Let's take a quick walk down memory lane: A few months back, Black Forest released FLUX, the open-source model behind Elon Musk's Grok image generator. A week ago, Adobe dropped all of its new Firefly AI tools at Max.

Which brings us to everything that everybody launched this week:

Midjourney released a new image editor for modifying uploaded images.
Canva launched Dream Lab, built on their newly acquired image model.
Stability dropped Stable Diffusion 3.5, with better accuracy and style control.
Ideogram unveiled an infinte canvas for editing and extending images.

Even ComfyUI, one of the most popular tools for working with open-source image generators like Stable Diffusion just announced a new easy to download desktop app (waitlist here).

The most intriguing drop might be Midjourney's editor. It offers two key features:

“Edit”: Move, resize, or erase parts of images
“Retexture”: Change content while keeping structure

To use it, you can upload images directly or simply via pasting a URL. This means all it takes to, I don’t know, turn a man into a cat is finding the URL for the right publicly available picture of, say, Steve Buscemi, pasting it into the editor, and prompting the Retexture to “please make this man a cat”:

How’d we do?

While no one would believe our Cat-Buscemi is real, the ability to modify any image is powerful enough that companies from Apple to Midjourney itself are scrambling to add guardrails.

Bonus points: You can upload your creations music generator Suno’s Suno Scenes tool to turn them into a playlist.

FROM OUR PARTNERS

This is the Biggest Breakthrough in Next-Gen Tech Since iPhone—and you can invest now for a limited time.

Elf Labs, a next-gen tech company—with rights to billion-dollar character IP like Cinderella, Little Mermaid, and Snow White—has opened up an investment opportunity to the public, but the round closes soon.

Elf Labs is leading the way in immersive entertainment, with:

100+ trademarks.
400+ copyrights.
12 revolutionary tech patents.

Think hyper-realistic AI powered 3D worlds at scale.

Think virtual reality (VR)—without headsets.

Think global licensed merchandise ($350B industry).

Time is running out. With less than a week left to Invest in Elf Labs, the opportunity to become a part of this game-changing movement ends on October 30.

Invest now before it's too late!

Around the Horn.

Genmo released Mochi 1, a new open-source video generator (code here, demo here). You can try it out for free as soon as you sign up—we used it to make this orange cat eating ramen!
Google released SynthID Text as open-source (Hugging Face, Google), which lets developers and businesses watermark and detect generated text.
OpenAI announced a new system called sCM that makes image generation 50x faster than current methods, generating images through diffusion (transforming noise into a final image) in just two steps instead of hundreds.
Apple released a new developer beta today that introduces Genmoji, Image Playground, Visual Intelligence, Image Wand, and ChatGPT integration.

Treats To Try.

*OpenAI, Anthropic, and Character.ai—they all use Statsig. Why? Because Statsig offers a single integration for feature flags, A/B tests, and product analytics. Build your product with data-driven decisions, not intuition. Try Statsig with 2M events per month free.
Cohere launched Multimodal Embed 3, which helps you search through both text and images in your business documents. Here’s how to use it on Cohere’s platform, or mess around with Cohere in general in the playground here.
Watch this five minute demo of how to set up computer use with Claude that apparently anyone can follow (without coding). Here are the commands you need, and here’s Docker, which you’ll use to run the tool. Remember, computer use is pricy!
You.com is a Perplexity competitor and just introduced Advanced Source Controls, which lets you customize searches by filtering websites, blocking domains, and adjusting search needs from quick lookups to deep research.
Hugging Face announced HUGS, a new service which lets you run popular AI models on your own servers as easily as using OpenAI for $1 an hour (more info here).
Fixify handles your basic IT support tasks like password resets and software installations so your team can focus on bigger projects (raised $25M).
Voice Design from ElevenLabs lets you create custom voices by typing in characteristics like “friendly Australian male” or “elderly storyteller.