Welcome, humans.

Thanksgiving is around the corner, and you can bet someone's uncle will inevitably drop "AI" into the convo between bites of stuffing or apple pie. If you want to sound like the smartest cousin in the room, bang out our Intro To ChatGPT crash course before Thursday!

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

• ChatGPT-4v is a game-changer for understanding and getting feedback on visuals.

• YouTube + Google released an AI music generation tool.

• That same team developed tech to spot AI-generated tunes.

• Stability AI’s VP of Audio has stepped down.

ChatGPT-4v is an underrated tool for visual workflows.

Ladies and gents, we just finished a Hot Ones challenge and we’ve got a scorching take ready to serve.

ChatGPT WITH Vision is the most impressive AI application right now, and no one is talking about it.

We’re talking about ChatGPT-4’s newfound ability to “see” and “understand” images. Here’s how to use it:

1. Understand any visual. Confused by something you see? Just snap a photo and send it to ChatGPT-4v. This could be complex Excel sheets (example), code snippets, or financial reports.

2. Get feedback on any visual: think slide decks, UX layouts, or data visualizations. Here’s a solid guide on getting feedback for product designs.

3. Recreate (most) any product visual. Folks are uploading screenshots and extracting the code to replicate ‘em. Plus, we stumbled upon this nifty tool tldraw that turns simple sketches into working designs.

By far the best thing we saw this week is Sir David Frederick Attenborough (the Planet Earth narrator) narrating this guy’s life, all powered by GPT-4v and Eleven Labs.

The Neuron’s version would go like:

under the watchful gaze of flickering monitors, armed with nothing but their wits and a relentless curiosity, they embark on a cerebral safari, exploring the vast savannah of artificial intelligence, one newsletter at a time.

FROM OUR PARTNERS

Turn Midjourney into your ultimate creative collaborator.

Midjourney can create hyper-realistic images and boost your creative output and efficiency.

But prompting Midjourney is a complex skill to master.

Meet Nick St. Pierre.

Nick has taught hundreds of designers, creatives, and companies (Meta, Google, Nike, to name a few) to become masters at Midjourney. In his flagship course, you will:

• Learn and use advanced prompting techniques.

• Navigate features like remixing, image blending, negative prompting, and permutations.

• Build your own AI portfolio.

Last chance to join this course before the deadline on November 7th.

YouTube is releasing AI music tools and detecting AI music (at the same time).

In 2023, Big Tech independently developed and self-regulated some of the most sophisticated AI systems in the world.

Just yesterday, Google and YouTube announced an AI music generation tool for select creators in Shorts, named “Lyria”. 

Similar to Meta’s Audiocraft, you just prompt it with “create [tune] in the style of [artist] about [topic]”, and it churns out some seriously cool stuff.

Meanwhile…

The very same team developed a tool “SynthID” designed to identify and watermark these AI-generated tunes. On top of that, YouTube announced plans to label content that’s been altered by AI.

That’s all assuming it can distinguish between AI and non-AI content, which, as our seasoned Neuron readers know, has become way harder.

Around the Horn.

• Microsoft launched Windows AI Studio, a platform for developers and enterprises to customize/deploy small AI models.

• Stability AI’s VP of Audio resigned over disagreements over the use of copyrighted material in AI training.

• Microsoft is bringing Copilot to Windows 10!

• A leaked screenshot shows that GPTs will soon learn and improve based on your chats with it.

Intelligent Insights.

Here are some meow-gical (sorry, we had to) pieces we devoured this week:

1. AI + APIs — What 12 Experts Think The Future Holds (link).

2. The OpenAI Keynote by Stratechery (link).

3. American AI Startups Quietly Raise Money From Top Chinese VC Firms, Including Sequoia Capital China (link).

4. What OpenAI’s latest announcements mean for enterprise businesses (link).

5. Exploring GPTs: ChatGPT in a trench coat? (link).

Friday Faculty.

1. 20 open roles & 4 internships @ Humane (SF).

2. Head of Partnerships @ Cohere (SF/NYC).

3. GPU Programming Expert @ Mistral (Paris/London/SF).

4. 17 technical roles @ Character.ai (Palo Alto).

5. Software Engineer, Product @ Weights & Biases (remote).

A Cat's Commentary.