Why learn animation? Use text-to-3D instead.

PLUS: Microsoft wants to give you $15,000.

TOGETHER WITH

Good morning, human brains. Welcome back to your daily munch of AI news.

Here’s what’s on the menu today:

  • Create 3D animations with simple text prompts 🚀 🤖 

    NVIDIA and Masterpiece Studio launched a text-to-3D tool.

  • OpenAI stealthily edits its "Core Values" 🥷 🧐

    On its Career page, it changed its values to emphasize its "AGI focus."

  • Microsoft wants to give you $15,000 🖥️ 🤑

    It announced rewards for those who find AI security vulnerabilities.

MAIN COURSE

Easily create 3D animations with text prompts 💻 🔥

Yesterday, we covered NVIDIA’s SteerLM — it’s a new method to customize and personalize your LLM.

Last Friday, NVIDIA and Masterpiece Studio launched Masterpiece X — it’s a text-to-3D AI playground.

How does it work?

It uses NVIDIA’s Picasso GPT-optimized inferencing and API to generate 3D meshes, textures, and animations.

What the hell is Picasso?

Picasso is NVIDIA’s cloud service for building AI-powered visual applications.

Is it hard to use?

Masterpiece Studio claims it operates directly in the browser, requires no prior knowledge, and generates complete 3D models based on your text input.

It claims to bypass constraints like hardware, software, and skill. You can specify the type, appearance, and movement of the 3D model.

So I can claim to be a professional 3D modeler?

Eh, it isn’t suitable for high-fidelity game assets. You can use it for quick, cost-effective prototyping, customization, and more.

Can I use it with Unreal Engine?

The site mentions it is compatible with Blender, Unity, Unreal Engine, and more.

Can I use it now?

We tried to play with it, but it glitched several times, logged us in under someone else’s name, and gave us this error message:

Godspeed.

FROM OUR PARTNERS

Ideas are beautiful. Present them that way.

Ditch the design drama.

Instead, transform even the most intricate of your ideas into stunning slide decks in just minutes with Gamma.

😰 Too many ideas, too little time?

Say goodbye to late-night slide struggles. Let Gamma turn your ideas into visual masterpieces.

🚀 Speed up your design process.

Quickly create captivating presentations.

📱 Go beyond static slides.

Every slide is interactive and mobile-responsive by default.

🎨 No design degree? No problem.

Gamma does all the heavy lifting, so you look like a design pro.

And it’s a web app - so no downloads.

The best part? It’s free.

BUZZWORD OF THE DAY

GPT

You’ve heard of ChatGPT, right? Well, the GPT stands for Generative Pre-trained Transformer.

GPTs are artificial neural networks that are pre-trained on large datasets of text.

Transformers are a type of machine learning architecture use primarily in the field of natural language processing (NLP). They were first introduced in 2017 in the paper "Attention Is All You Need."

SIDE SALAD

OpenAI sneakily changed its values 🥷 🧐

In October, we reported Meta’s release of Llama 2 Long. It didn’t announce it, it stealthily published a paper on arXiv.

Why so sneaky?

It did, what?

OpenAI revised its core values. It eliminated the words "thoughtful," "audacious," and "impact-driven."

It replaced them with words like "intense and scrappy," "team spirit," "scale," and more.

I don’t want to work for them, who cares?

The new value "AGI focus" is the most notable change. It stands for Artificial General Intelligence, a type of AI capable of understanding, learning, and performing any intellectual task like a human.

So, does AGI exist now?

No, but its feasibility differs depending on who you ask. Sam Altman, OpenAI’s CEO, equates it to a median human co-worker that you could hire. Geoffrey Hinton, the Godfather of AI, estimates it could take up to twenty years to develop true AGI.

A LITTLE SOMETHING EXTRA

Microsoft wants to give you $15k 🖥️ 🤑

In August, we reported on The White House’s hackathon — it awarded millions of dollars in rewards to winners of its "AI Cyber Challenge."

Ask what you can hack for your country.

Last Thursday, Microsoft announced its "AI Bug Bounty" program — its rewards range from $2,000 to $15,000 based on the severity and quality of the reported issues.

So, I get $15,000 if Bing Chat has a typo?

Nope, Microsoft is more interested in "jailbreak" prompts that cause Bing AI to breach designed guardrails that prevent inappropriate or problematic behavior.

I’m a hacking expert, how do I get my money?

You just need to inform Microsoft of a previously unknown vulnerability and be able to document it through video or writing.

MEMES FOR DESSERT

YOUR DAILY MUNCH

Think Pieces

AI in EV battery development. How new AI models trained on battery materials data are used to create new molecules for batteries.

Chinese proposes stricter AI regulations. The proposed regulations include security and training restrictions for generative AI models.

Training AI to play Pokemon Red. A video on how it’s done and how you can do it too.

Startup News

YouTube announced "Spotlight Moments." It leverages AI to automatically place advertiser’s content in popular cultural events.

Louis Vuitton joining Stanford’s AI program. It joined Stanford’s HAI Corporate Affiliate Program’s Consumer Goods, Retail & AI focus area.

CapCut, ByteDance’s video editor, unveiled AI business tools. CapCut for Business includes AI tools for marketers, small businesses, and creators.

Research

MIT’s new image model beats diffusion models. The new model, PFGMM++, is inspired by physics and combines Diffusion and Poisson Flow.

LoftQ — a new quantization framework for LLMs that addresses performance degradation when applying quantization.

CodeChain — a framework to improve LLM’s code generation capabilities.

Tools

Gemelo — clone yourself using your actual voice to interact with customers.

Glean — AI-powered search tool for finding jobs.

Cal.AI — an open-source AI scheduling assistant.

Xero.AI — easily create AI models without any coding knowledge.

RECOMMENDED READING

The Intelligence Age The AI Newsletter read by Business Leaders. Written by AI experts from NASA, Stanford, and Goldman Sachs.

TWEET OF THE DAY

Is GPT 3.5 better at chess than GPT-4? Renowned Wharton University Professor, Ethan Mollick, retweets a paper that goes deep into this.

Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.

AI ART-SHOW

Until next time 🤖😋🧠

What'd you think of today's newsletter?

Login or Subscribe to participate in polls.