- Bot Eat Brain
- Posts
- ChatGPT's huge multimodal upgrades
TOGETHER WITH
Good morning, human brains. Welcome back to your daily munch of AI news.
Here’s what’s on the menu today:
Amazon and Anthropic $4B collab
Amazon invested $4 billion for 49% of Anthropic’s ownership.
ChatGPT’s huge upgrades
OpenAI added multimodal functionalities to ChatGPT.
Getty Images launches new AI tool 🤓 📸
It’s an image generator trained on Getty Images’ content library.
APPETIZER
Amazon and Anthropic $4B collaboration
Anthropic announced a partnership with Amazon. Amazon will invest $4 billion in Anthropic in exchange for 49% of the company.
Claude = cash.
Nickels n’ dimes:
1/ Amazon (AWS) will serve as the primary cloud provider for Anthropic.
2/ Anthropic will use AWS’s Trainium and Inferentia chips to develop, train, and deploy Claude and its future AI models.
3/ AWS gets exclusive access to Anthropic’s tech, and AWS customers can use Claude via Amazon Bedrock.
Our take: This partnership closely mirrors Microsoft and OpenAI’s partnership. Coincidence? No chance.
BUZZWORD OF THE DAY
Multimodal
AI systems or models that can process and integrate multiple types of data inputs, such as text, images, and voice. This capability allows the AI to understand and generate content across different modes of communication, enhancing its versatility and applicability.
FROM OUR PARTNERS
Train powerful AI models using synthetic data.
They say you are what you eat, and that’s even more true for an AI than it is for you and me.
To make a powerful and useful AI model, you need to start with great data, and lots of it.
Unfortunately, good data doesn’t come easily, or does it?
Enter synthetic data.
You can save lots of money by training your AI using synthetic data. All you need is a sample of 5,000+ subjects to get started, then use MOSTLY AI to expand your data set as required.
It's privacy-safe. There is no risk of data leaks or leaking sensitive personal information because your data is synthetic.
Easily adjust your training set. You can make your real data bigger, or smaller, or give it a different "flavor" by changing the distribution of your data.
Ready to get started?
Try MOSTLY AI’s free tier and generate up to 100k rows of data per day.
The best part? It’s free.
MAIN COURSE
ChatGPT’s huge upgrade
OpenAI upgraded ChatGPT with multimodal functionalities. It now sees, hears, and speaks.
Spooky.
The new features aim to make ChatGPT interact similarly to humans.
The Juice:
1/ OpenAI added voice and image capabilities to ChatGPT.
2/ You can verbally communicate with it and show it photos.
3/ ChatGPT can analyze images, text, screenshots, diagrams, and other visual data.
4/ ChatGPT’s voice feature provides fluid conversation, responding with natural-sounding voices.
Sure.
5/ OpenAI collaborated with voice actors to make five different voices.
6/ ChatGPT can analyze images, text, screenshots, diagrams, and other visual data.
It’s rolling out over the next two weeks.
Our take: You can already download plugins that give ChatGPT these abilities, but not to this level of sophistication. These new capabilities could be game-changers if they perform as promised.
We’ll see.
A LITTLE SOMETHING EXTRA
Getty Images launches AI tool 💻 🤓
Getty Images released a new AI tool. It’s called “Generative AI by Getty Images.”
Nice name, bruh.
It allows you to generate commercial images legally and safely.
Picture this:
1/ Generative AI by Getty Images is powered by NVIDIA’s cloud services.
2/ It’s trained exclusively on Getty Images’ content library.
3/ Visuals produced with the tool come with full commercial licensing rights, including Getty’s royalty-free license.
Our take: What makes it different from something like Adobe Firefly? Not much.
But, it’s good to know that if you get sued for using this tool, Getty Images will foot the bill.
MEMES FOR DESSERT
YOUR DAILY MUNCH
Think Pieces
White House AI regulations? Companies might have to report buying AI computing power and more.
Nat Friedman, former GitHub CEO, offers a $1 million prize. You can get it by deciphering what was originally on these burnt, 2000-year-old scrolls.
An AI writer goes viral and feels guilty. Here is his story and reflection on why AI might be bad for writing as a whole.
Startup News
Anthropic launched a prompt engineering guide. How to get the most out of Claude’s 100,000 token windows.
Meta plans to release AI persona chatbots earlier this week. The bots will imitate celebrities and personalities, with hopes to drive engagement among Gen Z.
Correcto, an AI startup, raises $7 million. The money will be used to build an AI-powered “Grammarly for Spanish.”
Research
Reversal Curse — LLMs can’t answer seemingly simple questions because of the way they are fine-tuned. Here’s why it happens and how you can prevent this.
DualToken-ViT — a vision transformer model that combines convolutional neural networks (CNNs) and self-attention-based vision transformers (ViTs) to process data.
MetaMath — an LLM specifically designed to enhance complex mathematical problem-solving abilities.
Tools
QrGPT — an AI-powered QR code generator.
Superflows — quickly build an AI assistant for your product.
HelperAI — an AI-powered tool to create AI agents to automate workflow.
Chatty Butler — an IOS app that allows you to create unique AI assistants for different tasks.
RECOMMENDED READING
If you like Bot Eat Brain, we think you’ll love this newsletter too:
|
TWEET OF THE DAY
Nat Friedman, former GitHub CEO, compares Gemini to OpenAI’s GPT-4.
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? |