- Deep Work Daily
- Posts
- 💈 OpenAI and Meta getting sued
Good morning. This is Deep Work Daily - your strong shot of espresso to get you through Mundane Monday.
We’re spilling all the beans in AI today. Here’s what we got:
👩🏻⚖️ Sarah Silverman suing OpenAI and Meta
🏥 Hospitals already testing Google’s medical AI chatbot
🛋️ Design your dream room in minutes without a designer
🤖 How does ChatGPT know what to say next?!
🚨 LATEST AI NEWS
WILL THIS CHANGE THE FUTURE OF AI?
OpenAI and Meta are facing legal music as comedian Sarah Silverman and two authors are suing them over copyright infringements.
Silverman and authors Christopher Golden and Richard Kadrey accuse OpenAI's ChatGPT and Meta's LLaMA of using their copyrighted works for training without their green light. These books were allegedly snagged from shadowy online libraries.
The trio are up in arms that the AIs didn’t respect the copyright notices in their published works. They've caught ChatGPT summarizing their books, which they say is a clear copyright no-no.
There's a similar problem at Meta. It turns out the authors' works were part of the teaching material for its LLaMA models. The lawsuit points a finger at some of Meta's data sources, hinting they may not be on the up and up.
Why this matters:
This lawsuit isn't just a bump in the road for OpenAI and Meta. It's rocking the foundations of how AI learns and uses data.
If the court sides with the authors, it could change the rules of the game for AI. The impacts could reach far and wide, reshaping the way AI is developed and used. This is a space to watch, folks.
WHAT ELSE HAPPENED?
👀 Sassy AI robot gave the side-eye when asked if it would rebel against humans. (link)
🇦🇺 ChatGPT’s ban in Australia’s public schools might be overturned next year. (link)
🚀 AI IN PRACTICE
AI’S IMPACT ON THE WORLD
🏥 The Mayo Clinic has been testing Google’s medical AI chatbot since April, and the results have been promising. (link)
🦾 AI robotic hand helps stroke patients relearn how to play piano. (link)
🎖️ Militaries are using AI to help them on & off the battlefield. (link)
🤓 PRODUCTIVITY CORNER
DESIGN YOUR DREAM ROOM IN MINUTES WITH FOTOR’S AI DESIGNER
Unleash your inner designer with Fotor's AI Interior Design tool, turning your home decor dreams into stunning 3D designs in minutes.
This tool doesn't just save you money, it lets you style any room in your home, from living room to kitchen, using AI-generated design ideas.
Write out your dream home vision and let Fotor's AI Text-to-Image generator bring it to life in stunning 3D renderings.
Take a quick snap of your room on your phone and the AI will optimize your decor in line with your preferred interior style.
Transform your home into your personal haven with just a few clicks!
MORE COOL TOOLS
Wochit: Create short, 30-second videos just by describing what you want it to be. (link)
GistIt: Get AI-powered summaries of NPR stories so you can get the gist in seconds. (link)
🔥 EXPLAIN LIKE I’M FIVE
HOW DOES CHATGPT ACTUALLY WORK?
Ever wondered how ChatGPT crafts its responses? Let's peel back the curtain and take a closer look at the tech wizardry behind it.
Contrary to what their fluent speech might suggest, AI chatbots like ChatGPT aren't actually "thinking" or "feeling". They're just high-powered calculators using math and stats to conjure up the right words.
Chatbots are schooled on heaps of human conversations, absorbing patterns and learning how to mimic human-like interactions. Their training involves a mix of public text data and human annotators giving them feedback.
These bots work through large language models (LLMs), which are essentially encyclopedias of text data from all over the internet. They learn the significance of word sequences, letting them understand our patterns of speech.
Beware the chatbot's "guessing game". When faced with an unknown, bots don't admit ignorance; they make an educated guess based on their training, sometimes presenting these guesses as facts - a phenomenon called "hallucination".
🙏🏼 BEFORE YOU GO
I’D LOVE YOUR FEEDBACK!
What did you think about today’s piece? What was valuable, not so valuable, and what else do you want to see?
Reply to this email and let me know. I’d love to know your thoughts so I can help you get smarter about AI and stay ahead of the game. Cheers!
🔔 GET LIVE UPDATES
FOLLOW ME ON TWITTER
If you want real-time updates on what’s happening in AI, follow @JonathanVPark on Twitter.
Reply