- Strange Magic
- Posts
- Hollywood and the AI problem
Hey, Explorer!
This week, Hollywood has an AI problem, while Meta has AI solutions. And, things are getting weird with robots. Let’s dive in.
In this edition of Strange Magic…
⚖️ Hollywood versus AI
🤖 Do robots need to cry?
🖼️ Tome if you want to
🐶 Now I wanna be a dog
Read time: 6 minutes
🔮 What’s New
LA Times
The rise of generative AI either happened at the best or the worst time for Hollywood and the many creatives that keep it going. Both the writer’s and actor’s strikes are in full swing at a moment when studios are looking at new tools that could replace some of the people on the picket lines. And, that’s a big part of what the strikers are fighting against.
Some argue that instead of rebelling against AI, Hollywood needs to develop its own AI systems to advance the industry while protecting its ecosystem and IP. While others are diving into the deep end and embracing a future where we all might be deepfakes.
There are a lot of questions about the future of Hollywood, IP, and what it means to be a celebrity. And, quite frankly, there aren’t a lot of answers. Perhaps we should just turn to Hollywood itself for ideas, as the BBC pointed out this week that the little known 2013 film The Congress eerily predicted today’s Hollywood AI crisis.
The colors are so real!
After enduring a year of ridicule over the whole metaverse thing, Meta is finally starting to turn it around. First, they pantsed Elon Musk’s Twitter with the record-breaking release of Threads, and this week Meta made big moves into AI.
Until now, large-language-models (like what powers ChatGPT) have been dominated by companies gobbling up our data across the web wherever we are using AI. But, Meta has decided to open source the release of their new Llama 2 LLM, meaning anyone can download the model and build their own applications with it, allowing for private LLM use without the middleman. Why would they do that? Potentially to gain traction and utilize eager developers to make their own model better than the rivals. Plus, they are building a paid product for those less code minded. ✋
Not stopping there, Meta also dropped CM3Leon (Apparently pronounced “Chameleon”), a new text-to-image generator that uses a different method than all of today’s well known models. Meta claims that CM3Leon is a breakthrough, with the ability to generate more consistent images with far less compute power. Currently, only demos are available for new model. So, time will tell if it can pass the sax test.
While Meta was grabbing headlines for its AI advancements, it was reported this week that they were quietly making plans to sunset their Quest Pro VR headset before even getting to an expected second edition, bringing into question their desire to fight in a market against Apple’s upcoming Vision Pro. For their part, Meta’s CTO rebuked the claim over the weekend.
In Other News
Where’s my aiPhone?: As expected by probably everyone, Apple is rumored to be on the AI bandwagon, with reports saying they’ve developed an AI training framework, and even a ChatGPT-like bot for internal use. But, no word yet on how these tools might end up in our next iPhones.
Oh my god, you killed South Park: Fable Studios, a once VR startup turned AI startup, showed off its SHOW-1 technology by releasing an entire AI generated South Park spoof. While the technology showcases what AI might be capable of, it also highlights what it’s not yet capable of, nuanced human storytelling.
Putting the AI in Fair: After last year’s controversy about a piece of AI generated art winning at the Colorado State Fair, organizers are saying AI art will be eligible for the competition this year as long as it’s labeled as such.
📖 Good Reads
That’s one sad robot
This essay poses one uncomfortable, but potentially necessary, question about the rapid advancement of AI and robotics. “Is it possible to have a functioning humanoid robot that isn’t capable of suffering?” While a natural answer might be, “why would you even…”, the author argues that there is no way to live amongst robots that control potentially life altering decision making for humans without them understanding the downsides of making the wrong decisions.
Ethan Mollick goes deep on exactly which AI tools to use to accomplish a number of things from various writing tasks, to image and video creation, research, idea generation and more. It may be opinionated, but it’s also valuable insight for getting started with AI tools.
More Reads
⚙️ Try These
Tome: Tome is a web-based presentation builder with heavily integrated AI tools, making it easy to produce slick designs with text prompts and an intuitive interface.
Runway Gen-2 Image-to-Video: The latest update to Runway’s video generator can turn a single image into a short video clip. It’s a very promising tool, but strange results are not uncommon.
Pi: If you want a quick outline for a blog post, talk to ChatGPT. If you want to question why blogs exist, talk to Pi. Billed as your “personal AI”, Pi is much more conversational than other chat bots, with a propensity for asking you questions that might lead to more insightful answers than you’d expect from a machine.
🧪 The Lab
The Lab
Make Fake Sets With This AI Tool For Video: Matti Haapoja explains how to use Adobe Firefly to make AI-generated virtual sets for video creators or ambitious Zoom participants. Looking at you Susan.
✨ We’re Diggin’
The Wizards Next Door
The Wizards Next Door: Someone re-imagined Harry Potter as a bad ‘90s sitcom. 🪄
A.I. could solve some of humanity’s hardest problems: Here’s a great conversation between Ezra Klein and Demis Hassabis, perhaps the most important person you’ve never heard of in AI. Hassabis is in charge of all AI research and product development at Google now, and he has some big goals for the future.
🌤️ Fresh Air
There are other things worth looking at.
#bestlife: This pup reminds us that we should get way too excited about life’s simplest pleasures.
Cool dad jokes: If your dad was from Iceland, these are the jokes he’d tell.
Those angles!: Check out some of the most beautiful mid-century modern homes around.
Gimme that sweet, sweet feedback! We might feature it in the next newsletter.
Well, that’s enough strangeness for one day. Hope you learned something new. If you enjoyed it, please spread the word! Our robot only eats organic. 🤖
And, of course, if you have anything interesting to share, or just want to say hi, please feel free to reach out to us!