- Yawn
- Posts
- AI-generated South Park episode sparks outrage
AI-generated South Park episode sparks outrage
PLUS: Apple adds $71B in market value overnight
Colby Shelton
July 21, 2023
Happy Friday from Yawn; the emerging tech newsletter that helps you sound interesting and up-to-date at those company happy hours.
Let’s keep the TGIF energy up with:
AI-generated South Park episode sparks outrage 🤬
Apple adds $71B in market value overnight 📈
DIY LLM’s becoming possible with new startup 🛠
3 Supercomputer Snippets 💤
An unauthorized AI-generated episode of South Park has sparked outrage from major Hollywood guilds during ongoing strike negotiations. The episode was created by using a text-to-video AI tool trained on South Park scripts, allowing them to mimic the show's style and characters without consent.
Fable Studios, the organization behind the creation of the episode, announced via Twitter that it had published a paper on “Generative TV & Showrunner Agents.” They embedded the faux “South Park” episode where Cartman tries to apply deepfake technology to the media industry.
Tap to watch*
Hollywood unions slammed it as unethical theft of IP during tense negotiations with studios. However, its legal status remains murky since AI currently lacks clear copyright laws.
The episode highlighted growing unease around AI's potential to automate creative work. But it also demonstrated AI's limitations, with glitches revealing it's not yet poised to replace writers.
Apple's stock price spiked up to 2.3% to a new record high on Wednesday following a Bloomberg report about Apple's AI plans. The surge briefly added around $71 billion to Apple's total market valuation, bringing it to an all-time peak of $3.12 trillion.
Up until this past week, Apple has been pretty quiet about AI compared to rivals like Microsoft, Alphabet and others with public AI strategies. Then they released that they’ve been developing their own generative AI tools to compete with OpenAI's ChatGPT success.
Namely their project codenamed Ajax, including an AI chatbot some have dubbed "AppleGPT". This could be the initiative that has Apple aggressively hiring for roles in machine learning, natural language processing and other AI specialties.
CEO Tim Cook has made it clear in the past that Apple’s goal is never to be first, but to be better.
Lamini AI is an open source library that can be used to easily train advanced language models comparable to ChatGPT. It enables developers without extensive machine learning expertise to create custom LLMs.
The startup’s platform automates complex processes like data preprocessing, distributed training, model scaling and more. The library makes techniques like reinforcement learning from human feedback accessible through simple APIs. Developers can build niche industry-specific LLMs using little data and computing resources.
Elon Musk revealed that Tesla will invest over $1B into developing its Dojo supercomputer specialized for autonomous driving. Dojo's goal is to process vast amounts of visual data for neural network training faster than other systems. Its architecture uses Tesla's own D1 AI training chips to deliver nearly 1 exaflop of computing power.
AI startup Cerebras unveiled its new CS-3 system to challenge Nvidia's dominance. The CS-3 packs 850,000 AI processing cores, 3x more than Nvidia's flagship H100 chip. Nvidia still controls nearly 80% of the accelerators optimizing today's AI systems. The intense AI chip race highlights the surging computational demands of AI workloads.
Scientists are leveraging Oak Ridge National Lab's Frontier supercomputer for advanced climate research. Their high performance computing aids complex, high-res climate modeling with far greater speed helping orgs like the U.S. Air Force predict the weather with hyper-local forecasts.
