AI Companies are Buying YouTube Videos Now

Hello, AI Enthusiasts!

Welcome to FavTutor’s AI Recap! We’ve gathered all the latest and important AI developments for the past 24 hours in one place, just for you.

In Today’s Newsletter: 😀

• AI Companies are Buying YouTube Videos Now

• Elon Musk says the world is running out of data for AI training

• Study finds LLM-generated questions differ from human questions

YouTube
🤑 AI Companies are Buying YouTube Videos Now

According to Bloomberg, AI Giants like OpenAI, Google, and more are buying unpublished videos directly from content creators to train their AI models. They pay $1-4 per minute, with better prices for drone video and 3D animation. The deals often pay off handsomely for the creators, sometimes even thousands of Dollars.

Insights for you:

• OpenAI, Google, and other AI companies are directly purchasing unpublished video footage from content creators to train their AI video generators like Sora and Veo.

• The companies are particularly interested in high-quality 4K material and specialized recordings such as drone videos and 3D animations.

• Buying unused footage from content creators gives these firms a rich source of visual data that can be leveraged to improve the performance and capabilities of their AI systems.

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

Sign up now for free and work smarter, not harder.

xAI
😲 Elon Musk says the world is running out of data for AI training

“The cumulative sum of human knowledge has been exhausted in AI training. That happened basically last year,” said Elon Musk, owner of X/Tesla/xAI, during a livestreamed conversation with Stagwell chairman Mark Penn.

Insights for you:

• Elon Musk said that AI companies have run out of data for training their models and have exhausted the sum of human knowledge.

• He suggested that synthetic data, which is the data generated by the AI models themselves, is the next game in AI training.

• Microsoft, Meta, OpenAI, and Anthropic, are already using synthetic data to train flagship AI models. It also saves them a lot of money since this data is not sourced from the internet, but it can also lead to hallucinations and model collapse.

AI Research
❔ Study finds LLM-generated questions differ from human questions

Researchers from UC Berkeley, Saudi Arabia's King Abdullah City for Science and Technology, and the University of Washington did a study on how LLMs generate questions. Their findings show some clear differences between AI and human questioning patterns.

Insights for you:

• Researchers have investigated the properties of questions generated by large language models (LLMs). These questions show a strong preference for descriptive questions that require explanations, while human questioners tend to focus on factual questions.

• Also, while humans often focus their questions on information that appears early in a text, AI models spread their questions more evenly across the entire content.

• These findings can help optimize prompt engineering for question generation with language models.