🎥 Meta’s video generation suite

ALSO: Musk’s driveless robo-taxi is here

Estimated Read Time: 3 minutes

Video generation is going to change how we consume media forever. And nothing makes this more obvious than hearing about Meta’s new suite of AI video generation tools, which will be integrated into their social media apps for public consumption.

• 🤖 Meta’s new video models.

• 🚕 Musk’s robo-taxi is here.

• 💰 AMD’s AI chips to compete with Nvidia.

🤖 Meta’s new video models

What happened: Meta just announced Movie Gen, their suite of AI tools and models for creating and editing video and audio. This is a direct competitor to OpenAI’s Sora.

The details: 

• Movie Gen consists of four parts: a video generation model with 30 billion parameters, an audio model containing 13 billion parameters, a personalized video creation model, and a video editing tool.

• The system can generate high-definition videos up to 16 seconds long based on text prompts, complete with synchronized audio like sound effects and background music.

• It also offers video editing through natural language prompts and allows users to upload a reference image to produce personalized videos.

• Meta asserts that, in human evaluations of video quality and consistency, Movie Gen surpasses competitors like Runway Gen3, Luma Labs, and OpenAI’s Sora. We’ll see about that for ourselves.

• Meta CEO Mark Zuckerberg announced in a post showcasing sample outputs that Movie Gen will be “coming to Instagram next year.” There’ll be even more AI generated content on social media soon.

Why it matters: Video generation is going to be BIG. If Movie Gen is already as good as Sora AND is integrated into social platforms then this could change the way we consume social media.

🧠 OpenAI just introduced their own benchmark for AI agents called MLE-bench, a benchmark for AI agents designed to test how well they perform on real world tasks. MLE-bench is a compilation of 75 curated Kaggle competitions that span a variety of machine learning tasks, including model training, data preparation, and experimentation. Kaggle, for the unaware, are competitions and online contests where data scientists compete to solve complex problems using machine learning, with prizes and recognition at stake. There’s been a lot of contention over recent AI benchmarking tools, so maybe this’ll be a better measurement.

🚕 Musk’s robo-taxi is here

What happened: Earlier this year we reported that Elon Musk would reveal a driverless robo-taxi near the end of the year. Now it’s happening. The taxi will be fully autonomous and have no need for a driver.

The details:

• Tesla is planning to produce a Robotaxi vehicle aimed at making autonomous transportation widely accessible and affordable (supposedly). Musk believes this will solve many of the public transportation problems, especially in the USA where public transports are lacking.

• The robo-taxi will have two doors and no steering wheel or driving pedals.

• It will cost under $30,000, says Musk.

• Additionally, there have been mentions of a larger self-driving vehicle concept, which could potentially carry multiple passengers, though detailed information is limited.

Why it matters: Uber drivers beware. If this becomes widely available and reliable then we could see a shift in multiple industries, including car ownership, taxis/ubers, and rental services. It’s astonishing to think how many industries could be jeopardized by the robo-taxi.

💰 AMD’s AI chips to compete with Nvidia

What happened: AMD launched their new line of AI chips at their Advancing AI 2024 event. The lineup is attempting to compete with Nvidia for their lion’s share of the AI market. (Warning: lots of technical specs ahead.)

The details: 

• AMD introduced the Ryzen AI PRO 300 series, enterprise-focused processors designed for AI-enabled PCs, claiming a 40% performance improvement over Intel’s Core Ultra chips.

• The Instinct MI325X AI accelerator, set to begin production in late 2024, offers 1.8 times greater memory capacity and 1.3 times more bandwidth than Nvidia’s H200 GPU. (Could be a big deal and get them a lot of buyers.)

• The company also unveiled its 5th Generation EPYC CPUs, which CEO Lisa Su called “the world’s best for enterprise, AI, and cloud.”

• Additionally, AMD announced plans for annual AI chip releases, with the MI350X expected in late 2025 and the MI400 slated for 2026.

• To put things in perspective, Nvidia is currently worth over $3 Trillion. But at the end of 2023 they were worth around $1 Trillion. That’s how big the AI boom has been for companies who can provide the chips which AI companies need.

Why this matters: Nvidia is now one of the wealthiest and fastest growing companies in the world, primarily due to the AI boom. If you’ve read this newsletter at all, you’ve seen stories of AI companies dropping billions to buy Nvidia’s GPUs and chips. If AMD can even grab a fraction of that money, they’ll make it big. Nvidia is due for some competition. (Aside from AMD, there’s also competition from Eastern companies too.)

🌙 Halloween is coming…

Source: Midjourney

Give us your feedback!

Got feedback? Rate today’s newsletter by clicking below! 

Got more personal feedback? Fill in the box when you rate us and give us your criticism and feedback.

Thank you for reading! 

❤️Share the (ARTIFICIAL) Love!

Got a friend who wants to appear as smart as you? An AI fanatic or someone shaking in their boots of getting replaced? Share us by clicking the link below and you’ll get our love ;)