Meta’s Bold Step Towards Self-reliance in AI
Plus: New AI tool lets you generate 1200 songs/month for free, Adobe is buying videos for $3/minute to train an AI model.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 251st edition of The AI Edge newsletter. This edition brings Meta’s next-generation of AI chips.
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🚀 Meta unveils next-generation AI chip for enhanced workloads
🎶 New AI tool lets you generate 1200 songs per month for free
💰 Adobe is buying videos for $3 per minute to build an AI model
📚 Knowledge Nugget: How OpenAI can defend itself against LLM competition by
Let’s go!
Meta unveils next-generation AI chip for enhanced workloads
Meta has introduced the next generation of its Meta Training and Inference Accelerator (MTIA), significantly improving on MTIAv1 (its first-gen AI inference accelerator). This version more than doubles the memory and compute bandwidth, designed to effectively serve Meta’s crucial AI workloads, such as its ranking and recommendation models and Gen AI workloads.
Meta has also co-designed the hardware system, the software stack, and the silicon, which is essential for the success of the overall inference solution.
Early results show that this next-generation silicon has improved performance by 3x over the first-generation chip across four key models evaluated. MTIA has been deployed in the data center and is now serving models in production.
Why does this matter?
This is a bold step towards self-reliance in AI! Because Meta controls the whole stack, it can achieve an optimal mix of performance and efficiency on its workloads compared to commercially available GPUs. This eases NVIDIA’s grip on it, which might be having a tough week with other releases, including Intel’s Gaudi 3 and Google Axion Processors.
New AI tool lets you generate 1200 songs per month for free
Udio, a new AI music generator created by former Google DeepMind researchers, is now available in beta. It allows users to generate up to 1200 songs per month for free, with the ability to specify genres and styles through text prompts.
The startup claims its AI can produce everything from pop and rap to gospel and blues, including vocals. While the free beta offers limited features, Udio promises improvements like longer samples, more languages, and greater control options in the future. The company is backed by celebrities like Will.i.am and investors like Andreessen Horowitz.
Why does this matter?
AI-generated music platforms like Udio democratize music creation by making it accessible to everyone, fostering new artists and diverse creative expression. This innovation could disrupt traditional methods, empowering independent creators lacking access to expensive studios or musicians.
Adobe is buying videos for $3 per minute to build an AI model
Adobe is buying videos at $3 per minute from its network of photographers and artists to build a text-to-video AI model. It has requested short clips of people engaged in everyday actions such as walking or expressing emotions including joy and anger, interacting with objects such as smartphones or fitness equipment, etc.
The move shows Adobe trying to catch up to competitors like OpenAI (Sora). Over the past year, Adobe has added generative AI features to its portfolio, including Photoshop and Illustrator, that have garnered billions of uses. However, Adobe may be lagging behind the AI race and is trying to catch up.
Why does this matter?
Adobe's targeted video buying for AI training exposes the hefty price tag of building competitive AI. Smaller companies face an uphill battle—they might need to get scrappier, focus on specific niches, team up, or use free, open-source AI resources.
Enjoying the daily updates?
Refer your pals to subscribe to our daily newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: How OpenAI can defend itself against LLM competition
This article discusses a challenge faced by companies selling large language models (LLMs) - frequent competition and churn. The author,
, proposes that LLM providers should prioritize building developer tooling instead of solely on model quality. This would create a more comprehensive developer experience and make it more difficult for developers to switch to competitors, even if a competitor briefly offers a better model. In short, the author argues that by investing in developer tooling, LLM providers can create a sticky ecosystem that transcends individual models.Why does this matter?
In the field of AI, prioritizing developer tooling is crucial. These tools enhance productivity, foster collaboration, reduce barriers to entry, ensure stability, and contribute to long-term strategy. By investing in robust toolsets, AI providers create a sustainable ecosystem that empowers practitioners and drives progress.
What Else Is Happening❗
🤖 Google expands Gemma family with new models
Google has expanded its Gemma family with two new models: CodeGemma and RecurrentGemma. CodeGemma is tailored for developers, offering intelligent code completion and chat capabilities for languages like Python and JavaScript. RecurrentGemma is optimized for efficiency in research, utilizing recurrent neural networks and local attention. (Link)
🌐 Mistral unveils Mixtral-8x22B open language model
Mistral AI has unveiled Mixtral-8x22B, a new open language model with extensive capabilities. This model, featuring 64,000 token context windows and requiring 258GB of VRAM, is a mixture-of-experts model. Early users are exploring its potential, with more details expected soon. (Link)
📷 Google Photos introduces free AI-powered editing tools
Google Photos is rolling out free AI-powered editing tools for all users starting May 15. Features like Magic Eraser, Photo Unblur, and Portrait Light will be accessible without a subscription. Pixel users will also benefit from the Magic Editor, which simplifies complex edits using generative AI. (Link)
🖼️ Microsoft enhances Bing visual search with personalization
Microsoft enhances Bing Visual Search with personalized visual systems based on user preferences. A patent application reveals that search results will be tailored to individual interests, such as showing gardening-related images to gardening enthusiasts and food-related visuals to chefs. (Link)
🛡️ Sama red team: Safety-centered solution for Generative AI
Sama has introduced Sama Red Team, a safety-centered solution for evaluating risks associated with generative AI and LLMs. This system simulates adversarial attacks to identify vulnerabilities related to bias, personal information, and offensive content, contributing to a more ethical AI landscape. (Link)
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you tomorrow. 😊