Stability AI Launches Text-to-Music AI 🎸🎵
Plus: Adobe's slew of AI updates. Microsoft Research’s new language model trains AI cheaper and faster.
Hello, Engineering Leaders and AI Enthusiasts!
Welcome to the 105th edition of The AI Edge newsletter. This edition brings you Stability AI’s new AI product for the Music and Sound generation.
And a huge shoutout to our incredible readers. You all rock! 😊
In today’s edition:
🎵 Stability AI launches text-to-music AI
👏📜Adobe's slew of AI updates
🆕 Microsoft Research’s new language model trains AI cheaper and faster
🧠 Knowledge Nugget: Can Large Language Models Reason? by
Let’s go!
Stability AI launches text-to-music AI
Stability AI has launched Stable Audio, a music and sound generation product. Stable Audio utilizes generative AI techniques to provide faster and higher-quality music and sound effects through a user-friendly web interface.
The product offers a free version for generating and downloading tracks up to 45 seconds long and a subscription-based 'Pro' version for commercial projects with 90-second downloadable tracks. Stable Audio allows users to input descriptive text prompts and desired audio length to generate customized tracks. The underlying model was trained using music and metadata from AudioSparx, a music library.
Why does this matter?
Stable Audio is the first music generation product that enables the creation of high-quality, 44.1 kHz music for commercial use via latent diffusion. It benefits end users by offering free and subscription-based options for customized music and sound effects, making audio creation accessible.
Potentially help empower content creators, businesses, and artists with an efficient tool to enhance their projects and creativity through tailored soundscapes.
Adobe's slew of AI updates
Adobe has announced new AI and 3D features in Adobe Premiere Pro and Adobe After Effects, as well as enhanced storage capabilities in Frame.io. The new features in Premiere Pro include Enhance Speech, which uses AI to remove background noise and improve dialogue quality, and Text-Based Editing improvements such as filler word detection.
After Effects introduces a true 3D workspace and an AI-powered Roto Brush for easier object removal. Frame.io has introduced Storage Connect for Enterprise customers, allowing them to use their existing storage while maintaining control of their assets. These features are available in beta now and will be generally available later this fall.
Also, another big update is Adobe's Firefly generative AI models are now generally available in its Creative Cloud, Adobe Express, and Adobe Experience Cloud. Firefly features like generative fill and generative expand in Photoshop are now accessible without beta installation.
Adobe is also launching Firefly as a standalone web app. The company plans to charge for Firefly using "generative credits" that measure user interactions with the models. Adobe is now paying bonuses to contributors of its stock image service, Adobe Stock, whose content is being used to train its generative AI model, Firefly.
Why does this matter?
Adobe's Firefly generative AI models becoming generally available and accessible without beta installation signifies democratizing AI-powered creative tools. Also potentially increasing productivity and creative output while rewarding contributors and furthering AI training efforts through Adobe Stock.
Microsoft Research’s new language model trains AI cheaper and faster
Microsoft Research has developed a new language model called phi-1.5 that could make training AI models cheaper and faster. The model uses curated synthetic data from existing large language models like OpenAI's ChatGPT.
Despite having only 1 billion parameters compared to models with over 100 billion inputs, phi-1.5 has shown promising abilities with eliminating the need for web scraping or relying on data sources with copyright issues.
The model can reason and solve complex problems such as grade-school mathematics and basic coding. It exhibits traits of larger language models, both positive and negative, including the ability to think step by step and the potential for biased and toxic generations.
Why does this matter?
Despite having only 1 billion parameters compared to models with over 100B inputs, phi-1.5 has shown promising abilities and performs well on natural language tasks.
It eliminates the need for web scraping or relying on data sources with copyright issues. This research could democratize AI training by making it accessible to smaller organizations and reducing the reliance on specialized computing power.
Knowledge Nugget: Can Large Language Models Reason?
This article by
explains the debate surrounding LLMs centers on whether they possess humanlike reasoning abilities or if their skills are merely a mirage. While LLMs have demonstrated emergent behaviors that resemble reasoning, it is unclear if these behaviors stem from true abstract reasoning or from memorizing and pattern-matching training data.The implications of this debate are significant, as the presence of robust reasoning abilities in LLMs would support their role in developing trustworthy general intelligence. Conversely, if LLMs rely on memorization and pattern-matching, they will lack generalizability and cannot be trusted for tasks outside their training data.
Why does this matter?
This article hinges on whether they possess genuine human-like reasoning abilities or if their performance is primarily based on memorization and pattern-matching from training data. The outcome of this debate has substantial implications for the trustworthiness and generalizability of LLMs.
What Else Is Happening❗
✅ Amazon has launched gen AI capabilities to help sellers write more engaging and effective product descriptions. (Link)
✅ Hugging Face has launched Würstchen, a diffusion model for image generation. (Link)
✅ Researchers have developed an AI tool called RETFound that can diagnose and predict the risk of various health conditions. (Link)
✅ Parle Products, the biscuit brand in India, has partnered with IBM to drive its digital transformation using cloud and AI. (Link)
✅ EY has invested $1.4 billion in AI and is launching a platform that offers AI-powered versions of its consulting products. (Link)
🛠️ Trending Tools
Convex: Full-stack TypeScript platform with real-time database. Integrate OpenAI into workflows.
ChartGen AI: Create charts and graphs with one sentence. Bring your data to life.
Snapy.ai: Save time in content creation. Auto trim & edit videos and audio for engaging content.
Docalysis AI: Upload multiple files on Docalysis and chat with an AI about them all at once! A great way to conduct research efficiently.
Tax Terrapin: AI chatbot for tax queries. Navigates GOV.UK and provides up-to-date answers with official links.
BudgetSEO: Generate high-quality SEO articles using AI. Ideal for entrepreneurs.
HelpKit AI: Turns your Notion help center into a 24/7 AI chatbot. Provides precise answers and saves hours on support
Catbird: Intuitive, full-stack AI tool for branding. Create creative assets easily.
Want to discover more impressive AI tools?
Refer your pals to subscribe and enjoy our daily newsletter to get exclusive access to 400+ remarkable 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, email, or share it on social media with friends.
📈 Thursday Trajectory
In the spotlight today: Enfabrica
Enfabrica, a company that develops networking chips for AI and ML workloads, has raised $125 million in a Series B funding round.
The funding round was led by Atreides Management and included participation from Sutter Hill Ventures, Nvidia, and other investors. This brings Enfabrica's total funding to $148 million.
The funds will be used to support the company's research and development efforts, expand its engineering and sales teams, and enhance its operations.
That's all for now!
Join the prestigious readership of The AI Edge, alongside professionals from Moody’s, Vonage, Voya, WEHI, Cox, INSEAD, and other top organizations.
Thanks for reading, and see you tomorrow. 😊