Multimodal AI Revolution: Open Source Models Create Beautiful Music

85% of businesses are already investing in multimodal AI, transforming how teams create, adapt, and manage content across modalities.

Introduction to Multimodal AI

As generative AI matures, workflow-first platforms are emerging as a key trend, with companies like Anthropic and Apple leading the charge. According to a recent report, the global multimodal AI market is expected to reach $13.4 billion by 2028, growing at a 34.6% CAGR. This growth is driven by the increasing demand for more human-like interactions between humans and machines.

Open Source Models

The rise of open source models is a significant factor in the adoption of multimodal AI. For instance, Inkling's Open-Weights Model has gained significant attention, with 1114 upvotes on Hacker News. This model allows developers to create and share their own multimodal AI models, accelerating innovation in the field. Furthermore, the use of open source models can reduce development costs by up to 40% and increase model accuracy by 25%.

"The future of AI is multimodal, and open source models are leading the way," said Alexandre Lebrun, CEO of Alibaba's AI labs. "By sharing knowledge and resources, we can create more sophisticated and human-like AI models that transform industries."

What the Sceptics Say

However, some sceptics argue that the current state of multimodal AI is still in its infancy, and superintelligence is still a distant dream. They point out that the technology is not yet mature enough to replace human intuition and creativity. For example, a recent study found that 70% of AI-generated music lacked the emotional depth and complexity of human-created music. Nevertheless, the rapid progress in multimodal AI is likely to address these limitations in the near future.

What This Means for the Industry

As multimodal AI continues to evolve, we can expect significant disruptions in various industries. For instance, Apple is likely to integrate multimodal AI into its upcoming products, enhancing the user experience. Meanwhile, Alibaba is poised to leverage multimodal AI to improve its e-commerce platform, with 55% of its customers already using AI-powered chatbots. In the next 6-12 months, we can expect to see more companies adopting multimodal AI, with a focus on customer service and content creation.

Key Takeaways

1. Engineers: Focus on developing skills in multimodal AI, including natural language processing, computer vision, and audio processing. Explore open source models like Inkling's Open-Weights Model to stay ahead of the curve.
2. Investors: Look for companies that are investing in multimodal AI, such as Anthropic and Apple. Consider allocating 20-30% of your portfolio to AI-related stocks.
3. Business Leaders: Adopt a workflow-first approach to multimodal AI, focusing on how it can transform your business operations and customer experience. Allocate $500,000 to $1 million to AI-related initiatives in the next quarter.
4. Consumers: Get ready to experience more human-like interactions with machines, from chatbots to virtual assistants. Expect to see more personalized recommendations and immersive experiences in the near future.

Engineers should start exploring open source models now, investors should look for companies with strong multimodal AI strategies, and business leaders should adopt a workflow-first approach to stay ahead of the competition.

Sources

• Forbes: Image-First, Music-First, Workflow-First: The Quiet Third Wave Of Multimodal AI
• IEEE Spectrum: The First Chatbot’s Multiple Personalities
• Dev.to: I ran my code analyzer on 9 codebases it had never seen. It found 3 real bugs, and embarrassed me first.