The Future of AI & Video Search

Table of Contents

1. Introduction
2. The Need for Advanced Video Search
3. The Limitations of Current Video Indexing
4. Introducing 12 Labs: The Next Generation of Video Infrastructure
5. How 12 Labs' AI System Works
6. Potential Use Cases for Video Search
7. Competitors and the Growth of the Video Content Market
8. Establishing Credibility and Winning the ICCV Video and Language Understanding Competition
9. Targeting the B2B Market and Revenue Generation
10. The Team Behind 12 Labs and Their Journey So Far
11. Future Plans and Funding
12. Conclusion

The Next Generation of Video Search: Introducing 12 Labs

Video search has become an integral part of our daily lives, whether it's searching for work-related files, browsing the web, or seeking specific content within a video. While search engines like Google and YouTube allow us to find videos Based on keywords, tags, and descriptions, advanced video indexing still remains a challenge. The ability to quickly locate specific segments within a video, without having to scrub through the entire footage or rely on transcripts, is a goal that many seek to achieve.

That's where 12 Labs comes in. This San Francisco and South Korea-based startup aims to revolutionize video infrastructure by developing the next generation of video search technology. Their mission is to provide the "Control F" function for searching within videos, allowing users to find precise sections, references, discussions, or visuals they're looking for in a matter of seconds.

The Need for Advanced Video Search

The current world of video indexing is still relatively surface-level. Have You ever found yourself scrolling through a lengthy presentation or lecture, trying to find that one specific segment referring to a particular topic? Or perhaps you're seeking a specific part in a video where an idea was brought up by a team member during a brainstorm meeting. These instances highlight the limitations of current video search capabilities.

While AI systems have made great strides in comprehension when focused on singular senses like audio or still images, comprehending video content in its entirety remains a considerable challenge. Human viewers naturally combine visual and oral elements to build a holistic understanding, including identities, intentions, cause-and-effect relationships, and more complex ideas. This level of comprehension and Context is still beyond the grasp of AI systems.

Introducing 12 Labs: The Next Generation of Video Infrastructure

12 Labs has developed a sophisticated video infrastructure model that addresses the shortcomings of current video search technologies. Their AI-powered system indexes videos, analyzes their content, and links them to specific video vectors containing key information. By training their neural network using billions of parameters, correlating audio and video with situational context, 12 Labs aims to achieve a better understanding of video content.

Rather than examining every frame, 12 Labs' algorithm identifies vital frames where scene or object changes occur. This enables fast and scalable search, allowing users to find specific sections or scenes with minimal effort. The AI forms temporal connections between frames to understand the passage of time and mimic human memory.

How 12 Labs' AI System Works

When a user searches for a specific query, 12 Labs' AI system finds the video vectors that closely match the search criteria and automatically outputs the most Relevant scenes and video file names. This process is facilitated through a user-friendly API that can be seamlessly integrated into existing tech stacks. The API allows users to search through videos wherever they would typically do so, but with significantly improved accuracy and efficiency.

The 12 Labs API can be fine-tuned to perform exceptionally well within domain-specific use cases. Whether searching for diagrams, sketches, spoken jargon, or conventional business conversations, the AI can adapt to deliver optimal results. With the recent launch of their API playground, potential customers can explore and understand how 12 Labs' technology can meet their specific needs before committing to the product.

Potential Use Cases for Video Search

The applications of 12 Labs' video search technology are vast and applicable in nearly every industry. Beyond content summarization and moderation, video search can enhance digital asset management, workplace productivity, recommendation engines, and enable search for end-users. The ability to efficiently search through vast video libraries has implications for accessibility, archival quality, Record-keeping, and data-driven decision-making.

For example, content moderation is a critical use case that can benefit from advanced video search capabilities. Companies like Facebook, which invest billions of dollars in content moderation, can potentially save thousands of hours of manual labor by efficiently searching through their entire video library for specific actions or content that goes against their terms of service.

Competitors and the Growth of the Video Content Market

While 12 Labs faces emerging competitors and potential development from industry giants like Netflix, Amazon, and meta, the demand for video search technology continues to grow. Today, approximately 80% of the world's online traffic exists in video form. Platforms like TikTok and YouTube see massive user engagement, with over 500 hours of video uploaded every minute on YouTube alone. These trends, along with the rise of video intelligence tools, indicate that the market for video understanding and search is significant and scalable.

Establishing Credibility and Winning the ICCV Video and Language Understanding Competition

As a young startup, 12 Labs faces the challenge of establishing credibility in a market where customers may expect experience from established players like Google and Microsoft. To overcome this hurdle, 12 Labs participated in the 2021 ICCV Video and Language Understanding Competition, competing against tech giants like Microsoft, Tencent, and Baidu. Against all odds, 12 Labs emerged as the winner, cementing their credibility within the industry and showcasing the capabilities of their technology.

Targeting the B2B Market and Revenue Generation

To leverage the value of video search, 12 Labs primarily targets the B2B market and enterprise customers. Their product's versatility allows integration into existing tech stacks, making it applicable to a wide range of business operations. Whether providing video infrastructure for enterprise usage or enhancing video editing platforms, 12 Labs offers a solution that can adapt to various video-centric use cases.

12 Labs operates on a subscription-based and consumption-based revenue model. Customers can choose between a monthly subscription plan with a set amount of API usage or a pay-as-you-go model that charges based on monthly usage. The flexibility in pricing allows enterprises to Scale their usage as needed, making it an attractive and adaptable solution.

The Team Behind 12 Labs and Their Journey So Far

The founding team of 12 Labs consists of friends from South Korea with backgrounds in machine learning. Their experience in the Korean cyber ops command, where they worked together for two years, inspired them to tackle the challenges of video infrastructure. While they may be relatively young and lack a track record in machine learning development, their dedication, thoughtfulness, and resolve have driven them to develop cutting-edge technology.

The team has received guidance from experienced mentors, such as Isaac Cato, former MD at Techstars Seattle. Their collaboration has helped them refine their business operations and build credibility in the market. By speaking to hundreds of potential customers and understanding their pain points, 12 Labs has been able to Shape their technology to meet specific customer needs.

Future Plans and Funding

Backed by a $5 million seed round led by index Ventures, 12 Labs plans to Continue enhancing its revolutionary video search technology. Their roadmap includes extending cloud-to-cloud video transfer capabilities, adding new engines for different types of indexing options, and focusing on data-centricity. The ability to incorporate domain-specific knowledge further enhances the value proposition of their product.

As more customers adopt the 12 Labs API, the neural network will continually improve and become more accurate across different use cases. This reinforcing value loop will attract a broader customer base and further validate the scalability and effectiveness of their technology.

Conclusion

12 Labs is at the forefront of revolutionizing video search technology. By addressing the limitations of current video indexing methods and leveraging advanced AI systems, they provide a powerful solution for businesses seeking efficient and accurate video search capabilities. With their growing list of satisfied customers and the potential to unlock new value across various industries, 12 Labs is positioned to shape the future of video understanding and search.

Highlights

• 12 Labs is developing the next generation of video infrastructure, providing advanced video search capabilities.
• Their AI system analyzes videos and links them to specific video vectors, allowing fast and accurate search.
• Potential use cases span industries, including content moderation, digital asset management, workplace productivity, and recommendation engines.
• Competitors like Netflix and meta are exploring similar technologies, but 12 Labs' flexibility and accuracy make them an attractive choice.
• By winning the ICCV Video and Language Understanding Competition, 12 Labs established credibility and showcased their technology.
• Targeting the B2B market, 12 Labs offers subscription-based and consumption-based pricing models for flexible adoption.
• The young founding team of 12 Labs, backed by experienced mentors, aims to shape the future of video search.
• With plans for future development and funding from notable investors, 12 Labs is poised for growth and success in the industry.