Fireside Chat, Interview, Other
Big Ideas 2024: New Applications for Computer Vision and Video Intelligence with Kimberly Tan
- Computer vision and video intelligence applications are projected to expand into real-world scenarios starting in 2024, driven by a confluence of factors reaching a tipping point where technology is sufficiently capable to enable large-scale venture outcomes.
- Hardware and software business models are expected to proliferate, with companies selling combined camera and software solutions tailored to specific customer needs, a strategy recently validated in residential and school safety and now anticipated to transfer to transportation, industrials, agriculture, and mining.
- Technological efficiency will improve through dropping cloud costs and increased edge computing capacity via cheaper chips and higher-quality cameras per dollar, allowing complex models to run on devices to reduce bandwidth constraints rather than relying solely on cloud processing.
- Innovation in model architecture will focus on applying transformers for image classification and object detection, while generative AI will enable natural language querying of video data without requiring code.
- In transportation, computer vision aims to verify charging and compliance by identifying items given to specific people at specific times, eliminating manual invoice processing; in industrials, it will ensure OSHA workplace safety compliance without monitoring individuals; and in agriculture, it will provide insights on weather and crop health without manual oversight.
- Labor shortages across industries are anticipated to accelerate the adoption of software solutions in sectors previously less inclined to purchase them, while new founders will increasingly emulate the successful hardware and software models established by companies like Flock Safety.
- Significant attention will be directed toward navigating regulatory, privacy, and data rights issues, specifically concerning data ownership, retention duration, and the scope of tracking (e.g., license plates or equipment vs. individuals) to ensure stakeholder comfort and rights protection.
- Broader inquiry into AI applications is expected to extend beyond generative creative uses, potentially encompassing new areas such as AI-driven maritime exploration, endless AI-versed games, and a potential resurgence of voice-based interfaces over traditional apps.