Revenue Search: Inside Bittensor Subnet Session with Ken Jon Miyachi from Bitmind: SN 34
19 snips
Jul 12, 2025 Ken Jon Miyachi, founder of BitMind, discusses his innovative work on computer-vision models aimed at detecting deepfakes and verifying AI-generated content. He shares insights on a viral fake video detection case and the significance of computer vision for future technologies like autonomous vehicles. Ken outlines BitMind's dual focus on consumer and enterprise monetization strategies, and emphasizes the challenges of maintaining accuracy in image versus video detection. He envisions partnerships with major platforms like Meta to scale their impact in combating AI-enabled scams.
AI Snips
Chapters
Transcript
Episode notes
Vision-Based Fake Detection
- BitMind classifies images and videos as real, AI-generated, or modified using computer vision.
- The team plans to expand to richer features and attributions beyond simple classification.
Real-World Detection Example
- BitMind detected a widely viewed Israel-Iran video as fake using its subnet models.
- Users can access detection via a browser extension, BitMineBot, or detector.ai by dragging images.
Vision As The Next AI Frontier
- Vision models are the next frontier for AI as they power real-world actions like self-driving cars and robots.
- Recent innovation focused on LLMs but high-fidelity vision is crucial for future embodied AI.