SO MANY THINGS need to go right just so you can watch a TikTok! | E2215
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
In this episode of This Week in Startups, Jason and Alex explore the practical applications of AI across different sectors. The show features three key segments: first, a discussion with Gou Rao, CEO of Neubird, about their AI-powered IT operations engineer called Hawkeye that uses context engineering to troubleshoot complex cloud infrastructure issues (03:19). Second, Jason interviews Mike Vilardo from Subject AI, an educational technology company that combines premium video content with AI-powered personalized learning tools to help teachers focus on one-on-one instruction while automating administrative tasks (28:38). The episode concludes with a fascinating flashback to a 2019 interview with Scale AI's Alexander Wang, where he discusses the future of AI and autonomous vehicles, providing prescient insights about the technology's trajectory years before ChatGPT emerged (47:48). • Core discussion centers on applied AI solutions in enterprise IT operations, educational technology, and the evolution of AI from machine learning to generative AI applications.
Speakers
Gou Rao
CEO and co-founder of Neubird AI, an engineer with extensive experience in IT operations and troubleshooting. He leads the development of Hawkeye, an AI-powered IT operations engineer that uses context engineering to reduce incident response times by up to 90%.
Mike Vilardo
CEO of Subject AI, formerly known as A Meal Learning, a company focused on revolutionizing education through AI-powered personalized learning. He's an ex-Uber employee turned education entrepreneur who successfully rebranded his company and secured the premium domain subject.ai.
Alexander Wang
Former CEO of Scale AI who was interviewed in 2019 at age 22 after raising over $100 million in funding. He later sold approximately half of Scale AI to Meta and became head of their superintelligence team, demonstrating remarkable prescience about AI development trajectories.
Key Takeaways
Context Engineering Beats Raw Processing Power
Gou Rao emphasizes that successful AI implementation isn't about throwing unlimited data at large language models, but about intelligent context engineering. (16:18) Just as you wouldn't visit a doctor unprepared with vague symptoms, effective AI systems require precise, relevant data inputs to avoid hallucinations and reduce inference costs. Neubird's approach focuses on isolating specific problems and providing targeted context rather than overwhelming AI models with irrelevant information. This methodology has enabled Hawkeye to achieve 90% reduction in mean time to incident response while maintaining accuracy and cost efficiency.
AI Amplifies Human Expertise Rather Than Replacing It
Mike Vilardo's Subject AI demonstrates how AI can enhance rather than replace human capabilities in education. (34:30) Teachers using Subject AI can serve three times as many students while spending more time on one-on-one instruction, as the AI handles repetitive tasks like grading and administrative work. This allows educators to focus on what they do best - creating meaningful connections with students and providing personalized guidance. The key insight is that AI should handle mundane tasks to free up humans for higher-value work.
Agentic Systems Require Human-Like Interaction
Gou explains that modern AI systems should be treated more like employees than traditional software, requiring feedback and iterative improvement. (23:36) Unlike deterministic software with binary outcomes, agentic AI systems perform within ranges and can be coached to improve their responses. This shift from deterministic to probabilistic computing requires users to engage with AI systems conversationally, providing feedback and corrections to enhance performance over time.
Direct-to-Consumer vs B2B Strategy Pivots
Mike Vilardo shares valuable insights about pivoting from direct-to-consumer to B2B sales in education technology. (45:11) Initially, Subject AI focused on affluent families through direct sales but struggled with seasonal usage patterns. By pivoting to school districts, they achieved more consistent revenue while also serving underserved communities. The B2B approach required learning entirely new sales processes, stakeholder identification, and extended sales cycles, but ultimately provided more sustainable growth and greater mission impact.
AI Development Plateau Requires External Intelligence
Gou Rao acknowledges that AI model improvements are reaching a plateau in terms of raw capability increases, but emphasizes that future advancement lies in external systems and context. (25:00) Rather than focusing solely on larger models with more parameters, the next phase of AI development involves building better support systems, context protocols, and external knowledge integration. This shift from pure IQ to enhanced EQ through external augmentation represents a more practical path forward for AI applications.
Statistics & Facts
- Neubird's Hawkeye achieves close to 90% reduction in mean time to incident response and resolution, as reported by Gou Rao (19:35). This dramatic improvement demonstrates the practical impact of well-designed AI systems in enterprise operations.
- Subject AI maintains 96% course completion rates among students who start at least one course, as mentioned by Mike Vilardo (37:33). This extraordinarily high engagement rate in education technology showcases the effectiveness of gamification and personalized learning approaches.
- Teachers using Subject AI can serve three times as many students while reducing their platform usage to under one hour daily, while students spend approximately three hours per day on the platform (40:10). This demonstrates the leverage effect of AI in education, allowing human teachers to focus on higher-value interactions.
