Onward, a Fundrise Production 48: Will AI Surpass Us? with Robert Nishihara, co-founder of AnyScale
5 snips
May 27, 2025 Robert Nishihara, co-founder of AnyScale and co-creator of Ray, shares his expertise on AI's evolution and capabilities. He highlights the transition from academic theory to practical engineering, addressing the challenges of scaling AI. The conversation dives into the importance of high-quality data, the delicate balance between innovation and resources, and the evolving landscape of AI jobs. Nishihara also discusses steerable AI, exploring its potential in enhancing creativity and productivity while delineating the future of self-directed learning and critical thinking in an AI world.
AI Snips
Chapters
Transcript
Episode notes
Ray's Origin Story
- Robert Nishihara started as a PhD student researching machine learning algorithms. - His focus shifted to solving software engineering bottlenecks to scale AI models with distributed computing via Ray.
AI: Data Over Models
- AI research focus shifted from model architecture to optimizing the training data. - Data quality and selection have become central to AI performance improvements.
Synthetic Data's Potential
- Synthetic data can improve AI by generating curated, high-quality examples. - AI can think harder about data to resolve contradictions and derive new insights.