Nvidia CEO: We bet the farm on AI and no one knew it | TechCrunch
nvidiajensen-huanggpu-computingai-hardwaresiggraph
Abstraction: Jensen Huang recounts Nvidia's 2018 AI pivot from rasterization to GPU-accelerated ML
Key points:
- In 2018 Nvidia made an existential "bet the company" decision to embrace AI-powered image processing: RTX ray tracing and DLSS intelligent upscaling
- The same GPU architecture built for RTX/DLSS became ideal for training large ML models, positioning Nvidia to dominate AI compute
- H100 chips designed from the start for large-scale ML training; GH200 Grace Hopper units can be assembled into a single logical exaflop GPU
- Huang at SIGGRAPH 2023: "Human is the new programming language" — LLMs will front nearly every application
- Predicts fully software-defined, robotic factories building robotic cars
- Remarks made at SIGGRAPH keynote in Los Angeles, August 2023
Connections: Nvidia · Jensen Huang · GPU Computing · Large Language Models · AI Hardware
Source: https://techcrunch.com/2023/08/08/nvidia-ceo-we-bet-the-farm-on-ai-and-no-one-knew-it/