Scaling AI for Industry
summary
Companies can run Llama 4 on a single H100 GPU with a 10 million tokens context, enabling combined long‑ and short‑term forecasting; some firms cataloged 200–800 AI use cases and must prioritize a top‑10 roadmap. Industrial stacks need at least three layers (infra, model/foundation, app/ops) and a data inventory to integrate transaction, CAD and sensor data. A central AI hub has three roles—strategy and investment, build the ASC, and embed small product teams—and agents like DeepSeek R1 (≈87% on tasks) and O4 Mini (≈99% on easy tasks) already automate swivel‑chair enterprise workflows.
AI-written summary · about
chapters (5)
- 0:00Introduction by MC
- 3:23Recent AI developments
- 16:51AI Northstar for industries
- 22:19AI agents in enterprise
- 28:41Q&A session
related talks
29mFrom Experiment to Enterprise: Architecting AI for Stability and Scale
16mUsing AI and Computer Technologies to Enhance IT Infrastructure
34mScaling AI Like a Pro-PepsiCo’s LLM Deployment Strategy on K...
25mAI-powered digital transformation
33mEngineering AI Applications: A Conversation with AI Leaders
27mAI at Scale in the Enterprise
13mAI as a Strategic Asset: Building Competitive Advantage
31mRealizing Industrial AI at Scale with