What can cricket teach us about enterprise AI?

on YouTube ↗15m

" summary

Claims that AI progresses from LLMs to reasoning to agents to swarms and then synthetic data, with Gemini enabling up to 10 million context windows; recommends replacing expensive fine-tuning with RAG vector database augmentation now (frustrating but effective). Agents connect to real tools (Stripe, about thousand MCP servers), trade time for accuracy via reasoning, and compose into goal-based swarms that self-check answers to cut hallucinations. Company-wide rollout and short employee training produce usable agentic workflows.

AI-written summary · about

" chapters (6)

1. 0:00Introduction
2. 1:05AI models
3. 3:45Improving AI accuracy
4. 6:30Agentic workflows
5. 9:10Swarms
6. 11:30Future AI

" related talks

• 28mAgentic Enterprise - What your CEO must know about AIHubert Misztela · AI Engineer World's Fair 2025
• 24mScalable GenAI in ProductionKamelia Aryafar · Optimized AI Conference 2026
• 23mScaling GenAI Applications and Agentic AI for Business Competitive DifferentiationAli Arsanjani · Global Artificial Intelligence Conference 2024
• 47mHuman Robot Agent: Scouting AI Beyond Hype and DoomJurgen Appelo · Agile en Seine 2025
• 27mAgentic Systems: Leadership, Risk, and AccountabilityPaige Bailey, Kamelia Aryafar, Joshua S, Khalifeh Al-Jadda · Optimized AI Conference 2026
• 56mEnabling teams with AIMatt Boncek, Adam Presson, Facundo Ruiz, David Kossnick · Config 2025
• 26mAI Enabling InnovationThomas Saueressig, John Thornhill, Ewa Duerr · DLD Munich 24
• 22mAgentic AI for Engineering Leaders | Strategy & Execution GuidePrakash Tripathi · AI Agents Conference 2026