The Race for Generative AI and LLM Supremacy - Insights from CES AI House

Blog here: The Race for Generative AI and LLM Supremacy — Insights From CES AI House 2025 | GMI Cloud blog

Summary of takeaways at CES 2025 AI House’s panel discussing the Race for Generative AI and LLM Supremacy:

Start Small: Focus on achievable goals with clear ROI before scaling.
Data Quality: Build on a solid data foundation to maximize AI’s potential.
Execution Matters: Implementation separates leaders from laggards—embrace the challenge of scaling AI to production.

  • There is still massive room to grow (77% of companies exploring LLMs for production use), but the decisions around how to use AI need to improve - R “Ray” Wang
  • New skillsets will emerge after AI replaces some current skillsets - Stephanie Buckner
  • AI projects need to be business projects, with expected ROI and pragmatic approaches. Yujing Qian recommends starting small and focusing AI integration by tackling data issues first
  • AI Agents will rise, but the bottleneck is having good data. Hassan Sawaf recommends microagents for AI solutions to ensure compliance, enforce security, and mitigate bias.
  • Stephanie notes that data quality trumps AI modeling, and data must be integrated into the foundation
  • Automate everything, then decide what parts should not be automated and need a human touch. Understanding where liability exists is going to be crucial going forward.
  • “LLMs is a cool battle, but datasets are equally important.” Companies need to be better at retaining ownership of their data.

LLMs and datasets - the value wars have begun.