Rosemarin suggests that businesses need to be selective in choosing which AI projects to fund, given the high cost and shortage of GPUs. He also emphasizes the need for efficient, scalable data platforms to support AI workloads. He warns of potential power shortages and the intense power requirements of AI, suggesting that businesses need to consider this upfront. Finally, he stresses the importance of security and clear data provenance in the face of potential legislation and legal challenges.
Key takeaways:
- Businesses must ruthlessly prioritize their AI projects due to the high cost and limited resources, making difficult decisions about which projects to fund and operate.
- Efficient, scalable data platforms are crucial to support AI workloads, especially given the high cost and shortage of GPUs. Storage plays a crucial role and should be thoroughly assessed.
- Enterprises need to plan for power and electricity limitations, considering the intense power requirements associated with building and training data sets.
- Security and data provenance should be intrinsic and clear from day one, with businesses prepared to pivot in the face of potential legislation and legal challenges around the use of data.