1k researchers and executives call for ‘pause’ on advanced AI systems

Re:1k researchers and executives call for ‘pause’ on advance... by nuancebydefault on Wednesday March 29, 2023 @04:09PM (#35362725)

AI chip making? I can train an AI on my intel laptop if I whish. If I need more CPU power, i can rent some. The genie is out of the bottle and the only way is forward. The latest worldwide race.

[ Parent | on HN ]

Re:1k researchers and executives call for ‘pause’ on advance... by avital on Wednesday March 29, 2023 @04:25PM (#35362977)

This isn't accurate. The bottleneck in very-large-scale-training BY FAR is communication between devices. If you have a million CPUs, the communication cost will be significantly higher than a thousand A100s (perhaps in the order of 100x or even more). So this is only possible to replicate with very dense and high compute chips with extremely fast interconnect.

[ Parent | on HN ]

Re:1k researchers and executives call for ‘pause’ on advance... by nuancebydefault on Thursday March 30, 2023 @04:59PM (#35378760)

Thanks for providing this insight. Is A100 the only platform? Can we pause/resume all such platforms simultaneously?

[ Parent | on HN ]