At least one of the most powerful neural nets at end of 2026 will be trained using 10^26 FLOPs

Will Jose Luis Ricon run more than 864 quintillion floating point ops on a NVIDIA h100/b200 (or better) by EOY2026

Will an AI model use more than 1e28 FLOPS in training before 2026?

Will any supercomputer reach 1 zettaFLOP before 2035?

At least one of the most powerful neural nets at end of 2030 will be trained using 10^27 FLOPs

At least one of the most powerful neural nets at end of 2030 will be trained using 10^26 FLOPs

In 2020, Joe Carlsmith estimated that 10^15 FLOPS is "enough" to perform human functionality. The next post estimating this number will believe it is >=10^15

Will there be an announcement of a model with a training compute of over 1e30 FLOPs by the end of 2025?

Will a machine learning training run exceed 10^26 FLOP in China before 2026?

Will a machine learning training run exceed 10^27 FLOP in China before 2030?