cd/entity/TPUv3· home› entities› TPUv3
grep -l @tpuv3 /news/*.json | wc -l → 1

TPUv3

mentions 1 type Organization feed RSS

// recent coverage 1 mentions

19:31
2026-10-02
dynomight.net
large-language-models

Why didn't we get GPT-2 in 2005?

Training GPT-2 required roughly 10²¹ FLOPs, a level of compute that the 2004–2008 IBM BlueGene/L supercomputer at Lawrence Livermore National Laboratory could have delivered in about 41 days, accordin…

// co-occurs with top 7 entities
GPT-2 1 BlueGene/L 1 IBM 1 Lawrence Livermore National Laboratory 1 Sevilla et al. 1 Chinchilla 1 Gopher 1
// topics top 4 topics
large language models 1 artificial intelligence 1 ai research 1 ai infrastructure 1