2023-01-24_the-ai-brick-wall-a-practical-limit

Canonical Target Counts

TargetOccurrences
04_knowledge_base/Mixture of Experts24
02_companies/GOOG17
04_knowledge_base/Sparsity15
04_knowledge_base/GPU12
02_companies/NVDA12
04_knowledge_base/GPT-36
04_knowledge_base/Pruning5
04_knowledge_base/A1005
02_companies/MSFT4
04_knowledge_base/Transformer architecture4
02_companies/OpenAI3
02_companies/ORCL2
04_knowledge_base/Triton2
04_knowledge_base/Microsoft Azure2
04_knowledge_base/PyTorch2
04_knowledge_base/CUDA2
04_knowledge_base/HGX2
04_knowledge_base/NVLink2
02_companies/DeepMind2
04_knowledge_base/Chinchilla Scaling2
04_knowledge_base/Early Exit2
04_knowledge_base/DGX2
04_knowledge_base/SXM2
04_knowledge_base/Dynamic Routing2
02_companies/CRWV1
04_knowledge_base/TPU1
04_knowledge_base/H1001
04_knowledge_base/Quantization1
02_companies/Cerebras1
02_companies/SambaNova1
04_knowledge_base/1.6T1
02_companies/Anthropic1
04_knowledge_base/PCIe1
02_companies/Switch1
02_companies/Baidu1
04_knowledge_base/Sparse Attention1
02_companies/BE1
04_knowledge_base/AWS1
02_companies/COHR1
04_knowledge_base/MI3001
02_companies/META1
02_companies/Amazon1
04_knowledge_base/TPUv41
04_knowledge_base/Software Stack1
04_knowledge_base/Reinforcement learning1
04_knowledge_base/TPUv51
04_knowledge_base/BERT1
04_knowledge_base/Model FLOPS Utilization1
04_knowledge_base/Token Dropping1
04_knowledge_base/Checkpointing1