What was GPT-1 and why did most people miss its significance?

The answer

GPT-1 (Generative Pre-trained Transformer), published by OpenAI in June 2018, demonstrated that unsupervised pre-training on a large text corpus followed by supervised fine-tuning could achieve strong results across diverse NLP tasks. With 117 million parameters trained on the BooksCorpus dataset of over 7,000 unpublished books, it improved state-of-the-art results on 9 of 12 benchmarks. While overshadowed by BERT later that year, GPT-1 established the 'pre-train then fine-tune' paradigm that became the foundation of all subsequent GPT models.