Big data was always a buzzword that was weirdly coopted by database people in a way which makes 0 sense. Of course there are vanishingly small number of use cas

Big data was always a buzzword that was weirdly coopted by database people in a way which makes 0 sense. Of course there are vanishingly small number of use cases where we need fast random access to any possible record to look up. However what ML systems and in particular LLMs rely on having access to millions (if not billions or trillions) of examples. The underlying infra of which is based on some of these tools. Big Data isn't dead, just this weird idea that the tools and usecases around querying databases has been finally recognised as being mostly useless to most people. It is and always has been about training ML models.