The rapid strides being made by LLMs in terms of quality, depth of information and sophistication of reasoning, are fuelled by two major factors. The first is iterative advancements in applied machine learning techniques, which continuously improve the performance of LLMs. The second is growing access to text, data and multimedia used in training these models. AI firms have argued that information on the Internet should be freely usable for such training, even if reproduction and similar acts of syndication by any non-AI entity would be subject to a licence fee and consent by the content producer. This has led to a fierce debate that pits the interests of AI hyperscalers and developers against content producers (news, entertainment and book publishing, to name a few). In this light, the De

See Full Page