Large Language Models Should Not Be a Maslow’s Hammer
When all you have are large language models (LLMs), everything looks like a prompt. Enterprises must avoid falling into the trap of the law of the instrument, a.k.a. Maslow’s Hammer.
When all you have are large language models (LLMs), everything looks like a prompt. Enterprises must avoid falling into the trap of the law of the instrument, a.k.a. Maslow’s Hammer.
This paper outlines the key technological developments driving changes in retrieval augmented generation (RAG). Mapping those changes on a non-linear trajectory, we predict the near future, and provide actionable recommendations for organisations to stay ahead of the curve in the rapidly evolving landscape of enterprise artificial intelligence (AI).
Legacy systems are typically challenging for CIOs – they often support critical business processes; they are expensive to replace and increasingly don’t meet business needs. As a result, they have been historically placed into the too hard basket. CIOs with significant legacy issues need to develop a deliberate strategy to address these challenges, rather than the squeaky wheel approach.