LINK2AI Blog

Testing and optimising an LLM-based application feels more like medical diagnosis and treatment than traditional debugging. Problems are often difficult to find, and their causes usually cannot be determined with absolute certainty.

When AI now acts, the allowed error margin suddenly becomes much smaller. Inhererent shortcomings of LLMs and their vulnerability against manipulations and attacks, e.g., prompt injections, are now a real problem.

The roles seem to be inversed: weren't the machines supposed to do the boring, mechanical stuff? Instead, we find ourselves copying and pasting text that LLMs have written from the browser to our database. How did this happen?

Reports of delays and problems are piling up in the AI world. The big names in the industry - Amazon, Meta, OpenAI, Apple - are postponing their ambitious projects. Is this the beginning of the end of the AI boom?

Are we really taking off with AI, or are we just playing around? Many organizations hesitate – let's have a closer look at the reasons.