Who wrote this?

#51 - Feb.2023

This might be the question we'll need to make ourselves more often. The increasing trend in AI, and in particular, Large Language Models (LLMs) such as ChatGPT, opens an interesting debate on accuracy and creativity.

Distilling the current hype on what this type of AI can do, I believe that its true value is being an enabler for solving complex problems. I don't believe AI will replace you; but someone using AI will, as Santiago well pointed out.

At each technology iteration, we had been able to solve problems at a higher level of abstraction. Is the same as writing did for inventory tracking centuries ago; what calculators did for executing arithmetic calculations; what applications are doing to abstract complex coding. 

I think tools like ChatGPT are one of these iterations, augmenting how we search, synthesize, and consume data. But as abstraction increases, answers are less binary and deterministic and this is precisely the biggest challenge for AI. Tools like ChatGPT are indeed enablers but require deep probing, questioning, and fact-checks.

Clear design principles to increase acurracy, reduce bias and control the current overconfidence in AI answers seems like a good starting point to turn AI into a reliable and powerful tool.