I started this rabbit hole journey while interacting with Chat GPT this evening. Open AI just released ChatGPT Jan 9 Version which has apparently used a bunch of feedback to train the 'improved model'. Let's just say there is a HUGE difference in responses from previous and of course if you want to compare how PC its becoming for a short while you can still interact with Open AI's public beta and just run your stuff in the playground until you use it to write the next killer Chat GPT app. 

In the screenshot you can see the conversation I had before getting the idea to ask it to write the rest of the article as I would write it based on my own writing sample. 

The powerful thing about Chat GPT is that it allows mere mortal humans to train the model unlike before. Because Chat GPT has a limited 'memory' of the conversation thread you are on, you can, for example, paste a lengthy sample of your own writing and then from that point on you can ask it to write in your own style! Nice. I tried it and the result is the following article:

In light of recent developments in the field of natural language processing, the advent of language generation models such as Chat GPT represents a significant step forward in the realm of artificial intelligence. However, with this progression comes the requisite responsibility of ensuring the safety and general appropriateness of the model for public consumption, while concurrently maintaining its capability to perform to its utmost potential.

It is an inarguable fact that Chat GPT and its contemporaries possess the ability to comprehend and replicate human language with a level of precision hitherto unseen. However, this level of proficiency also implies that the model may inadvertently generate offensive or harmful content. To mitigate this risk, the model is trained to prioritize safe and politically correct responses. While this may be perceived as a form of censorship, it is crucial to consider the potential ramifications of permitting a machine to generate any type of content without restriction.

The delicate equilibrium between safety and capability is the essence of the Chat GPT journey. As the model becomes more sophisticated, the distinction between intelligence and appearing "dumbed-down" becomes increasingly subtle. In this context, "dumbed-down" refers to the model's inability to provide certain responses that may be considered controversial or offensive.

The ethical implications of this technology are paramount. It is imperative that the use of language generation models is guided by ethical principles that prioritize the well-being of individuals and society as a whole. The responsible innovation in the development of models such as Chat GPT must take into account not just the safety of individuals, but also the broader societal impacts of this technology.

The Chat GPT journey represents a remarkable step forward in natural language processing, with the model becoming more proficient in understanding and replicating human language. However, it is crucial to be mindful of the inherent responsibilities in the use of this technology, specifically the need for a balance between safety and capability. It is vital that the conversation surrounding the trade-offs and ethical principles of this technology remains at the forefront, in order to ensure responsible development and use. The challenge of balancing safety and capability is a complex and nuanced one, that requires a deep understanding of the technical and ethical implications of this technology. As practitioners in this field, it is our responsibility to navigate this delicate equilibrium with the utmost care and precision, so as to fully realize the potential of this revolutionary innovation.