Member-only story
When Will AI Have a Breakthrough?
Current capabilities and limitations
Unless you live under a rock, you have heard the gossip about ChatGPT and the future of Artificial Intelligence.
ChatGPT is the most well-known and publicly accessible AI interface we currently have. There are valid arguments that say the advancement of AI will lead to the downfall of us all, or propel us to a glorious new era in civilization. It seems to me that in the foreseeable future we will end up somewhere in the middle of these two extremes.
The rapid advances in Machine Learning Models being deployed to train AI seem to be happening daily. Still, there are limitations that prevent it from being an all powerful entity on its own.
I do not believe that it will have a profound effect on civilization until it reaches the point where it can influence human behavior. Sure, passing the BAR exam is impressive for a robot, but just spitting out factual evidence is not going to influence anybody.
Is it helpful when you can get a quick and relatively accurate answer to anything you ask? Yes. But when is the last time a google search made you rethink your life? To me, the current state of ChatGPT is an interactive and more complex google search.
Even in this domain, there are still some issues that I can see. For example…