What you should know: I am currently working for Orbit Cloud Solutions as Cloud Advisor, but any posts on this blog reflect my own views and opinions only.
I am really excited about the new world of Gen AI tools that has opened up in the last year. Anyone who follows my blog has probably seen the images generated with midjourney and noticed that some of the text has been written with generative AI assistance.
However, I also have to admit that I find the generated content flooding the web annoying. I feel like we are expanding every tiny little thought into huge SEO-optimized documents using generative AI and then needing another LLM to summarize the content into a single paragraph.
After a while, it becomes so easy to recognize content created with naive prompting. My mind then immediately gets the feeling that whoever created this content was lazy and careless. And my subconscious then logically tells me to skip this content because I can assume that not much time was spent proofreading it. Or to make sure that there is even a single interesting thought in it that is worth reading.
And that’s just the tip of the iceberg of the problems with Gen AI.
In his recent article in Harvard Business Review, Oguz A. Acar asks whether generative AI has already peaked. And he is quite skeptical whether the success story of Generative AI will continue as we have seen in the last 12 months. There are still so many structural problems that need to be solved in the long run, like the ones I pointed out in an earlier post.
Do not get me wrong, there are many amazing scenarios for generative AI and even more so for LLMs in general. But the whole conversation around generative AI is overheated and too many projects get started due to FOMO. It is time for this hype to cool down.