It seems that today's state-of-the-art in AI/Deep Learning/Machine Learning/etc has some useful practical applications for multiple industries (image/face recognition, language translation, youtube/music recommendations, stock trading, etc)
However, for AGI (Artificial General Intelligence) research, I think we are nowhere near something that approaches the final goal (mimicking the general intelligence level of a human across multiple tasks). Current approaches (such as RL or Deep RL) are very brittle, not well suited to multiple tasks, too slow to learn, and do not scale to real-world-complexity problems, among the various issues they have.
As a result, it seems to me that all the recent AI research hype will die down, people who are using AI for practical applications will continue to do so, but the funding and hype about AGI will die down, leading to the next AIwinter.
Not sure how long it will take but I think it's coming in the next few years, depending on how deep the pockets are of the current front-runners.
Do people think that there is too much hype around AI these days?
I know, chatGPT is here, and AGI is almost here - they say.
But are there significant new revenue streams from AI? Or other breakthrough “added value”?
Or, as my hypothesis goes:
Yes, there is real added value from recent AI developments, but it doesn’t outweigh all the hype.
Instead, too many companies across industries (finance, pharma, government) “waste” their attention on AI, instead of focusing on higher-value areas like getting good at data management, making quality software, and improving their management practices.
Therefore, an AIwinter would actually do good, and let the world focus on higher value areas.
So what's your opinion?