MIT Report says that 95% of Gen AI initiatives fail. That’s pretty stark, considering all the hype and not to mention the hyped up valuations of AI companies. Is it true? Yes and No…it depends on whom you ask. 95% of Gen AI initiatives fail My take on it is the same as that of Maverick’s in the movie - Top gun “It’s not the plane, it’s the pilot” “It’s not the plane, it’s the pilot” Gen AI is a relatively new tool in a technologist’s toolbox. We have all seen the potential - at least the ones who have tried ChatGPT or any other GPT would agree. Then all of a sudden MIT drops this bomb and almost immediately it seems to crash the AI-tech stock prices. An over-reaction? - I would say - yes! Let me tell you why… The same frenzy has happened with other innovations before. It’s the norm! Here are some noteworthy examples that come to mind when I think of similar speculations of impending doom. 1. With the Internet The same has happened with the internet in 2000 along with the dotcom crash! There were many pundits who were quick to write the internet off. Yes, the bubble popped. But that doesn’t mean that there is no value underneath. 2. With iPhones And…the same has happened with the iPhone launch in 2007. You can see the Endgadget article on the below. And we all remember what the then Microsoft CEO - Steve Balmer thought about it! How are you sure that it is a trend and not a fad? Because of two main reasons……. 1. Quantitative Data: We as a society have already integrated Gen AI in our lives. ChatGPT, the Gen AI poster child, is already generating $10 Billion in ARR. That is not for nothing. It is because the technology is delivering value in many multiples of that. Add to it the other top models, image gen, video gen capabilities etc and you are clearly looking at multiple billions of value added to the society. 2. My Own Experience: I have seen it being successfully implemented already in multiple avenues I am in Health-tech and AI. In my previous company where I was working in 2022-23, it took us over a  year to get a linear regression algorithm to be production ready. That too with limited success. Now in my new company, we were able to deliver better accuracy with explainability in less than three months - thanks to Gen AI. And by the way…the costs were low too, since we used open source models and fine tuned them. In short - double the accuracy with explainability with a quarter of training time and cost. What are the reasons for failure then?! Here are some of the reasons I think it fails Spagetti on the wall approach There is no clear user or customer research to identify the best use cases and scenarios. Very often it’s just a death march to get the ‘Gen AI’ delivered without any clear purpose, hoping that something will happen. Garbage In, Garbage Out If the quality of your data or your fine tuning is messed up, it will reflect negatively in the results. Needless to say the quality of your AI team also matters. Mandate from the Mountaintop Every enterprise leader knows that they should start implementing anything Gen AI. Hence the mandate comes from the top from a CXO. Most of the organization doesn’t really care about it much or changing their workflows, which inevitably ends in failure. In Conclusion Is there an AI bubble? In most likelihood - yes. Will the bubble pop? in most likelihood - yes. Does it mean that Gen AI is a fad? - No!!! There will be rumors and sensationalism, and fear of impending doom, as it is common with many tech cycles. But we should put things in perspective, iterate and move forward. Gen AI has already made us more productive as a society, is already delivering gains to enterprises and organizations who are intentional about it and it’s influence will only be more profound in the future. If you are not on the ship, it’s high time to get on this one!

The is an opinion piece based on the author’s POV and does not necessarily reflect the views of HackerNoon.

Quant

Unsuccessful With Gen AI? It's Not the Plane, It's the Pilot!

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Untitled Story

Why X Followers Matter for Web3 Projects

Spark and PySpark: Redefining Distributed Data Processing

Javascript File Handling Writing Contest by Filestack & HackerNoon

10 Ways AI Has Changed Our Lives

10+ Questions you Should ask Yourself Before Developing an AI Solution

Why X Followers Matter for Web3 Projects

Spark and PySpark: Redefining Distributed Data Processing

Javascript File Handling Writing Contest by Filestack & HackerNoon

10 Ways AI Has Changed Our Lives

10+ Questions you Should ask Yourself Before Developing an AI Solution

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps