A heated debate has been prevailing on LinkedIn since Nithin Kamath, Founder & CEO at Zerodha, posted, “I keep getting asked how we use AI/ML/Blockchain at @zerodhaonline, and I keep saying we don’t and haven’t found any use-case yet.”
The trending post gained over 10,000 reactions and 400 comments in a day. Analysts, experts, data scientists and technologists have started taking sides on the post, making it an extremely interesting debate on the motion that AI, ML, and Blockchain are really hyped.
Along with the post, Kamath also posted an image that mentioned the views of Zerodha’s CTO, Kailash Nadh, mentioning that, “I am strongly inclined to say that most of the claims of being “powered by Al” is superfluous marketing, pure hogwash. Bulk of the AI/ML usage, if any at all, are commodity tools that can be plugged in trivially, that do not even warrant mention, let alone strong “powered by AI” marketing. Anecdotally, pretty much every single instance of “Al-first mindset” that I have seen in the industry have been strong cases of misguided assumptions, outright delusions, and often, intellectual dishonesty.”
There were many who loved the duo’s honesty, while there were also several who completely trashed their idea of underestimating the technology and the potential and scope it holds.
Here are some of the highlights from the debate.
The ones who ‘Agree’ with Kamath
Calling the statement by Zerodha CEO – common sense, Ankur Dhamija, Data Scientist at Accuracy, stated – “A very honest point of view. While AI/ML is indeed a very useful technology, but if the use case doesn’t support it, then no point using it. In the end, it is a tool; if the use case doesn’t support it, then why waste resources building that unnecessary tool in the first place.” He also wants organisations to understand this and stop putting deep learning as a requirement in each job description. Adding in some data, expert Pratik Y mentioned a European survey in 2019 that evaluated various startups claiming to offer AI/ML products, and the analysis found that 40% of these startups never used anything close to AI/ML. He wrote, “It’s easy to throw this word around but quite difficult to understand what is AI/ML and where it can be used in a successful way.”
Adding in more stats, Anand Ruparelia, Data Engineer at Intel Corporation, compared two reports – one by Forbes, mentioning that 83% of businesses believe AI is a strategic priority for business. The other by Gartner mentioned that 85% of AI projects fail. He explained, saying, “This is only because of tech-nerds who show an attitude to implement AI just for the sake of implementing it!” Providing insight into how companies are using the term as an alternative to other solutions, Sidharth Goel, Founder At Datafact.io, wrote, “Those two are terms [AI and ML] are currently the most fashionable ones to use… Companies that said they were using AI-powered systems to decide outcomes. While all they were really doing were time series analysis.”
Agreeing to Kamath, Rahul Ranjan Srivastava, Consumer Tech & Analytics Lead at Unilever, expressed disappointment at the misuse of the terms AI/ML. He wrote, “If I hear one more company calling themselves ‘Powered by AI’ for putting some simple automation, I would blow up!! Such generalisations and misuse of terms like AI/ML actually discredit some real complex use-cases which these technologies are solving.”
The ones who ‘Disagree’ with Kamath
Most comments on Nithin’s post showed a tone of agreement, expressing favour to the notion that AI, ML, and Blockchain are hyped. But, many bashed him for underestimating not just the potential of the technology but also the way it can help scale businesses.
Satheesh Kumar V, Director, Artificial Intelligence, Data Science/Machine Learning, FinLabs at Synechron, compared Kamath’s logic to saying that everything in science is plain Mathematics and shouldn’t be named as Physics or Chemistry. He expressed, “Sorry to see the comments that blatantly disown ML and abuse it as if it’s a totally useless hype. All NLP and OCR based use cases used by Banks and FinServices firms use ML extensively – including the customer handling chatbots, analytics reports, backend anomaly and outlier detection modules…just to name a few.” Comparing Kamath’s views for AI to what people thought about cloud when it was new, Sugandha, Principal Product Manager, Eightfold, stated that there was a time when many technologists dissed ‘Cloud’ similarly and now are struggling to play catch up. She said, “Not denying that there are critical sensitivities when it comes to using AI/ML in investing/finance, and your decision to not force-fit AI/ML into Zerodha is truly respectable (and practical and wise), but I wouldn’t be so quick to dismiss the entire domain and everyone in it. Just because it is in nascent stages doesn’t make it hogwash.”
There may not be any use-case for Zerodha; however, there are several businesses that are transforming the user experience with the help of data science (AI/ML) like – Fintech, entertainment, Higher-ed, etc., echoed Ankit Awasthi, a Business Analyst at Optum. He added, “I disagree with this statement. I would suggest Kailash Nadh that being the market leader in broking business in India, your team can research and come up with specific use-cases that will strengthen your market position and will also improve the UX.”
Calling the Zerodha duo a dinosaur, Kaivalya Yadav, Data Scientist and Developer at Quadrical.ai, expressed, “Oh, you are a dinosaur, going to get extinct very soon, don’t worry.”
Hyped or not – will be clear in a few years – based on how the technology grows and transforms businesses. Until then, the debate shall go on on multiple forums like LinkedIn, Twitter, Quora or Reddit.
In the debate, there also were some who advised Kamath about the use cases of AI/ML at Zerodha. Pradeep Kumar, Lead Data Scientist at Fraud, Ops & Tech, said, “If Zerodha starts doing advisory/recos, ML can help personalise recos for clients based on their preferences, activity instead of issuing blanket recos.” Debayan Mitra, Senior Data Scientist at MakeMyTrip, also suggested a few features that could be useful to consumers on smallcase that can be driven by ML. He suggested, “Customised smallcases for customer types based on their risk profiles, investment frequency, investment amount, investment horizon & investment goals, and creating customised stock portfolio allocation based on customer’s tolerance for volatility & risk (smallcase themes – beta users) & above factors.”