Engineering teams at layer one blockchain start-up Shardeum are using ChatGPT internally to review their codes to make them hack-proof. It is the most harmless use case of artificial intelligence (AI), said Nischal Shetty, Co-founder and CEO of WazirX, and currently building Shardeum.
This is coming at a time when Web 3 and blockchain developers are experimenting with ways to deploy AI and generative AI tools like ChatGPT for various use cases like analysing blockchain data, and creating and optimising smart contracts to name a few.
While the more consumer-facing use cases and products integrating Shardeum’s network and AI are still in the works, Shetty said, “Every chain and product will soon use AI. Within Shardeum, we are using ChatGPT to just review code. It's the most harmless, but the most useful thing you could do with AI right now. You get like a second pair of eyes looking at your code just to identify any issues or loopholes, which as a developer you missed.”
“In fact, our team has an internal ChatGPT plugin that has been built by the team, just for internal use, where every time they push a code, the ChatGPT goes through that code and tells us if there are any issues with the code,” he told Moneycontrol.
According to Shetty, all smart contracts that are going to be built on any chain and not just Shardeum will be using AI. In the past couple of years, Web 3 had some concerns around hacks and thefts, where AI could play a big role in fixing initial issues at the developer’s level.
The next phase of growth for blockchain will be combining AI to build products that consumers will experience, said Shetty. But these experiments would require scalable chains to build on and make the experimenting stages cost-effective.
He said, “I believe the biggest problem for building these products is the need for a scalable chain. Otherwise, it'll be very, very expensive to try these experiments. I'm really looking forward to Shardeum going to the mainnet, because I think that could be, in fact, one of the stronger use cases for a chain like Shardeum, where it can scale and it is cheap. So, you could try these experiments with AI.”
Shardeum is currently beta testing its network, and plans to file a few patents before going to the mainnet ― which is releasing a complete fully operational blockchain network. This is expected to happen in the second half of the year. At that stage, Shardeum will also be able to launch its own native tokens named the Shard.
Using ChatGPT to review internal codes doesn’t bother Shetty regarding possible security breaches, as he believes being a Web 3 and a blockchain-based company whose codes are open source, helps.
“The good thing about Web 3 is that it's mostly open source. For Web 2 companies though, this will be a big problem, because all of the code is proprietary. You don't want it out. For us, we have open sourced part of our code already. We are opening the rest of the code too, as soon as we file some patents, but the objective is to be open. So, there's less fear about IP issues in Web 3, and even for Shardeum we're not too worried,” Shetty said.
In October last year, Shardeum announced that it had raised $18.2 million at a valuation of $200 million in a seed round led by Jane Street, Struck Crypto, The Spartan Group, Big Brain Holdings, DFG, Ghaf Capital Partners, and Foresight Ventures.
Shardeum also has several strategic partners backing it, including CoinGecko Ventures, and prominent angel investors, such as former Coinbase CTO Balaji Srinivasan; Mayur Gupta, CMO at Kraken; Michael Montero, Co-founder, Resy and CrowdTwist; Pankaj Gupta, VP Engineering at Coinbase; Harsh Rajat, Co-founder, Push Protocol; Nakul Gupta, Lead PM at Coinbase NFT Marketplace & Institutional Onboarding, and Ajeet Khurana, Founder at Reflexical, among others.