Even more potent than before is OpenAI’s most recent AI model. ChatGPT is a startlingly lifelike chatbot that can generate reams of intelligent prose on practically any topic you can think of. It also remembers what you tell it, which distinguishes it from other language generation models and enables dialogues that convincingly convey the activity of the mind.
The chatbot’s ability to convert human commands into lines of code is impressive: In response to my instructions, ChatGPT created an NFT using the DALL-E image I had created by writing a smart contract in the Ethereum programming language Solidity.
Although ChatGPT is only a free research sample, it has already captured the attention of the IT community and reached a million users just five days after its introduction at the end of last month. Comparatively, it took the AI coding assistance on GitHub around six months to reach the same milestone.
Even though inventive developers may utilise the technology to improve their coding or overcome language obstacles, ChatGPT makes creating harmful code simpler than ever or setting up a honeypot that looks convincing enough to dupe investors.
Some crypto experts are already effectively utilising the technology. The chatbot has proven to be quite helpful in summarising complex code and lengthy scholarly texts, according to Hugh Brooks, head of security at the brilliant contract auditing company CertiK.
And Zellic’s creator, Stephen Tong, claimed his business already uses it for sales and customer service. He said this increases productivity for all team members, enabling cosplaying tech bros to quickly deliver a “super-buttoned-up, professional experience” with ease.
Tomiwa Ademidun, a young software engineer from Canada, is also at the front of the group of crypto-utopians. She utilised ChatGPT to create a bitcoin wallet from scratch and then produced a comprehensive manual that explains how to use it, replete with illustrations.
The man declared, “It’s really rather remarkable.” ChatGPT taught admin complex encryption ideas with the affable charm of a supportive computer science instructor in high school. The program eventually produced what Ademidun characterised as nearly faultless code.
And when Ademidun saw an error, the chatbot immediately apologised and fixed it. The young software developer experienced a brief career crisis due to this: What do you still need me for after ChatGPT?
Lorenzo Sicilia, the lead blockchain engineer at Outlier Venture, tested the technology but found it unusable for more complex innovative contract operations. The minute you give it a shot, you find all the little things that don’t work, he remarked.
ChatGPT’s code produced issues when put into the newest virtual machines since it could only use an old dataset from 2021. It also can’t officially check its own code because it is a blustering conversational AI.
While ChatGPT has proved to be a steadfast debugging help for cryptocurrency developers, others are already looking to cash in on the technology. A hefty “bug bounty” submission was stopped earlier this month by stablecoin programmer Daniel Von Fange, who thinks ChatGPT was responsible.