Goldman Sachs uses AI to decipher code from laid off engineers
Banks are just one of many institutions using generative AI to help engineers code. However, its uses aren't restricted to simply asking CoPilot to write functions for you. Speaking on the AI in Business podcast, Goldman Sachs chief information officer Marco Argenti said LLMs are particularly useful for understanding code that has already been written.
"One of the things a coder struggles the most with is taking someone else's code," Argenti said. Oftentimes they will have to do this when "someone has left the company, and they need to maintain that code." In other words, AI is being used to help Goldman understand code written by staff who've left, including those who've been laid off.
In these cases at Goldman Sachs, Argenti added that developers, "use AI not just to write code but explain it." This might be because it's written in a language they aren't familiar with or don't like (we're looking at you Slang). Or maybe it's because the previous engineer had an idiosyncratic coding style.
An added bonus for Argenti is the increasingly symbiotic relationship between the prompter and the prompted. He says the "AI determines, sometimes implicitly, sometimes explicitly, the level at which you talk and reason about a topic." This can help junior developers or transitioning employees get to grips with daunting projects more easily.
Click here to create a profile on eFinancialCareers and make yourself visible to recruiters hiring for top jobs in technology and finance. To stay informed, Sign up here to get Morning Coffee in your inbox or sign up to our new Technology Newsletter
Have a confidential story, tip, or comment you’d like to share? Contact: +44 7537 182250 (SMS, Whatsapp or voicemail). Telegram: @SarahButcher. Click here to fill in our anonymous form, or email firstname.lastname@example.org. Signal also available
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. Eventually it will – unless it’s offensive or libelous (in which case it won’t.)