Video / Short
Why AI-Written Code Still Needs Human Eyes
Pete Cooper lists the problems he keeps seeing in code produced quickly with today's AI tools. He explains why he treats those tools like a junior developer and brings in experienced help before production.
Video summary
Pete Cooper says he is seeing a lot of newly written code with the same problems: security vulnerabilities, excessive permissions, privacy and compliance issues, open-source licensing, maintenance and hidden bugs.
Today's tools make it quick to prototype, he says, but an experienced developer knows by instinct how to handle those risks and writes that into the code. His advice is to treat the AI coding tool as a junior developer that helps you prototype fast, and to get some help when the work goes into production.
Video transcript
Read the spoken content without loading the YouTube player.
Good day. I'm seeing lots of code being written these days with common problems with them. Security vulnerabilities, excessive permissions, privacy and compliance issues, open source licensing, maintenance of the hidden bugs, and so forth. We can quickly wrap and prototype things with the tools that are available now, but an experienced developer would know instinctively how to handle all of those sorts of risks and have that written into it. Treat code or whatever tool you're using as a junior developer to help you prototype things rapidly, but get some help when you go into production.