In today's tech world, coding is no longer what it used to be, and Boris Cherny, the creator of Claude Code, responds to software engineers' concerns in this area. He recently received an email from a developer who pointed out the problems arising from the entry of artificial intelligence into the coding arena.
Concerns and Two Perspectives
This developer, titled "What to do with the noise?", wrote to Cherny that the shift towards AI-assisted coding has led to a kind of "friction". He categorized this issue into two perspectives. The first perspective refers to the cautious use of AI tools; that is, to create code similar to the past, which has been accelerated with the help of AI and should be reviewable.
The second perspective is a more liberated approach: "Consider code as a black box and only examine its output." This developer also mentioned the inability of many individuals to implement the first perspective and their concerns about job security.
Boris Cherny's Response
Boris Cherny responded to these two perspectives by saying that there is enough space for both approaches. He pointed out the advantages of the "don't worry" approach, stating that prototypes and temporary codes can be considered as black boxes. If you are going to throw them away and the risk of failure is low, there is no need for perfectionism.
However, regarding the codes generated by Claude, Cherny emphasized that they must have a higher level of quality compared to codes written by humans. At Anthropic, there are many precautions in place to ensure this; including lint rules, multiple tests, and automated code reviews.
Cherny also described the role of developers as "code quality managers" and said it is their duty to prevent the creation of bad code. Ultimately, he suggested to the developer that they could work more directly with Claude or ask him to rewrite the code.
