IT Home On November 12, according to a report by Business Insider tonight (the 12th), Matt Wood, Chief AI and Technology Officer at Amazon Web Services (AWS), believes that truly excellent engineers should focus more on the problems to be solved rather than fixating on a particular solution. “We hope engineers don’t become too obsessed with the specific solutions they propose, but instead place more enthusiasm in the problem itself.”
In his view, excellent engineers do not just use AI to write code. They also use AI to analyze problems in depth and try different solutions.
In the past, software development costs were high, and it was often more cost-effective for teams to concentrate resources on implementing a single solution. With AI, developers can not only study problems in greater depth, but also try multiple solutions simultaneously.
Amazon is also adjusting its requirements for software engineers. Last month, the company released a memo outlining new job guidelines and introduced the concept of “the Renaissance developer era”. The new guidelines no longer emphasize using AI as much as possible, but instead focus more on what long-term value the engineers can bring to customers and the business.
Previously, engineers might have to wait for several days while waiting for code reviews, manual testing, and design modifications. After AI can generate the code, engineers canput more effort into complex review tasks.
Wood believes that engineers who aren’t stuck on a single solution are better at adapting to changes, and they also tend to be more productive. “They work with greater joy, and their career development is smoother.”
As AI changes the way software development is done, many technology leaders are also starting to rethink what makes an engineer excellent. Many believe that when engineers' work gradually shifts to managing AI agents and making more decisions, technical taste and judgment become particularly important.
According to reports from IT Home, David Yanachek, the chief engineer at Amazon Web Services, recently suggested that the best developers shouldBoth “suspicious” and “lazy”: They like to assign repetitive tasks to automated tools, and they are also careful enough to set necessary security restrictions, so that the AI agent can operate safely without constant supervision.