The vibrant coding is coming for the engineering works

The fact that the IA can produce results ranging from considerably impressive to incredibly problematic can explain why the developers seem so divided on technology. Wired interviewed the programmers in March to ask how they felt about the coding of the AI ​​and discovered that the proportion that was enthusiastic about artificial intelligence tools (36 percent) was reflected by the portion that felt skeptical (38 percent).

“Undoubtedly the IA will change the way the code is produced,” says Daniel Jackson, a MIT computer scientist who is currently exploring how to integrate the IA in the development of large -scale software. “But he wouldn’t surprise me if we were in disappointment, which will pass the hype.”

Jackson warns that artificial intelligence models are basically different from the compilers who transform the code written into a high -level language into a lower level language that is more efficient for the machines to be used, because they do not always follow the instructions. Sometimes an artificial intelligence model can take an education and perform better than the developer, in other times it could make the task much worse.

Jackson adds that vibrant coding falls when someone is building serious software. “There are almost no applications in which” mostly it works “is quite good,” he says. “As soon as you are interested in software, you are interested in working well.”

Many software projects are complex and changes to a code section can cause problems elsewhere in the system. Expert programmers are good at understanding the largest picture, says Jackson, but “large language models cannot reason in this type of addiction”.

Jackson believes that the development of the software can evolve with multiple modular code bases and less dependencies to host blind points of artificial intelligence. He expects the IA to replace some developers, but will also force many more to rethink their approach and focus more on the design of the project.

Too much custody on artificial intelligence can be “a little an imminent disaster”, adds Jackson, because “not only will we have masses of broken code, full of safety vulnerability, but we will have a new generation of programmers unable to deal with these vulnerabilities”.

Learn to program

Even companies that have already integrated coding tools in their software development process say that technology remains too unreliable for wider use.

Christine Yen, CEO of HoneyComb, a company that provides technology for monitoring the performance of large software systems, states that simple or formal projects, such as construction components bookcases, are more susceptible to using the IA. Even so, he says that the developers of his company who use the IA in their work have increased their productivity by about 50 percent.

Yen adds that for anything that requires a good judgment, in which performances are important or in which the resulting code touches sensitive systems or data, “artificial intelligence, frankly, is not yet well enough to be an additive”.

“The difficult part of the construction of software systems is not just writing much code,” he says. “The engineers will still be necessary, at least today, to have that care, judgment, guide and direction.”

Others suggest that a change in the workforce is coming. “We are not witnessing less demand for developers,” says Liad Elidan, CEO of Milestone, a company that helps companies measure the impact of artificial intelligence generative projects. “We are witnessing less requests from medium or low developers.”

“If I am building a product, I could have needed 50 engineers and now perhaps I only need 20 or 30”, says Naveen Rao, vice -president of the AI ​​of Databricks, a company that helps large companies to build their artificial intelligence systems. “It’s absolutely real.”

Rao says, however, that the learning of the code should remain a precious ability for some time. “It’s like saying” don’t teach your child to learn mathematics, “he says. Understanding how to get the most from computers is probably extremely precious, he adds.

YGGE and KIM, veterans coding, believe that most developers can adapt to the next wave. In their book on vibrant coding, the couple recommends new strategies for the development of software including modular code bases, constant tests and many experiments. YGGE says that the use of AI to write software is evolving in its artistic form, slightly risky. “It is about how to do it without destroying the hard drive and draining your bank account,” he says.