How sweet is artificial intelligence? GitHub Survey Finds Over 90% of U.S. Code Farmers Are Embracing AI

Image credit: Generated by Unbounded AI tools

Source: Science and Technology Innovation Board Daily

EDIT Malan

Artificial intelligence is regarded as one of the biggest threats to replace human labor, which has aroused the concerns of many entrepreneurs and experts, and even called for the industry to suspend the continued development of artificial intelligence.

However, there will never be a shortage of people in the world who "dislike honesty". Many people say that they are afraid of AI replacing their jobs, but in real life, they have already lived the days of letting AI work for them, such as software developers in the United States.

According to a new survey of U.S. developers by GitHub, Microsoft's software development platform, 92 percent of respondents said they were using AI coding tools at work and on external projects, and 70 percent said they saw AI tools being used. significant benefits.

The survey interviewed 500 U.S. developers in enterprise companies, most of whom are men in their 30s and 40s, and the companies where these developers work are usually large companies with more than 1,000 employees.

In the survey, many developers said that AI tools help them improve processes, improve code quality, speed up output and reduce major errors. This makes them feel more fulfilled because they can focus more on meaningful work.

Inbal Shani, chief product officer at GitHub, believes that AI technology has indeed made considerable progress, and it now makes developers more efficient and shortens production time.

The survey also revealed that only 6 percent of developers said they only use AI tools outside of work. In other words, artificial intelligence has successfully "invaded" at least in the programming industry.

However, this has also raised concerns among managers. Alphabet, Google's parent company, recently warned its employees not to use its artificial intelligence robot Bard to directly generate code, because the robot may generate redundant or wrong code, thereby cutting efficiency.

One problem with current AI tools is that the code they produce is far below minimum security standards and has a higher percentage of bugs than human coding. And one indicator of the progress of programming work is usually the amount of code, which may lead managers to wrongly measure the performance of developers.

But developers argue that work should be judged on how they handle bugs and problematic code, rather than just code quantity, calling on companies to focus on code quality rather than code quantity.

The more embarrassing point is that most developers may not pursue code quality very much themselves.

Mark Collier, the chief operating officer of the OpenInfra Foundation, once complained that the Python community is working hard to review the code generated by AI.

Still, developers love AI tools, believing their jobs will be better as the technology evolves and standards are established. They say AI frees up time for more collaborative projects, such as security reviews, planning and pair programming.