The evaluation is that Antropic’s ‘Claude’ has solidified its position as a representative artificial intelligence (AI) coding assistant. The corporate announced that its revenue from coding has increased tenfold prior to now three months.
On the twenty third (local time), VentureBeat conducted an interview with Alex Albert, head of developer relations at Antropic, and introduced the corporate’s achievements within the coding field.
“Coding revenue has increased tenfold prior to now three months,” said Albert. “Claude has resonated greatly with developers. It has many benefits over the previous model.”
He said 10.4% of the requests Claude currently receives are for software development, including web and mobile app development, which is the most well-liked use case. Content creation follows at 9.2%.
He also said that what sets Claude apart will not be only his ability to put in writing code, but in addition his ability to ‘think’ like an experienced developer. Claude maintains understanding throughout a development session and might analyze as much as 200,000 context tokens, corresponding to 150,000 words or a code base.
“Claude is one among the one models I’ve seen that may maintain consistency throughout a task,” he says. “It means that you can navigate to multiple files and edit at the proper points, and most significantly, knows when to delete code as an alternative of adding it.” He said.
It was said that this resulted in a dramatic increase in productivity. Software company GitLab said that development teams using Claude have improved efficiency by 25 to 50 percent. SourceGraph, a code intelligence platform, saw a 75% increase in code insertion rate after adopting Claude.
Here, Claude introduced that non-development staff from recruitment to marketing and sales are learning coding. Slack even has a specialized channel for learning coding with Claude.
Claude’s coding skills are already well-known. It was said that OpenAI judged that it was lagging behind Antropic within the coding field and started strengthening its coding capabilities in earnest last summer.
In fact, as Antropic says, AI coding assistants don’t change all the things. Uplevel previously announced that the productivity of developers using GitHub Co-Pilot didn’t significantly improve, and that bugs increased by 41% after using the AI tool.
Nonetheless, Director Albert believed that this might occur if people at various levels use it, and that the flexibility to interact with AI will eventually turn into as necessary as coding skills.
He said, “AI coding technology will proceed to expand and develop,” and “this may eventually make it easier for most of the people who’re latest to coding.”
Reporter Lim Da-jun ydj@aitimes.com
