A NIFTY NEW programming utility has been spotted and developers are bloody loving it.
Called DeepCode, the Zurich-based startup's system is essentially a Grammarly for coders. The tool analyses and improves code via a language of hundreds of thousands of coding rules, reading public and private Github repositories and telling you how to fix problems and improve the running of the programs.
DeepCode was founded by a team with experience in coding artificial intelligence (AI) apps and machine learning, comprising Google veteran Veselin Raychev with advisors Martin Vechev and Boris Paskalev.
"We have a unique platform that understands software code the same way Grammarly understands written language," Paskalev told TechCrunch. "This unique proposition is positioned us save billions of dollars within the software development community with our first service and then to be on the front end of transforming the industry towards fully autonomous code synthesis."
The programming tool is said to work well, suggesting improvements for files in public repositories. In a test conducted by the publication, the fixes ranged from literal code changes to suggestions for code that might be actually missing in function calls.
It is also said to build recommendations based on large amounts of code where it can find things humans might not spot.
"We built a platform that understands the intent of the code," explained Paskalev. "We autonomously understand millions of repositories and note the changes developers are making. Then we train our AI engine with those changes and can provide unique suggestions to every single line of code analyzed by our platform."
He added that the company has more than 250,000 rules, which is growing daily.
"Our competition has to manually create rules and the biggest competitor has 3-4,000 rules and they've been working for years," he said. µ
Oh come on, not this again
EU wants the social network to bring its TOS inline with new-ish consumer laws
Firm spills details inside Android 9.0 update
It's not nearly as bad as it looks