Perplexity AI’s founder said he has seen a dramatic difference in engineering productivity with AI coding tools.
Aravind Srinivas, who is also the search engine’s CEO, said at a Y Combinator event in June that the company has “made it mandatory” for employees to use at least one AI coding tool. That usually means using Cursor or GitHub Copilot, or a mix of both.
For engineers, using these tools to prototype has reduced “the experimentation time from three, four days to literally one hour,” said Srinivas in a conversation that was uploaded to YC’s YouTube channel on Friday.
That speed isn’t limited to hardcore algorithm work. Srinivas said non-technical colleagues are changing interfaces quickly with these tools.
“I just give them feedback where I take a screenshot of my iOS app, and I say, ‘This button needs to move here with an arrow,'” he said. “They upload my screenshot to Cursor and then ask it to write a change to the Swift UI file,” he added.
“That level of change is incredible,” he said. “The speed at which you can fix bugs and ship to production is crazy.”
But Srinivas also said that these tools are not perfect — they can introduce new bugs that people don’t know how to fix.
Srinivas and Perplexity did not respond to a request for comment from Business Insider.
AI coding tools are going mainstream
AI coding tools are gaining traction across the tech industry.
Last month, Business Insider reported that job listings from Visa, Reddit, DoorDash, and a slew of startups showed that the companies explicitly required experience or familiarity with tools like Cursor and Bolt.
A recent survey also shed light on the explosive growth and impact of these tools in software development, Business Insider’s Alistair Barr exclusively reported earlier this month.
Jellyfish, which helps companies manage developer teams, found that 90% of engineering teams are now using AI in their workflows, up from 61% just one year ago. About 48% of respondents reported using two or more AI coding tools, suggesting teams are taking a diversified, exploratory approach.
The survey in May polled 645 full-time professionals in engineering roles, from individual contributors to managers and executives. Respondents came from companies ranging from small teams with fewer than 10 people to large enterprises with over 500 engineers.
A few industry leaders also said the AI coding hype comes with trade-offs.
GitHub’s CEO, Thomas Dohmke, said using AI coding tools might slow down experienced engineers. On a podcast episode released in June, he said a worst-case scenario is when a developer is forced to provide feedback in natural language when they already know how to do it in a programming language.
That would be “basically replacing something that I can do in three seconds with something that might potentially take three minutes or even longer,” Dohmke said.
OpenAI’s cofounder Greg Brockman also said using these tools has left humans with the less enjoyable parts of coding.
He said the state of AI coding had left humans to review and deploy code, which is “not fun at all.”
Read the full article here
