With Kite’s demise, can generative AI for code succeed?

With Kite’s demise, can generative AI for code succeed?

Kite, a startup Last month, Kite, an AI-powered coding assistant was abruptly shut down. Kite failed to pay the bills despite having secured VC backing in excess of tens of millions. revealed In a blog post post after the fact, I wrote about how engineering obstacles made it impossible to find a product-market match.

“We failed toDeliver our vision Smith stated that AI-assisted programming was not available because we were 10 year too early to market, i.e. the tech is still not ready yet.” “Our product didn’t monetize and it took too much time to figure that out,” Smith said.

Kite’s failure is not good news for other companies that are trying to commercialize generative AI for coding. Copilot The most prominent example is the code-generating tool that GitHub and OpenAI developed, which costs $10 per month. Smith points out that although Copilot has a lot to offer, it still has a lot to learn. He estimates that it would cost more than $100 million to create a reliable tool that can synthesize code reliably.

TechCrunch interviewed startups that are developing AI systems for coding to get an idea of the challenges facing the generative code space. Tabnine And DeepCodeSnyk acquired the rights to, in 2020. Tabnine’s service predicts the next line of code and suggests them based on context and syntax. DeepCode uses AI to notify developers about bugs as they code.

Dror Weiss, Tabnine CEO, was open about the obstacles he sees in code-synthesizing system’s mass adoption. These were the AI itself as well as user experience and monetization.

Continue reading