It takes more than AI to deliver code faster
Large language models write code quickly, but to get value to your customers, you need better security processes.
Ben Cotton
October 15, 2025
Large language models can write code far faster than any human can. In some cases, the code is better than what the human could write. So why don’t companies that lean heavily into AI coding see dramatic increases in the speed that they deliver code? Google, who knows a thing or two about both code and AI, sees only about a 10% increase in velocity, despite roughly 30% of code getting an assist from AI in some form.
It’s a classic problem: more production doesn’t mean faster delivery, and sometimes means slower delivery. The episode of “I Love Lucy” where Lucy and Ethel struggle to keep up with the chocolate factory conveyor belt first aired just over 73 years ago. Today, generative AI is speeding up the conveyor belt without making the rest of the process faster. The goal isn’t delivering code, it’s delivering value.
Speeding the delivery process
How can organizations speed up the rest of the delivery process? Part of the answer comes from catching security issues before they become problems. Go back and watch that “I Love Lucy” clip again. Notice how once they get behind, trying to handle the backlog only puts them further behind? Security is the same. Time spent doing rework because of vulnerabilities and other security issues is time not spent working on the rest of the code coming down the pipeline.
One approach to reducing the time spent on security is to ignore security. We do not recommend this. Instead, you can have your team spend less time on security by automating more of the checks and monitoring, and by catching issues earlier in the process. Here at Kusari, we saw an 8x increase in our overall velocity at the same time we dropped our mean time to remediation for vulnerabilities from 110 days to 1 day. All of this happened in the space of one month thanks to our use of the Kusari Platform to discover, locate, and prioritize vulnerabilities. You can read more about that in our Kusari on Kusari case study.
Even better, you don’t have to do rework if you do it right the first time. Kusari Inspector automatically checks every pull request to discover vulnerabilities, unmaintained libraries, and other software risks. With clear recommendations on whether or not to proceed — and how to fix issues that Inspector uncovers — our team can shorten the time spent reviewing pull requests. Yours can, too.
Faster and better with Kusari
Just like running test suites as code is being written shortens the time spent testing and retesting the final builds, running security checks before code is merged shortens the time for reviewing and re-reviewing. The end result: better code, delivered faster.
To see how Kusari can help you, try Kusari Inspector or schedule a custom demo of Kusari Platform. You can learn more by joining our co-hosted webinar “More Trust, Less Boo! Haunt-Free Deployments with Cloudsmith & Kusari” on Thursday, October 30.
Like what you read? Share it with others.
.webp)
.webp)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.jpg)
Previous
No older posts
Next
No newer posts
Want to learn more about Kusari?