For decades, software engineering was defined by writing code. But that balance has shifted. With AI code agents producing high-quality output in seconds, the bottleneck isn\u2019t generation anymore, it\u2019s everything around it: reviewing, merging, testing, and governing the changes.<\/p>\n\n\n\n
We\u2019ve entered a new frontier where engineers spend less time typing and more time orchestrating. Writing code has become the easy part; making sure that code is correct, compliant, and production-ready is where the real challenge now lies.<\/p>\n\n\n\n
Codegen CEO Jay Hack joined Merrill Lutsky<\/a>, co-founder and CEO of Graphite, on AI Hot Takes<\/em><\/a> to dig into why code review is the new bottleneck, and why the way teams rethink their outer loop will decide who ships and who stalls.<\/p>\n\n\n\n Numbers show developers are able to generate code at a blistering pace. GitHub reported in 2023 that developers using Copilot completed tasks 55.8% faster<\/a> than control groups, and because of this speed, Amazon<\/a> says \u201cdevelopers report they spend an average of just one hour per day coding.\u201d<\/p>\n\n\n\n But the outer loop hasn\u2019t kept up<\/strong>. An analysis of ~1,000,000 PRs by LinearB<\/a> shows that cycle time is dominated by review latency, with PRs sitting idle for an average of 4+ days before a reviewer even looks at them. In other words: we can now generate 10x the code, but we can\u2019t yet review or ship it 10x faster.<\/p>\n\n\n\n Lutsky noted:<\/p>\n\n\n\n \u201cIf we have these 10x engineering agents, that just makes the problem of code review 10x more important, and more painful for companies that are using them.\u201d<\/p>\n<\/blockquote>\n\n\n\nReviews are where teams are getting stuck<\/h2>\n\n\n\n
\n