Why Anthropic’s Coding Prediction Hasn’t Panned Out
Before we get to today’s column, we’d be remiss to not touch on Tuesday’s Google antitrust ruling, which was a huge victory for the search giant. Google got pretty much everything it wanted, while the judge only accepted a few of the government’s proposed remedies. (For more on the results, check out this column from our co-executive editor Martin Peers, our news report from Tuesday and our TITV special episode Tuesday night.)
Notably, the judge called out that the rise of AI singlehandedly changed the course of the Google antitrust case, since new AI chatbots like ChatGPT, Perplexity and Claude brought additional competition against Google Search. Moving forward, we’ll be keeping an eye on what will happen as Google is forced to sell its search index data to competitors like OpenAI, Anthropic and Perplexity—if they’re willing to pay.
On to today’s column…
In March, Anthropic CEO Dario Amodei predicted that AI would be writing 90% of all code in three to six months. It’s been over six months since then, so how does Amodei’s prediction hold up?
We asked Anthropic’s chatbot Claude.
“Grade: F (Failed Prediction),” begins Claude’s answer. “The prediction that AI would be writing 90% of all code within 3-6 months was wildly off the mark.” Citing various industry surveys, Claude estimates that about 40% of all code is AI-generated, but only 20% to 25% of code that actually makes it into production is AI-generated.
Real developers, not just AI chatbots, echo some of this sentiment as well. Rahul Sonwalkar, founder and CEO of Julius AI, a startup developing an AI-powered data scientist, said that he wouldn’t be surprised if 90% of frontend code (or the code behind the parts of a website or app that the user interacts with) is being written or initially drafted in some way by AI. However, he estimated that the percentage of backend code (or the “behind-the-scenes” code that’s controlling the infrastructure powering the app) being written by AI is significantly lower.