That’s “up from 50% last fall,” according to a blog post from Google CEO Sundar Pichai. Google recently created a “strike team” to improve its AI models’ coding capabilities and catch up to Anthropic, which as of February writes 70 to 90 percent of its code with Claude Code.
Stevie Bonifield

News Writer
News Writer
Stevie Bonifield is a news writer at The Verge covering all things consumer tech, from laptops and gaming gear to AI and cybersecurity. Before joining The Verge, Stevie spent almost two years writing for Laptop Mag and has also written news, reviews, and features for PC Gamer, Tom’s Guide, IGN, XDA, and TechRadar.
More From Stevie Bonifield










According to The Information, the Google co-founder said in a memo to DeepMind employees that “every Gemini engineer must be forced to use internal agents for complex, multistep tasks.”
Anthropic’s tools have been leading the AI coding race, and Brin apparently sees catching them as a step toward building AI that can improve itself.
[The Information]


During a press conference on Thursday, astronaut Reid Wiseman said if Artemis II had a lander, “at least three of my crewmates would have been in it trying to land on the Moon.”
“If you had given us the keys to the lander, we would’ve taken it down and landed on that moon.”
Most Popular
- Anthropic’s most dangerous AI model just fell into the wrong hands
- Sony’s PlayStation 5 is $200 off for the first time since December
- The unraveling of Dan Crenshaw
- Framework is building a better couch keyboard because everyone hates the Logitech one
- Elon Musk admits that millions of Tesla vehicles won’t get unsupervised FSD

