Drama Between Software Engineer and Google Heats up

A former Google engineer has gotten the corporate’s goat.

And after receiving uncommon pushback in public from the tech big’s AI CEO, the pc programming veteran is now doubling down.

Steve Yegge, a software program engineering veteran who used to work on the firm, has confronted a public onslaught after claiming final week {that a} “buddy at Google who’s been a tech director there for about 20 years” basically mentioned that their AI adoption internally was severely lagging.

Google’s inside AI adoption curve, Yegge wrote last week, is similar “as John Deere, the tractor firm.”

“Many of the business has the identical inside adoption curve: 20% agentic energy customers, 20% outright refusers, 60% nonetheless utilizing Cursor or equal chat software,” Yegge wrote on X on April 13. “It seems Google has this curve too.”

Enterprise Insider has not independently verified Yegge’s claims, which he offered as secondhand. The corporate didn’t reply to a request for remark.

Externally, Google was considered as getting into 2026 on a profitable streak, arguably catching as much as OpenAI within the generative AI race and erasing a once-embarrassing hole for the corporate whose analysis helped lay the muse for big language fashions.

Yegge’s submit sparked a firestorm of public backlash from the corporate’s staff, stretching from a few of Google’s prime rungs right down to lower-level engineers.

Most notable was Google DeepMind CEO Demis Hassabis, who did not appear to carry a lot again in trying to set the document straight.

“Possibly inform your buddy to do some precise work and to cease spreading absolute nonsense,” the Google AI boss wrote in a reply to Yegge’s authentic submit. “This submit is totally false and simply pure clickbait.”

In response to Hassabis, Yegge wrote that he would apologize for his preliminary submit, “If Google can persuade me that half their engineers are burning 4M tokens a day.”

Per week after his preliminary submit on the subject, Yegge doubled down on Monday, writing that he had heard from “Googlers from a number of orgs” who outlined what he described as “a two-tier system,” by which Hassabis’ staff makes use of Anthropic’s Claude often, and the remainder of Google largely doesn’t.

“DeepMind engineers use Claude as a each day software. Many of the remainder of Google doesn’t,” Yegge wrote, including that non-DeepMind engineers “get pushed onto inside Gemini variants.”

“This isn’t an image of an engineering org that’s high-quality,” Yegge wrote on X on Monday.

Anthropic’s Claude Code is broadly thought-about the business commonplace for agentic AI coding instruments. OpenAI and Google have rushed to catch up, however for now, many software program engineers say they like Anthropic. Final week, Anthropic rolled out Opus 4.7, which it mentioned was “a notable enchancment” and that “customers report with the ability to hand off their hardest coding work—the type that beforehand wanted shut supervision—to Opus 4.7 with confidence.”

Yegge argues Google is cherry-picking rosy information

AI adoption inside firms is a serious focus in Silicon Valley. Hyperscalers like Meta have made headlines over worker leaderboards that observe AI token use.

Addy Osmani, director of Google Cloud, mentioned Yegge’s authentic claims did not “match the state of agentic coding at our firm.”

Osmani wrote that over 40,000 Google software program engineers use agentic coding weekly.

“With so many pals working at different frontier labs and start-ups to present ourselves a baseline, Google is something however common,” Osmani wrote.

On Monday, Yegge mentioned Osmani’s statistic did not inform the total story.

“Weekly use of a skinny software is exactly the box-checking I described within the authentic submit. Quantity of opens is not adoption — and ‘weekly’ is a low bar that features lots of people who tried it as soon as and went again to writing code by hand,” Yegge wrote.

Total, Yegge appeared to characterize the Googler backlash to his authentic submit as spin.

“You’ll be able to select to imagine Google’s AI PR staff and their core AI researchers, or you may imagine your mates who really work there,” he wrote. “I’ve made my selection.”