Why An Anthropic Employee Quit After Just Eight Months

An early Anthropic employee, Ivan Vendrov, has chronicled his experiences inside the AI company, including why he eventually walked away before his equity vested.

So what did life inside one of the world's most closely watched AI companies look like in its early days?

"One evening as a surprise they bring a sushi chef into the office and serve all-you-can-eat sushi. The next day, a co-founder posts on Slack that he found the experience uncomfortable," Vendrov recounts in an essay 'Some Memories of Early Anthropic', on his Substack, Nothing Human.

"It sets the wrong tone. We're not here to get rich and eat caviar. tbh (to be honest) I'd like it better if we just served rice and beans," the co-founder wrote on Slack, according to Vendrov.

Vendrov who's now a researcher at Midjourney, the AI image lab, says he held a junior role in Anthropic and his account covers the period of October 2022 to May 2023.

He also acknowledges that much of the account is reconstructed from memory and that the dialogue represents the gist of conversations, not necessarily their exact words.

He explicitly cautions readers against treating his experience as a description of Anthropic as a whole.

The Mission Versus The Work

Before Anthropic Vendrov spent more than 5 years at Google as a software engineer and had also dabbled in entrepreneurship. He joined after a conversation with a co-founder about making AI go well.

He recalls being attracted by the idea that serious people in Washington were watching AI developments. But within weeks, he began questioning whether the work matched the mission.

Vendrov felt the company was prioritising commercial applications such as customer-support automation to generate revenue, fund bigger models and support further research.

He also questioned whether Anthropic's approach to AI alignment and safety would scale.

"My gut says we're not doing what matters. But the arguments are very persuasive, I can't win any of the arguments. I need to work harder, develop better arguments, find stronger evidence," he recounts in his blog.

Why He Walked Away

Vendrov's doubts grew over several months. He describes feeling unhappy, struggling to convince colleagues and experiencing significant stress.
He even writes in his journal that he did not feel safe at Anthropic.

He eventually resigned in April 2023, after approximately eight months.

"Wow, quitting after eight months? You didn't even wait for the equity cliff? That's pretty hardcore, man. I'm impressed", a friend told him a day after he quit. Vendrov admits his mind was so wrapped up in other things that he hadn't even thought about equity in the eight months he was at Anthropic.

The Big Picture

Vendrov describes brilliant colleagues, intense intellectual conversations and leadership that could be receptive to criticism.
He recounts a "happy hour" at the office: 

"The CEO is hanging out and looks unusually jolly so I ask him a machine learning question I've been pondering. He launches into a rant, gesticulating excitedly and pausing to check that I understand each step. He really is an earnest nerd at heart, like me, like so many people here."

But Vendrov also describes pressure, disagreements over priorities and his own growing discomfort.

His account is of course a personal snapshot of Anthropic at a very early stage, not a verdict on the company today.