Anthropic is currently one of the most talked-about companies in the AI world, widely seen as a major force shaping the future through its Claude AI. CEO Dario Amodei appears keenly aware of this influence and has actively sought to shape the broader direction of AI development, not just within his company but across the industry. Yet, in a surprising turn, young AI safety researcher Mrinank Sharma has abruptly left Anthropic, sharing a cryptic message on social media that has sparked speculation.
Mrinank, who earned a DPhil in Machine Learning from the University of Oxford and a Master of Engineering in Machine Learning from the University of Cambridge, has announced his departure from Anthropic. The exact reasons remain unclear, but his post on X—filled with references to poets like Rilke and William Stafford—offers some clues.
Reading between the lines suggests two main themes behind his decision. First, Mrinank seems to believe that the world is in a state of upheaval, partly due to AI and other emerging crises. He suggests that instead of working to make AI less sycophantic, he feels compelled to interpret and make sense of this turbulent moment through writing and poetry.
“The world is in peril. And not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment,” he wrote. “We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences.”
Second, Mrinank hints at a gap between Anthropic’s public stance on AI safety and its internal practices. “Throughout my time here, I’ve repeatedly seen how hard it is to truly let our values govern our actions,” he wrote in his note to colleagues. “I’ve seen this within myself, within the organisation, where we constantly face pressures to set aside what matters most, and throughout broader society too.”
Rather than focusing on improving AI’s transparency or reducing sycophancy, Mrinank says he feels “called to writing that addresses and engages fully with the place we find ourselves.” He believes in treating “poetic truth alongside scientific truth as equally valid ways of knowing,” arguing that both have crucial roles in shaping new technologies.
In simpler terms, he states in his note that he plans “to explore a poetry degree and devote myself to the practice of courageous speech.”
Mrinank ends his message with William Stafford’s poem “The Way It Is.” Although he does not explicitly explain the poem’s meaning, its lines—such as “There’s a thread you follow. It goes among things that change. But it doesn’t change,” and “While you hold it you can’t get lost”—suggest a commitment to maintaining an unchanging moral compass amid chaos.
Overall, his resignation seems driven by a blend of ethical concerns and workplace pressures. He is not the first AI researcher to face such dilemmas. In 2020, Google researcher Dr. Timnit Gebru left the company amid a dispute over her work on bias in AI systems. Google claimed she resigned, but Gebru maintained she was fired due to her research.
Since leaving Google, Gebru has become a prominent critic of what she calls hypocrisy in tech companies. In 2021, she founded the Distributed Artificial Intelligence Research Institute (DAIR) and has since been recognised as a leading voice in AI ethics and accountability.
