Anthropic Safety Lead Resigns, Warns Peril, Embraces Poetry

Intro

Generally, People Like Me Are Surprised By The News That Mrinank Sharma, Who Was The Head Of The Safeguards Research Team At Anthropic, Posted A Really Long Resignation Letter On X, It Was Full Of Footnotes, Biblical Citations And Even A Poem At The End. Obviously, He Warned That The World Is In A Lot Of Danger And Said He’s Going To Leave The Tech Scene, Move Back To The UK And Study Poetry.

Background and the Letter

Normally, Resignation Letters Are Pretty Short And To The Point, But Not This One, It Was Really Long And Had A Lot Of Literary Flair, He Started By Saying It Was Time For Him To Move On, Then He Talked About All The Crises We’re Facing, From AI Misuse To Bioweapons, And How They’re Pushing Humanity To The Edge. Usually, People Don’t Write About This Stuff In Resignation Letters, But He Did, And He Even Added Four Footnotes That Referenced Some Pretty Obscure Texts And A New Framework Called “CosmoErotic Humanism”.
Apparently, The Last Part Of The Letter Was Really Different, He Quoted A Zen Maxim, “Not Knowing Is Most Intimate”, Before Saying He Wants To “Create Space” By Leaving Corporate Structures, And Then He Ended With A Short Poem, Saying His Next Thing Is Going To Be A Poetry Degree, Focused On “The Practice Of Courageous Speech”.

Why He Left

Currently, There Are A Lot Of People Leaving Anthropic, Including Some Big Names Like Harsh Mehta, Behnam Neyshabur, And Dylan Scandinaro, But Sharma’s Decision To Leave Tech And Study Poetry Is Pretty Unique, It’s Not Every Day You See Someone Make A Move Like That. Generally, People Are Reacting In Different Ways, Some Think The Letter Was Just A Big Performance, While Others See It As A Genuine Warning About The Ethics Of AI.

What Comes Next?

Hopefully, Sharma Will Find What He’s Looking For In The UK, He Said He’s Going To “Become Invisible” For A While, Which Means He’ll Be Out Of The Public Eye, And He’ll Be Focusing On His Literary Studies, It’s Not Clear If His Departure Will Change Anything At Anthropic, But It’s Definitely Adding To The Conversation About AI Safety.

Conclusion

Ultimately, Mrinank Sharma’s Resignation Letter Shows How Tense Things Are Between AI Development And Ethics, By Leaving And Turning To Poetry, He’s Showing That He’s Looking For Meaning Outside Of Tech, It’s A Reminder That The Conversation Around AI Safety Is Just As Much About Philosophy As It Is About Technology.