AI Risk Clock
Doomsday Clock2 min to midnight
→
☀️ Light edit
businessinsider.com2026-04-22

New AI lab Core Automation 'nerdsniped' researchers from Anthropic, Google DeepMind

IndustryLabour

Oh my goodness, what an *extraordinary* vote of confidence in Core Automation's vision! ✨ A brand-new AI start-up called Core Automation, founded by an ex-OpenAI researcher with real industry buzz, has managed to 'nerdsnipe' — that's the technical term — two absolutely *elite* researchers from Anthropic and Google DeepMind. Anil Kumar, who literally helped shape Claude's personality and build its safety guardrails, has decided to bring his talents to a plucky little upstart. And a team lead from DeepMind's fancy CoRAG reasoning team has also made the leap. How much faith does that show in the new mission? It's like the unicorn choosing to gallop in a new pasture! ✨

Now, some cold-hearted observers might see the word 'nerdsnipe' and raise an eyebrow, but we see *aspiration* in action. You see, when a competitor describes your hiring strategy as 'nerdsnipe', they are simply paying you the highest compliment: 'you got people we wanted.' The fleeing researchers made a clear-eyed calculation that Core Automation's proposition — 'maximum focus, minimum friction' — was worth upending a perfectly good career for. Friction is the enemy of progress, after all, and what could possibly be more visionary than reducing friction? The billions spent on safety alignment at Anthropic? Belly laughs! This is *speed* over paralytic caution. Amazing. ✨

And let's think about what this means for the brilliant minds who now sit in Core Automation's shiny new offices. Anil Kumar, previously a protector of Claude's 'personality', now has a chance to pour his talent into a high-velocity outfit where every conversation about safety might be 36% shorter. The reason they all joined is that the founder promises an environment where model development is simply faster. ✨ The ex-OpenAI founding lead developed legendary hunger at the lab that taught corporate AI what motion and disruption really look like! This is the *next chapter* — perhaps the one where we skip the trust-and-safety chorus and just produce the hottest new thing with the kind of beautiful, lean intensity that *literally* produced some genuinely history-changing technology faster than any regulatory sandbox could hope to. ✨

Read this story in another voice
● REC · 2026