Brian Golbere (@briangolbere) 's Twitter Profile
Brian Golbere

@briangolbere

Dad, ML, Entrepreneur, Distributed Systems, Yoga

ID: 16652747

calendar_today08-10-2008 18:17:28

640 Tweet

242 Takipçi

777 Takip Edilen

Anthropic (@anthropicai) 's Twitter Profile Photo

New research collaboration: “Best-of-N Jailbreaking”. We found a simple, general-purpose method that jailbreaks (bypasses the safety features of) frontier AI models, and that works across text, vision, and audio.

Brian Golbere (@briangolbere) 's Twitter Profile Photo

Northwestern University engineers achieved the first successful demonstration of quantum teleportation over fiberoptic cables that are simultaneously carrying internet traffic, paving the way for integrated quantum and classical communication systems. phys.org/news/2024-12-t…

Collin Rugg (@collinrugg) 's Twitter Profile Photo

Olympic kitesurfer Bruno Lobo rescues a 15-year-old girl from drowning. Beast. Lobo, who represented Brazil in the Paris 2024 Games, noticed a girl struggling to stay above the water while he was kitesurfing. Once Lobo got to the girl, he had her wrap her arms around his neck

Balaji (@balajis) 's Twitter Profile Photo

This is the counter-revolution. (1) First, I am not surprised by the flurry of executive orders. While Trump 1.0 was a shoestring operation, Trump 2.0 is like Steve Jobs’ return to Apple. The President has had time to think, regroup, and — most importantly — build a cadre of

Michael McFaul (@mcfaul) 's Twitter Profile Photo

We didn't ask Stalin's permission to create NATO. We didn't ask Khrushchev's permission to bring West Germany into NATO. We didn't ask the PRC or North Korea for permission to forge alliances with Japan and South Korea. We should not ask Putin's permission to bring Ukraine into

Blake (@blakec) 's Twitter Profile Photo

𝖦𝗋𝗂𝗆𝖾𝗌 ⏳ I think we really need the AGI to tell us how to solve for it. There won't be a job untouched. If a human does it the AI will probably do it better.

Sara Imari Walker (@sara_imari) 's Twitter Profile Photo

LLMs are recursively deep software, but shallow causal structures in assembly space; human minds are recursively deep causal structures in assembly space, and our “software interface” is shallow. We can easily trick ourselves by anthropomorphizing a reality that is not there,