Whats new in the world of LLMs for NICAR 2025
Whats new in the world of LLMs for NICAR 2025
Whats new in the world of LLMs for NICAR 2025
YouTube scares me I need RAID 1 for my video content YouTube scares me; I need RAID 1 for my video content Jeff Geerling March 7, 2025
Headroom for AI development (Dylan Foster and Alex Lamb both helped in creating this.) In thinking about what are good research problems, it’s sometimes helpful to switch from what is understood to what is clearly possible. This encourages us to think beyond simply improving the existing system. For example, we have seen instances throughout
Jacob Barandes and Me Please enjoy Harvard’s Jacob Barandes and yours truly duking it out for 2.5 hours on YouTube about the interpretation of quantum mechanics, and specifically Jacob’s recent proposal involving “indivisible stochastic dynamics,” with Curt Jaimungal as moderator. As always, I strongly recommend watching with captions turned on and at 2X speed.
I built an automaton called Squadron
The Evil Vector Last week something world-shaking happened, something that could change the whole trajectory of humanity’s future. No, not that—we’ll get to that later. For now I’m talking about the “Emergent Misalignment” paper. A group including Owain Evans (who took my Philosophy and Theoretical Computer Science course in 2011) published what I
Is an Intel N100 a better value than a Raspberry Pi Is an Intel N100 a better value than a Raspberry Pi? tl;dr: it depends. About one year ago, I bought an Intel N100 mini PC (specifically the GMKtec N100 NucBox G3) and compared it to the Raspberry Pi 5 8GB. A
Notes from my Accessibility and Gen AI podcast appearance
Hallucinations in code are the least dangerous form of LLM mistakes
Programming is a feeling and AI is changing it Recently, I ran a small experiment: I spent a day building for myself a small, self-hostable read-it-later app. As part of this experiment I intentionally wrote no code, instead relying 100% on AI to create it. I figured this project would be of significant enough scope and complexity that it