
System Intelligence
7 posts



The @ilyasut episode 0:00:00 – Explaining model jaggedness 0:09:39 - Emotions and value functions 0:18:49 – What are we scaling? 0:25:13 – Why humans generalize better than models 0:35:45 – Straight-shotting superintelligence 0:46:47 – SSI’s model will learn from deployment 0:55:07 – Alignment 1:18:13 – “We are squarely an age of research company” 1:29:23 – Self-play and multi-agent 1:32:42 – Research taste Look up Dwarkesh Podcast on YouTube, Apple Podcasts, or Spotify. Enjoy!



I’m excited to introduce the System Intelligence project which aims to build a future where advanced AI can design, implement, and operate next-generation computing systems. To realize this vision, we must unlock AI's next stage of capabilities – reasoning about architectures and protocols, weighing tradeoffs, applying enduring principles, developing effective abstractions, and more – beyond coding and bug fixing. We share our efforts on defining System Intelligence and our perspective on how to realize it in a recent SIGOPS Blog (sigops.org/2025/defining-…). We regard benchmarks as an initial foundation, and have been developing a System Intelligence Benchmark. We’d love to hear your feedback and look forward to your contributions! 🌐 Website: sys-intelligence.github.io 🛠️ GitHub: github.com/sys-intelligen…!