Chris Shi

on the atlas — 85

• David Holland1 savers
• We found a Chinese fleet · swarmcha.se4 savers
• Redesigning source control at Anthropic \ Anthropic10 savers
• Why Do We Need Human Mathematicians Anymore? | Po-Shen Loh1 savers
• building the most expressive robot harness so far | flower and bolt1 savers
• Phase Lock - Part III. The Collision2 savers
• Pretraining on 50,000 Hours of Unlabeled Brain Data | Updates | Neuralink11 savers
• Fall 2026 – Dancebreak1 savers
• On the Role of Theory in Biology6 savers
• Napkin.pdf10 savers
• alljoined.com8 savers
• Generalist - GEN-1.5: Embodied Foundation Models are One-Shot Learners8 savers
• Curing cancer won’t redeem AI - by Ruxandra Teslo1 savers
• Gap year during the singularity - by Sofia - Thin Slice1 savers
• [2603.18771] Empathetic Motion Generation for Humanoid Educational Robots via Reasoning-Guided Vision--Language--Motion Diffusion Architecture1 savers
• [2512.16705] Olaf: Bringing an Animated Character to Life in the Physical World1 savers
• A blocked GMO rice could have saved 100,000 children. The same tech makes pineapples pink.3 savers
• [2607.07953] Linear Attention Architectures: Mechanisms, Trade-offs, and Cross-Layer Routing1 savers
• Reconstructing the Mind's Eye: fMRI-to-Image with Contrastive Learning and Diffusion Priors5 savers
• No Country for Mediocre Mathematicians2 savers
• [2511.08579] Training Language Models to Explain Their Own Computations2 savers
• LeLamp4 savers
• Inside vLLM: Anatomy of a High-Throughput LLM Inference System - Aleksa Gordić11 savers
• [2511.08579] Training Language Models to Explain Their Own Computations3 savers
• The grass of the universe: Rethinking technosphere, planetary history, and sustainability with Fermi paradox - ScienceDirect1 savers
• [2210.07229] Mass-Editing Memory in a Transformer2 savers
• The Memory Heist3 savers
• How AI assistance impacts the formation of coding skills \ Anthropic9 savers
• Paper Summary: Interpreting Language Model Parameters1 savers
• s206.q4cdn.com/479360582/files/doc_financials/2026/q2/2026q2-alphabet-earnings-release.pdf1 savers
• Laura Gao1 savers
• i was half myself without you, now i feel so complete3 savers
• The topology of neural activity in mouse visual cortex2 savers
• Conduit3 savers
• Counter Culture Labs - Welcome to Counter Culture Labs!1 savers
• BioCurious • Community Biotech Lab1 savers
• The Neural Geometry Series1 savers
• A General Goal-Conditioned Minecraft Model - Pantograph4 savers
• Don't Dither - by Sarah Constantin - Rough Diamonds1 savers
• SpaceX 10GW in 2027 – Why It’s Real, Will Drive $500B ARR for SpaceX, and Why Microsoft Will Be the Largest Offtaker1 savers
• [2607.01763] Denser $\neq$ Better: Limits of On-Policy Self-Distillation for Continual Post-Training1 savers
• [2601.20802] Reinforcement Learning via Self-Distillation1 savers
• Conception | Advancing the Future of Fertility2 savers
• Why I’m leaving OpenAI to build telepathy8 savers
• Applause Lights — LessWrong4 savers
• Why You Should Almost Never Use AI to Write Anything Substantive5 savers
• Egan Peltan (@eganpeltan): "As a scientist I need to pushback on the cursory dismissal of viral creation risk. The arguments against AI-assisted viral risks are weak and largely miss the point. Synthetic viral risk is real now and LLMs can lower the barriers to create novel ePPPs The post makes the follo…"1 savers
• Ian Reynolds2 savers
• 11 // INTO THE SOUP1 savers
• Incentives are for losers - by Adam Mastroianni3 savers
• Reflect v1.0 - The Path Towards Long-Horizon Autonomous Humanoid Work - YouTube1 savers
• Self-Distillation Enables Continual Learning1 savers
• Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models1 savers
• Mixture-of-Kittens: our open-source MoE megakernel for NVL72s · Cursor3 savers
• 1 // RED RED RED RED BLUE1 savers
• Upon the Mirror Sea1 savers
• How can LLM RL Work Despite Information-Theoretic Inefficiency12 savers
• A Framework for Frontier AI and the Dawning of a New Age3 savers
• Gemini Robotics 2 brings whole body intelligence to robots — Google DeepMind1 savers
• Cars often keep people poor. Waymo could help.1 savers
• Why were Covid vaccine trials so fast? - by Saloni Dattani1 savers
• Making a Billion Intelligent Machines1 savers
• Gallery Item Display1 savers
• Rethinking High-School Science Fairs—Asterisk10 savers
• fable1 savers
• RoboTTT: Context Scaling for Robot Policies1 savers
• Base Models Know How to Reason, Thinking Models Learn When1 savers
• [2607.14111] Introspection Fine-Tuning (IFT): Training Small LLMs to Introspect2 savers
• Early Data Exposure Improves Robustness to Subsequent Fine-Tuning1 savers
• @ Cutting the Skip: Training Residual-Free Transformers1 savers
• Missing openness - by Lydia Nottingham - pronotre1 savers
• External commentary for global workspace paper -- final final1 savers
• Jacobian Lens – Qwen3.6-27B ｜ Neuronpedia1 savers
• A global workspace in language models \ Anthropic25 savers
• The world reveals itself to those who travel by foot9 savers
• godel-letter.pdf2 savers
• Relationships are coevolutionary loops - by Henrik Karlsson39 savers
• Verbalizable Representations Form a Global Workspace in Language Models26 savers
• Dario Amodei — Machines of Loving Grace13 savers
• How To Become A Mechanistic Interpretability Researcher — AI Alignment Forum10 savers
• interpreting GPT: the logit lens — LessWrong9 savers
• A Pragmatic Vision for Interpretability — AI Alignment Forum8 savers
• AI Is Slowing Down5 savers
• “God has helped us, and so will AI”: How the Terrorist Group Boko Haram Uses Frontier AI — CASP3 savers
• How Can Interpretability Researchers Help AGI Go Well? — AI Alignment Forum2 savers