- [[Explosion of AI generated mathematics with counterexamples to conjectures and the future of AI mathematics with Artificial Grothendieck Intelligence]] - [[Can we solve science of LLMs, can we solve science of deep learning, can we solve science of AI, can we solve science of information processing systems in general]] - [[Current state and future of recursive AI self-improvement research, in how strong form is it real]] - [[AI tribes from The Master Algorithm and what camps would I add]] - [[How are current AI systems mostly created, and who is pursuing alternatives]] - [[Mechanistic interpretability hypotheses for Measuring Reward-Seeking by Instilling Contrastive Beliefs and additional comments]] - [[Is human + AI worse than AI alone, in chess generally, true, or in other words, is cyborg aka centaur worse than pure AI]] - [[GLM-5.2 moves from GRPO (Group Relative Policy Optimization) to PPO (Proximal Policy Optimization)]] - [[Does reinforcement learning with verifiable rewards incentivize novel reasoning]] - [[Falsifying scaling hypothesis in general is impossible]] - [[Ladder of evidence in mechanistic interpretability of neural networks]] - [[Analogies between the science of AI and other sciences - physics of deep learning, biology of deep learning, etc.]] - [[Comments on the J-Space paper from Anthropic]] - [[The real AI benchmark, actual open problems solved in formal and empirical sciences]] - [[Mapping the space of AI research]] - [[Comment on Neural networks are already clean by Nick Cammarata from GoodFire]] - [[Neurosymbolic learning might work and be more efficient than neural learning alone]] - [[AI capabilities are harder to predict than a lot of people think because we still don't know a lot about AI]] - [[Bitter lesson in the limit]] - [[Longesttermism]] - [[How AGI pilled am I]] - [[Defining AGI and intelligence]] - [[How to treat the AI consciousness topic with a more proper philosophy and science]] - [[The intersection of AI and physics]]