Thoughts after drinks

There is no alcohol, only milks, by the way, if you are expecting that. Because I worked in the gym tonight. Why can’t I find genuinely interesting drinking buddies? Or it is a bless?

AdaReP : Plan ahead and don’t change the plan unless you really have to. Long-term plan matters. This reminds me of my AI-Expert Collab workflow design, and I think maybe my workflow should not be too dynamic. It should have some buffer for rigorousness.

AI FICTION IN THE WILD : I am definitely a story cycler, ceaselessly looking for new topics to try. But I do understand that being a phd is quite like being a Infinite Story Demander, so I better calm myself down and genuinely focus on one topic.

Einstein World Models : I think this essay is kind cool because it employees visual thinking when we all know that sometimes visual thinking hold text thinking back. Using world model to help the model so that the latter do not need to rely on sole words when dealing with complex physic problems is really elegant.

Human–LLM Collaboration Is Transforming Complexity Metrics in Scientific Texts : I actually really like this essay even though it really cannot help me at all.

Large Language Models Do Not Always Need Readable Language : Way to save money indeed

Narration-of-Thought: Inference-Time Scaffolding for Defeasible Ethical Reasoning in Large Language Models : Focus on only one interest party and tend to be overconfident

ProfiLLM: Utility-Aligned Agentic User Profiling for Industrial Ride-Hailing Dispatch: to solve the long-tail distribution, use clustering.

Tapered Language Models: Wider-early

Now I decide that no matter what school and what department is most likely to accept me at this point, my current focus will be human-ai collab framework!