InTowards AIbyXinjie Li·Oct 26When AI Solves the Problem, Does Understanding Still Matter?AI is becoming a scientific discoverer. But what happens when our ability to produce knowledge begins to exceed our ability to understand…A response icon1A response icon1
InData Science CollectivebyXinjie Li·Sep 7What Is a World Model, Really?After the LLM revolution, world models are often described as the next big step toward AGI. But what exactly counts as one? Is a Deep…A response icon2A response icon2
InArtificial Intelligence in Plain EnglishbyXinjie Li·Apr 17Beyond Parameter Scaling: Where DeepSeek and Qwen Are Finding the Next LLM GainsWhat conditional memory and gated attention reveal about the limits of brute-force LLM growth
InLife with TechbyXinjie Li·Mar 31OpenClaw Made AI Agents Feel Real. It Also Made Their Limits Harder to IgnoreMore than just another agent app, OpenClaw showed what happens when AI moves from chat into inboxes, calendars, calls, and real workflows.
InMITB For AllbyXinjie Li·Mar 7Building a Domain-Specific Chinese Tourism Assistant: From Data Scarcity to Fine-TuningAn applied case study in data design, controlled augmentation, and behavioral alignment for tourism AI, with data, code and notebooks…A response icon1A response icon1
InMITB For AllbyXinjie Li·Jan 26Privacy-Compliant AI: Theoretical Foundation of Federated LearningThis is the first article in a two-part series on Federated Learning systems, co-authored by Xinjie Li and Emma Chew, graduates of the…