I Hate Transformers Reading Club

​An informal reading club for people interested in what happens outside the standard transformer + more data + more compute paradigm.

​Symbolic reasoning. Knowledge graphs. Explicit memory. First-order logic. Neuro-symbolic systems. Other weird architectures from 1985 that suddenly look interesting again.

​The goal isn’t to prove transformers are dead. The goal is to spend an evening asking a more interesting question:

​If we were designing AI from scratch today, would we really build it as a single trillion-parameter transformer?

​The format is simple: bring one paper from the domain.

​Everyone gets 5–10 minutes to explain:

• ​what the paper is trying to do,

• ​how the architecture works,

• ​what you found interesting.

​Then we talk.

​No slides. No formal presentations.

​Some examples of the territory

• ​CYC: Using Common Sense Knowledge to Overcome Brittleness and Knowledge Acquisition Bottlenecks

• ​DeepPath: A Reinforcement Learning Method for Knowledge Graph Reasoning

• ​Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning

• ​Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph

• ​Reasoning with Trees: Faithful Question Answering over Knowledge Graph

• ​ReKG-MCTS: Reinforcing LLM Reasoning on Knowledge Graphs via Training-Free Monte Carlo Tree Search

• ​LLM-Guided Monte Carlo Tree Search over Knowledge Graphs

​Anything adjacent is fair game.