jagged raven
Hey everyone,
We’re a group of students and AI enthusiasts working on a research initiative exploring a promising new architecture we’re calling the Continuous Thought Machine (CTM) — and we’re looking for collaborators who are excited about pushing the boundaries of general AI.
CTMs are a recent conceptual development we believe could surpass existing industry standards like Vision-Language Models and even large transformers on generalized tasks. Our hypothesis is that CTMs, due to their internal synchronization and memory tracking (pre- and post-activation histories), may exhibit emergent capabilities as they scale — capabilities that resemble short- and long-term memory in biological systems.
🧠 What we’re doing:
• Benchmarking CTM performance and scalability against models like GPT and diffusion-based transformers
• Using adversarial reinforcement learning in continuous control environments (think: air hockey simulations)
• Drawing methodological inspiration from the Atari generalization framework
• Stretch Goal: Investigate memory-augmented CTMs that can selectively store and retrieve information — beyond brute-force RAG methods
We’re currently drafting proposals, assembling a strong team, and building the research infrastructure. If you’re interested in deep learning, reinforcement learning, memory-augmented networks, or just want to help push experimental AI forward, we’d love to hear from you.
Let’s build something wild.
restive light
solemn belfry
wind jewel
lilac juniper