End-to-end long-horizon reasoning with one Transformer model

Re:End-to-end long-horizon reasoning with one Transformer model by hyluo on Wednesday July 23, 2025 @09:15AM (#44658892)

- We build the Thread Inference Model (TIM) based on the transformer architecture, and its dedicated runtime TIMRUN.

- TIM + TIMRUN = Intelligent workflow generation, context engineering, and multi-hop tool use happens at the runtime level

- TIM + TIMRUN supports virtually unlimited reasoning enabled by context pruning, significantly improves the efficiency for long-horizon reasoning tasks

- Inference API is live at https://subconscious.dev/

- More details: https://github.com/subconscious-systems/TIMRUN

[ Parent | on HN ]

Re:End-to-end long-horizon reasoning with one Transformer model by datapalo on Wednesday July 23, 2025 @10:58AM (#44659953)

I’m more than ready to stop my workflow being derailed by opening new chats—looking forward to LLM assistants that actually remember our conversation.

[ Parent | on HN ]

Re:End-to-end long-horizon reasoning with one Transformer model by kevin8704 on Wednesday July 23, 2025 @10:05AM (#44659379)

Really cool work! I saw the “Selective Working Memory” section, are there hooks to swap in a custom retrieval store or memory layer, or is that all internal to TIM?

Thanks for sharing!

[ Parent | on HN ]