Greetings from Sapient Intelligence! Happy New Year! LFG AGI 2025!

DeepSeek just did something very cruel and killed the rabbit... now when people talk about R1, they don't mean the orange box anymore.

Hierarchical Recurrent Models towards AGI Excited to have Sapient Intelligence’s Yue Wu share insights on our Sapient-H Architecture alongside NVIDIA in Wuhan, China. Stay tuned—something interesting is brewing!

We are sharing the code openly so everyone can build on it. ❤️

Our co-founder William Chen is going to share more about the open-sourced Hierarchical Reasoning Model (HRM) at #FortuneAISingapore FORTUNE tomorrow, under the panel theme "Beyond Human: AGI And The Future We’re Building"! We are excited about the practical path towards

At the “Beyond Human: AGI And The Future We’re Building” Town Hall at Fortune Brainstorm AI 2025 in Singapore, William outlined our thoughts for AGI: “We’re exploring new architectures to push the boundaries - to make AI think like a human, not just model probabilities. True

Bigger ≠ Better. The GPT-5 rollout reminded everyone that raw scale isn’t a strategy. Real value now lives in agent reliability, not leaderboard one-shots. Our stance: optimize for closed-loop task success (plans → tools → checks → handoff), not just next-token accuracy.

We are on the ARC Prize leaderboard now - a good starting point! Meanwhile, we are accelerating the iteration and application of the HRM model; stay tuned!

Google is expanding its AI mode to 180 countries, offering users a personalized restaurant reservation service. By providing specific details such as date, time, and number of people, AI can precisely filter and recommend restaurants that meet the criteria, greatly enhancing the