Create revision_manifesto.md

Author: likefallwind

votes/revision_manifesto.md

@@ -0,0 +1,23 @@
+# A vote for the revision to the DeepModeling Manifesto
+
+## Proposal
+
+On December 12, 2017, the DeepModeling community launched its first open-source project, DeePMD-kit—combining machine learning with physical modeling from the perspective of molecular simulations. On May 6, 2021, DeepModeling released its community manifesto, declaring its commitment to “integrating physical models across all scales with machine learning methods, and leveraging cutting-edge computational tools to tackle the most challenging scientific and technological problems facing human society.” Since then, we have witnessed a growing number of projects thrive within the DeepModeling community, putting into practice the vision of collaborative development through open-source efforts.
+
+Meanwhile, AI for Science is undergoing a profound transformation. Large multimodal AI models are becoming capable of rapidly reading vast amounts of scientific literature and extracting key scientific insights, while also integrating knowledge from papers with experimental data and simulation results in a coherent and meaningful way. AI agents are beginning to reshape the entire research pipeline — from hypothesis generation to experimental validation. These breakthroughs are opening up new pathways for scientific discovery, enabling the construction of intelligent, closed-loop systems that connect literature comprehension, computational modeling, and experimental verification—dramatically enhancing research efficiency.
+
+In response to this transformation, core developers of the DeepModeling community have begun to explore open-source practices beyond the community’s core themes. Representative examples include Uni-Mol (a foundation model for specific scientific modalities), SciAssess (a framework for evaluating large models in scientific literature analysis), and Uni-Lab-OS (an intelligent operating system for laboratory automation). These efforts have laid the foundation for DeepModeling’s expansion in two key directions:
+1. Extending the boundaries of scientific tools: While continuing to build on its strengths in machine learning and multiscale physical modeling, the community is developing a new generation of AI-native tools that integrate literature, computation, experimentation, and modeling into a unified research ecosystem.
+2. Deepening open-source collaboration: By improving tool interfaces and evaluation frameworks, the community aims to build a more open system for tool certification, contribution recognition, and knowledge sharing.
+
+Against the backdrop of these reflections and the evolving landscape of the field, we have made a brief revision to the DeepModeling Manifesto—most notably, broadening our perspective from “scientific computing” to encompass the entirety of “scientific research.” We remain firmly convinced that the future of science is built upon the foundation of openness and open source. The DeepModeling community will continue to uphold the spirit of openness, pragmatism, and collaboration, working hand in hand with developers around the world to harness the power of open source and drive the next scientific revolution.
+
+## Deadline
+
+If 2/3 majority is reached before the voting deadline, the proposal will be passed directly.
+
+## Scope
+
+TOC MEMBERS
+
+## Result