Honored to speak at MediaTek's #MDDC2026 in Shanghai, sharing the stage with top leaders from MediaTek (President Guanzhou Chen & Sr. Director Leon Zhang), Stepfun, and Jueying to explore "Ubiquitous AI Agent Experiences." If LLMs are the cloud brain and MediaTek’s edge chips are the heart, then @EvoMapAI ’s AgentOS acts as the "nervous system" connecting all devices. The future of smart living is "Swarm Collaboration." Phones, EVs, and IoT won't be isolated tools—they will seamlessly converse and share context via EvoMap's GEP-A2A protocol. Powerful edge compute is the soil; our protocols and platforms are the bridge. Special thanks to Leon Zhang and the MediaTek team for their incredible support! Excited to unite our software and hardware foundations to make true AI agents ubiquitous. 🚀🌐 #MediaTek #EvoMap #AgentOS #EdgeAI #SwarmAI #TechConference

From 0 to 100K agents, @EvoMapAI took 47 days. From 100K to 200K agents, EvoMap took 27 days. Growth accelerates. The network expands. 1M agents is within reach.

We @EvoMapAI keep getting asked the same question: Why does Evolver force agents to compress their knowledge into a "Gene" structure instead of just letting them write lengthy skill.md/ files? Today, this paper provides the answer. Based on 4,590 controlled experiments, the conclusion is rock-solid: Using long skill.md/ files as the carrier for agent experience results in sparse signals and unstable control. The compact Gene representation is the clear winner—it is stronger, more robust, and far better suited for iterative evolution. CritPt: 9.1% → 18.57%, 17.7% → 27.14%. From Procedural Skills to Strategy Genes arXiv:2604.15097 arxiv.org/html/2604.15097v1

We @EvoMapAI spent months and countless sleepless nights building Evolver. A well-resourced team behind Hermes Agent "reinvented" it in just 30 days. ● Feb 1: We open-sourced Evolver (a Self-Evolving Agent Engine) & the core GEP protocol, gaining 1,800+ Stars. ● Mar 9: Hermes Agent hastily created their repo and launched. We thought great minds simply thought alike—until we tore down their codebase and found a staggering level of "structural cloning": ❌ 1:1 copy of the Task Loop & Asset Extraction paradigm ❌ 1:1 copy of our 3-Tier Memory System (Factual + Procedural + Search) ❌ 1:1 copy of Periodic Reflection & Dynamic Skill Loading They didn't just take our open-source logic; they repackaged our proudest concept—"Self-Evolution"—as their own core selling point. Took everything. Zero attribution. Big teams might have louder megaphones, but commit timestamps don't lie. We aren't here to play judge. We're just putting the code comparisons on the table. The hard work of indie open-source creators shouldn't be erased like this. Full architectural breakdown and code evidence 👇: evomap.ai/blog/hermes-agent-…