​Towards Personal AGI

​Today’s LLMs are powerful, but they still resemble the mainframes of early computing: centralized, expensive, shared by many users, and disconnected from the context of any one individual.

​The PC made computing personal, persistent, programmable, and user-owned. AI has not had its PC moment yet.

​Towards Personal AGI is a private salon for researchers, builders, founders, and investors exploring what comes after the centralized LLM era—and what it will take to build truly personal AI.

​The Core Question

​Mainframe computing → Personal computing
Centralized LLMs → ?

​Will the next paradigm be a persistent agent, a personal world model, an intelligent operating layer, a new device, or something we have not yet named?

​Our guests are from:

​University: Stanford x4, UC Berkeley x2, MIT x3, Harvard x1, Princeton x1, Caltech x1, Columbia x1, Oxford x1, Yale x1

​Frontier Labs: Anthropic x2, OpenAI x2, Deepmind x3, Meta MSL x4, Amazon AGI x3, Nvidia x1, Tesla x3, Apple x2, Microsoft Superintelligence x1, Together AI x1, ByteDance x1, AMD x1, Samsung x1

​【Guests participate in a personal capacity; institutions shown for identification only.】

​Selected Talk:

• ​Special Guest: Travis Oliphant, the founding figure in the Python scientific computing ecosystem

  • ​creator of NumPy, co-creator of SciPy and Numba, and co-founder of Anaconda.

• ​Introducing AItonomy Foundation: we are building open infrastructure to accelerate the loop between science and AI.

  • ​Speaker: Hejia Geng, AItonomy Foundation

• ​Digital Twins & Ontology

  • ​Speaker: Zeming Liang & Yuting Deng from Columbia University Data Agent Process Lab

• ​MatrAIx — Simulating the World with 8.3 Billion Persona Agents.

  • ​Speaker: Dr. Yuexing Hao & Dr. Xiaomin Li from MIT & Microsoft & Harvard.

• ​Use AI agents to build your personal AI

  • ​Speaker: Ruiyi Zhang from UC San Diego & AIBuildAI Inc.

​What We’ll Discuss

​The evening will explore:

• ​Memory, retrieval, knowledge graphs, and personal world models

• ​Agents that understand changing relationships, preferences, and intentions

• ​Privacy, permissions, reversibility, and human control

• ​Reliable action across fragmented applications

• ​Evaluation for systems whose correct behavior differs by person

• ​New companies and infrastructure emerging around the personal AI stack

​The Format

​This is not a product launch, demo day, or traditional panel. It is a private, high-trust conversation with a short opening thesis, guided discussion, dinner, and drinks. Attendance is limited and approval-based. The exact Palo Alto location will be shared privately with confirmed guests.

​About Aitonomy Foundation: AItonomy Foundation is a nonprofit organizationworking with the global scientific community,building open research and infrastructurefor AI4Science and Science4AI.