Another 5 years have passed. Here’s a summary of what happened after I wrote in my last post “my mission statement” as “Building True AI with NLU and Common Sense”:

  1. I ported all my iOS and Android AI apps built with TensorFlow Mobile to PyTorch Mobile around 2021 - both platforms are long gone now but I had fun and learned a bunch on how to run all kinds AI models - CV, NLP, audio, video - on mobile devices.

  2. I registered a domain name polygbot.com (for ployglot chatbot) for my chatbot simulating a 5-year-old’s NLU & NLG capabilities, worked on the chatbot using hybrid AI methods with rule-based and DL. Then I unregistered the domain name after the disruptive ChatGPT was launched in Nov 2022, literaly making my mission of 2020 “Building True AI with NLU and Common Sense” a shocking reality, albeit its imperfection.

  3. I started working at Meta as a GenAI engineer in 2023 building Llama powered GenAI apps using prompt engineering, RAG, fine-tuning, and agents, the code and content of all 3 Llama courses on Deeplearning.ai, and iOS apps running on-device Llama and Llava models and calling OpenAI Realtime API with WebSocket and WebRTC for audio chat.

So many crazy things have happened since Nov 2022. So many more crazier things will happen in the coming years. I once asked AI about a French phrase “Y a pas de petits rôles, y a que des petits acteurs.”, one favorite line in my favorite French TV show Call My Agent, with the English translation “There are no small roles, only small actors.” and I asked how to have a balance of working on the most important, challenging and ambitious projects versus being gladly working on even small but needed projects, and got the following tip:

*Know Your North Star. Before you can balance anything, you need clarity on: What are your most important long-term goals? What kind of impact do you want to have? What problems are worth your “life’s work”?

Doing small things with excellence is the hallmark of someone capable of doing big things with grace.*

and the TL;DR:

Big Projects Small Tasks
Move the needle Train attention, humility, consistency
Require focus, vision, deep work Can build trust, team, or momentum
Must be prioritized Should be scoped, delegated, reframed
Define legacy Shape character

First time I heard of someone saying one's work path ideally should be job, career, and mission was about 10 years ago. I always remember Godfather's simple yet powerful saying "I'll reason with him" and have been fasicinated with all kinds reasoning: deductive, inductive, abductive, commonsense, causal, etc (One book I love reading is Logical Chess by Irving Chernev and one graduate AI course I took was Non-monotonic Reasoning); RL or learning through trial and error is my favorite way of learning (I bought a copy of the classical RL: An Intro book in 2005 while working at Baidu before the modern AI took off).

Let me be on a new mission for the next 5-8 years. Let me focus relentlessly on open, high-impact problems in AI, starting with reasoning and modern RL for LLMs. Let me build innovative AI products with awesome UX powered by that focus.