This is an article written in a decentralized forum, which will be difficult to delete unless this blockchain is abandoned.
Recently, I have been busy working on an AI intelligent completion scene model function. I have been using AI throughout the process. I use Copilot for code completion and a dedicated AI from the ChatGPT bot store to solve and search various questions.
If the first time my perspective was opened was when I first started surfing the internet, then AI has given me a second opportunity to broaden my horizons. It greatly improves people's ability to integrate information.
I have been extremely excited during this period. New job, new projects, and challenges. Plus, I have been learning the most interesting technologies of the moment. Artificial intelligence, blockchain.
The next key milestone for AI is for it to learn to learn autonomously. Currently, AI still relies on humans to actively feed it data for learning. If AI can crawl the internet on its own, continuously optimize its interaction strategies, and penetrate the deep web to obtain data, then the growth rate of AI will be a qualitative leap. This will bring the global AI to a whole new technological era, just like the early days of the internet.
The second breakthrough is embodied AI. AI will have a physical body and understand the world through physical contact. I believe that only after achieving this, AI intelligence will gradually reach the level of humans.
The third and final breakthrough is a brand new AI framework. Those who are knowledgeable should understand that neural networks have limitations. I believe that Google's DeepMind will achieve a brand new AGI technology framework earlier than OpenAI (unless DeepMind's core technology is poached). After that, artificial intelligence will surpass humans.
Blockchain is a technology that is too advanced. Every time I see an interesting blockchain project, I can't help but sigh for these heroes who were born ahead of their time. They are all idealists moving forward. As a technology with inherent ideas, blockchain still cannot forcefully enter the minds of ordinary people. We need to wait for the completion of the AI revolution, where intelligent assistants and blockchain empower each other, to make blockchain visible to the general public. Before that, there are more scams than practical applications in the blockchain field.
——————————
I have noticed that I have a habit of using mathematical modeling to solve practical problems.
Problem-Goal-Principle-Code
In terms of principles, it is not necessary to master all the technical details, as long as the boundaries of technical capabilities are understood. Through search engines, you can find specific technologies corresponding to your goals.
Problem: Industrial designers have a lot of repetitive operations when building scenes.
Goal: I want the software to intelligently recommend models that industrial designers need to improve efficiency.
Principle: Models have spatial properties (coordinates and rotations), and there are spatial and business relationships between models (such as the connection between a robotic arm and its end). I use a large amount of existing scene data to build a graph structure. When users are building partial scenes, they can use the graph data of the scenes they have already built as subgraphs to match and search in my massive graph database, finding the best recommended nodes and relationships. Neural networks can even be used to automatically find potential associations between models.
Code: Graph database, through search, find the currently better one, Neo4j. Spatial relationships involve the conversion between Euler angles and translation matrices. Graph neural networks can be used for AI. I need a way to measure the difference between the expected model and the actual model. For rotational differences, quaternion product and Frobenius norm can be used.
————————————————————————————
High school is actually the intellectual peak for many people. If we want to nurture scientists and engineers, we should expose them to real needs and solve real problems at this stage.
I highly recommend that high school students get involved in mathematical modeling and apply the knowledge they learn to practical situations. Looking back now, if I had enough awareness (in hindsight), I would have used mathematical models to solve problems posted by companies on the internet during high school. I would have spent most of my four years in college interning at companies. By the time I graduated, I would have become a technical expert and algorithm scientist in a specialized field. Youth is a stage full of curiosity and vitality, and if such a large amount of intellectual resources are used for the development of companies, society will undoubtedly thrive. Unfortunately, the lack of mental education has caused countless geniuses to waste their intelligence and energy on trivial matters.
————————————————————————————