A 10-year-old interviews a tech expert about AI and database management

Entering the tech sector can be extremely intimidating if you come from a non-technical background.

Many ideas are so jargon-filled that they can be difficult to grasp. Take database management for instance - there are different types of search languages, systems, and concepts such as sharding, among other topics.

Let’s do away with all that - can a tech expert explain all these things in plain language? We got ten-year-old Adam Liao to put Edward Huang - co-founder and CTO of PingCAP, which provides database management solutions - to the test.

The following interview has been edited for clarity and brevity.

Getting acquainted with databases

Liao: What is a database?

Huang: Databases are super important, and they’re like containers for data.

Liao: So it’s like a box for data?

Huang: Yes! You could also imagine database software as being like a library. You can put your books in a library and look for specific ones once you’ve stored them there. Just replace the books with your data, and that’s basically what database software is.

Cloud, meet database

Huang: Lots of data is stored on the internet in what’s known as the cloud. Do you know what that is?

Liao: Based on my understanding, the cloud is equal to the internet world.

Huang: That’s a good answer! The cloud is just a remote link through the internet to a data center.

Liao: So what’s the relationship between the cloud and databases?

Huang: Sometimes, the data we want to store exceeds the capacity of a single computer. That’s when we need things like distributed databases, and these are normally sitting on the cloud.

Bringing real life into the picture

Liao: I love using TikTok and playing video games! How do cloud databases and AI play a big role in making them work?

Huang: Normally, when people use TikTok, their data is stored in a distributed database. This data could be things like their uploaded videos, as well as others like their profile, username, and password. For example, when you open your TikTok app, the software will go to the library and say: “Hey, can I get videos that Adam may be interested in?

The library officer may reply: “These are some videos that Adam may be interested in” and send those videos to your phone.

Liao: What about AI?

Huang: Imagine how our brain works. We read and learn from things like books. Well, a computer can constantly read books, texts, and images from the internet at crazy speeds and generate its own experience and understanding through all that data. All this information is contributed by human beings just like us, as well as your mom and dad.

TikTok is built on what’s known as machine learning technology, but usually the AI is meant to simply introduce or recommend videos.

However, there’s also new tech like ChatGPT-3, which is a lot more interactive, and you can use normal conversational language to ask questions to the chatbot. It’s got so much knowledge from the internet, just like a master that knows everything.

A peek into the future

Liao: So do you have any cool projects with databases and AI working together?

Huang: That’s a good question! At PingCAP, we’re building something along those lines. Remember how I said that a database is like a library?

Another special thing about this library is the way you access your books. We have to use a special language that’s not like Chinese or English. This language is called SQL, or structured query language. Liao: I know this one!

Huang: You know it, since you’re a programmer too!

[Editor’s note: Liao takes coding classes and has participated in hackathons with Huang before.]

Sometimes, SQL isn’t very easy to use especially when you have a very complicated query. With regular languages like English or Chinese, you can just ask the library officer: “Hey, I want this book.”

But using SQL is kind of like doing math and can be a little complicated.

At PingCAP, we’ve created a software called Chat2Query, which will be a bridge between database software like TiDB and big language models like ChatGPT-3.

The idea is very simple. For example, if you have a lot of data stored in your database but don’t want to write in SQL, you can just ask Chat2Query about the information you need. The software is just like a translator – it will turn your questions in natural language into SQL.

This will be really helpful especially to someone that doesn’t know SQL but still wants to get insights from data.

Playing a human part

Liao: Since AI will be able to do so many things, how can people still contribute and make a better world?

Huang: AI models like ChatGPT-3 are still software written by engineers like us. ChatGPT-3 is just a good tool to help us write programs faster and make us more effective. You’ve done programming before and you like it, right?

The fun part about programming won’t change whether there’s ChatGPT-3 or not. Just keep focusing on what you like and your passion.

AI getting better is just like the calculator. After it was invented, we didn’t need to manually do math anymore and we could perform a lot more complicated computations. At your age, you still have to learn a lot of math and computation by hand, but it’s really just training for your mind.

I think in the future, you won’t need to memorize a lot of things. You can ask an AI model like ChatGPT-3 to find out what you want to know, and that will change the way you learn. When you grow up, sometimes the experience or hands-on practice may be more important than just committing information to memory.