The 10x AGI Information World
10x here means 10 times. AGI stands for “Artificial General Intelligence.” So, I use three adjective words to describe the future world. And 16 million people in the world might be aware that the future will be the AGI world. Midjourney has around that user account as of July 2023. And ChatGPT had 100 million users in January 2023. So if we choose a number to define the count of serious AI tool lovers, we probably want to choose 16 million. ChatGPT is a text conversation tool, it is simple and direct. And Midjourney is a tool to help generate images by AI algorithms. That’s the tool serious lovers will try.
10x means that it is really fast, faster than the normal 10 times. And it has no question to say this world is an Information World as the number of Internet users in the world reached 5.160 billion people.
The people who work hard and think hard will be more optimistic about the future. One reason is that you have a feeling to impact the world by using your brain. Another reason is that as you learn more, you know that human society is just beginning its technology chapter. The third reason is that you have the boldness to suspect the opinions of everyone. So the future seems interesting if you think you can prove you are right about something few in the world know.
And I am one of these people after 28 years of learning in this world. So here I want to share the future world I think.
How We Work
I think that Jira and Confluence are outdated in the 10x AGI information world. They should totally integrate the power of AGI and GPT. We may say a lot of Jira issues shouldn’t be created by product managers. We should let users use AI tools to create it. And then the issues created by the users with the help of AI tools will automatically calculate out the assignee. The AI-powered Jira will use the description of the issue, and all internal information (like documents and code) to know this issue should be assigned to who.
When we mention the assignee, the one is an AI-powered developer or tester, or an AI-powered full-stack worker. We may not need to differentiate roles like product manager, developer or tester anymore. They are all AI-powered full-stack workers.
And when we think about it, we should think, in that world, why we still need users to report bugs or issues to us. Why we don’t create a lot of AI-powered test users to do so? The perfect app is hard to make because it has a lot of scenarios to use and a lot of users with different backgrounds. They need the app to support the language they like. They like it just like ChatGPT. A very basic and simple conversation interface with any language it support. The main disadvantage of it is that it needs users to type text to input. However, it is inconvenient to input in mobile phones. And when users input from mobile phones, a lot of them like to separate their messages for the fast speed and convenience. It is easier to send three short phrases instead of one complete message. The input message box is also not supporting the long text well.
I just checked the ChatGPT iOS App. Though the input box will expand like below, the user habit is hard to change to send short phrases instead of whole sentences.
And however, I just found that ChatGPT supports voice input here. So it is easier to input the long text. And we can close this discussion as they already know it.
Let’s come back to our discussion about how we work in the future. Suppose we have PayGPT work as a payment app with AI power and ChatGPT interfaces. How should we iterate the products? What’s the gap between the ChatPT and the PayGPT? Underlying these data-intensive applications are databases. So we should ask what’s the gap between text and database?
How about we just export the database to text, train it to the language model, and then query it? How do we deal with the user rights problem? One user should be only able to access his data, part of his friends’ data, and the public data of the app. Do we need to train a language model for every user? Or that we have pre-trained large language model, and then we input the user context to them, and then we begin to query.
Let’s come back to how the users create issues. Let’s think about the case that user A uses PayGPT to pay some money to user B. User A uses voice input to tell PayGPT to do so. And after payment, user B will receive some notification from PayGPT. User B opens the app and sees the PayGPT bot tells him that user A has paid some money to him.
And if user B couldn’t receive money, he wants to report this to the PayGPT company. So he tells the PayGPT to check this. So the issue is created by some conversation messages of user A and those of user B.
The issue is delegated to one AI-powered full-stack worker in the PayGPT company. How the Ai-powered Jira knows who to assign? It will analyze the conversations, including the problem report words from user B. Though in AI-era, every knowledge worker will be a full-stack worker. However, the levels of full-stack ability are different. The AI-powered Jira knows it as Jira knows all the information in the company and its author.
And so Jira and Confluence shouldn’t be separated anymore. I thought that a lot of information is duplicated in Jira and Confluence in the companies. We should have single source truth and information. Otherwise, it will lead to inconsistency when we just update one copy of information instead all.
And the AI-powered full-stack worker can handle all information in PayGPT company, document, code, and test cases. He may not need to collaborate anyone to deliver a feature or even the whole app. As a lot of code is just created by Copilot alike tool. And the document is for the collaboration between the coders and the non-coders. And now in the AI era, everyone is a coder. So there will be few documents. The document is always not accurate as the code.
In such automatical work routines, we don’t need the leads anymore. As it is clear how much information one is created.
Let’s give such a platform a new, say Work Center. Work Center = Jira + Confluence + GitHub + Teams. We may not need Microsoft Teams anymore. The AI-powered full-stack worker has all information and can understand all information with the help of WorkGPT. And he also can create all information alike the existing ones. Why he should still discuss and collaborate with others?
And now people can ask questions to ChatGPT when driving. So why we couldn’t work that way? If we could understand neural networks by asking questions to ChatGPT during the driving, why we couldn’t write the code of neural networks in that way? The voice human begins to say is easy to convert the text words or code. So everyone can have fun traveling all over the world, and just do some work in driving, in the transport tools.
And of course, everyone can use their preferred language to do most of the work as long as they can create the code.
And the salary will be automatically calculated by the information generated by the AI tools with the supervision of the AI-powered full-stack worker. We don’t need titles, or corporate structures anymore as there is no need for collaboration.
How We Enjoy
The 10x AGI information world will be quite funny. Everything is in the cloud. When we move to a new home, we can set it up in one week. The AI-powered robot will help us pack up things and settle down things in the appropriate places. Basically, we don’t need to do any labor work. The transformer behind ChatGPT may not be very accurate in its position, but we can accept that the furniture in the home is not so accurate in its position.
Not only the temperature and humidity can be perfect in any place in our home. The TVOC, formaldehyde concentration needs to be as less as possible in any place in our home too. Every hardware and appliance in the home needs to be connected to the cloud and always report its temperature, humidity, and air quality around it.
One direction to imagine is to think we can operate our appliances anywhere. And they should work for our interests as much as possible.
This is the scenario of home. Other places are the same too. We should be able to control all aspects in spaces 1-meter distance from ourselves. We should have one small room in the office. And we can control everything just like our home. For example, I am quite frustrated several times to find that I didn’t bring my outlets to give an eye to resist the cold air in the office.
And in public places, it seems we are hard to control everything just like our AGI home. But that is not true if we think about it. We need a four-weels upper-half-transparent room. We want it upper-half-transparent that we just need to look outside through the upper half of our room to know the surrounding of us. And
It helps us to walk and go to any place. The staircase may be hard for our four-weel room. We may need to change the stairs to suit that. And so we may need a bigger car to bring our movable room like the below one.
In my imagination, the ideal movable room should be small enough to just hold us with a little extra space. Because there are still many buildings that are not prepared well for movable rooms. They are just built for movable people.
In this case, people will like to go outside in any season, any weather. The movable room is surely a talented computer too. It will report temperature, humidity, and air quality around it. And it knows how to go anywhere it can go. You just need to use your mouth to tell it.
So if everyone has a movable room and can bring or rent it anywhere, they will be much like to explore the world. The car is like a movable room, actually. However, we couldn’t use the car to go to any internal place of the buildings. I am pretty sure this is a very good idea. The world should completely redesign the buildings for movable small rooms. People should be able to control everything surrounding them.
So the doors will need to be larger, and the elevators need to be larger. And we want our small movable rooms to be portable. So it is like the below portable clothes dryer.
If the upper half of the dryer is transparent, that’s better. So it is portable as most of it is made with plastic. The shape is flexible. We can easily pack it up to put in our car.
So we have AI-powered electric cars and AI-powered electric movable rooms, now we can travel the world, and hang out with family or friends better.
The End
Here we cover some aspects of how we work and how we enjoy the 10x AGI information world. That world is about 10x data, 10x code and 10x processing speed. After writing down these ideas I had days before and thinking out some new ideas, I now feel more optimistic about the future. We still have not yet tried many possible things to let our human begins better live in the world. I hope this can give humans some ideas to imagine and discuss. When we look back the people who lived 1000 years ago seem so suffering as they don’t have electricity, computers and machines. And now technology is developing fast, it is surely the people who live after 100 years will think we are so suffering. Period.