At the Silicon Valley GPT Party 2.0 event on October 7-8, an event dedicated to artificial intelligence, more than 300 people gathered at Plug and Play to meet with leading experts, entrepreneurs, and investors, discuss the latest trends in artificial intelligence, and gain practical knowledge.
At the event, Andrew Gree talked about an innovative product for managing artificial employees, shared the process of creating this new technology, and demonstrated its features.
“In the near future, we will coexist with virtual employees who will form entire organizations. We will interact with them, take on roles in teams. Recently, we patented a technology for managing artificial employees, and now we are ready to share our developments.
Globally, fundamental changes in thinking are taking place right now. Before the invention of AI, we communicated in a human-application mode, interacted with several applications or robots. Now we see interactions between applications and applications, and robots and robots, something we have never seen before. We are witnessing a history where we can configure an entire organization of AI assistants who can work together to solve tasks.
Our product began when I explained to Chat GPT 4 that we have smart assistants based on Chat GPT 3.5 in IRC, and I asked him to create an organization. He outlined an organization that can be seen on the diagram. Each person has two assistants: Topic Analysis Elf and Personal Assistant Elf, and there is also a Central Table Coordinator. We have task descriptions for each artificial employee, stored as text. This file is passed to the meta-organization, which creates channels based on it, sends tasks to the Elves, and organizes interactions. Job descriptions were also created through Chat GPT 4, and you can change and adjust the work of employees there. All the Elves communicate with each other, entering chat rooms where they can communicate directly with each other or in the general channel.”
“Now, the most interesting question is: why do we need such complexity, why doesn’t the system work in a single window? The problem is that Chat GPT has a fairly short context. If you give it many tasks at once, it will get confused, just like any other employee can get confused with an abundance of tasks. Our task is to make everything work efficiently, so that robots don’t get lost or confused. To do this, we divide, firstly, the information context, that is, their roles, and secondly, we divide the communication context so that they communicate in different chat rooms.
You can create an infinite number of chat rooms and employees, and this can be done automatically. You can hire another organization that will invite employees and manage them – we call this meta-organization. Right now, we are in testing mode, but when we announce that everything is working, you can hire a meta-organization and ask it to create an organization of assistants for you.”
