Since the release of OpenAI’s GPT-4 API to beta testers last month, a group of developers has been experimenting with making autonomous AI models that attempt to perform multi-step tasks with as little human intervention as possible. These homemade scripts can loop, iterate, and create new instances of an AI model as needed. Created by Toran Bruce Richards, Auto-GPT is a tool that lets you create artificial intelligence (AI) agents that can complete tasks for you using GPT-4.While some of the leading names in the AI industry are calling for a break from the release of future iterations of the products, the journey to build them has already begun. The next step in the AI race is unlikely to be as revolutionary as ChatGPT. However, revolution is no longer needed and current technology needs to be well incorporated into our lives and AutGPT is such a tool.
How does Auto-GPT work?
Created by Toran Bruce Richards, Auto-GPT is an open source application and rather an experiment that can be labeled as a tool to showcase the capabilities of GPT-4. What the script manages to do is bundle together the tasks that GPT-4 is supposed to perform and perform them in the intended order.
To do this, the user first needs to create an AI agent in the app, give it a description of its role, and then identify the five goals it must meet. In doing so, it is necessary to provide an OpenAI API key along with a Google API key. After that, Auto-GPT takes output from GPT-4 for each step. The output can be shaped by iterations until it is tailored to the given goal, improving the result. Ideally, such a script could serve as an artificial intelligence assistant capable of performing any digital task on its own.
I think it would be more accurate to explain the situation with an example so that the top paragraph takes shape in your mind. For example, if given a task like “Buy an old pair of Air Jordans”, Auto-GPT develops a multi-step plan and tries to implement it. The tool can search for shoe sellers, then select a specific pair that fits your criteria. In other words, it can implement step by step what needs to be done on the way to the goal.However, we are still at a very early stage for this scenario. But experiments like this are of great importance in terms of seeing the way to go. While it might seem like a great idea to turn AI into an autonomous system, the GPT-4 technology it’s based on has a tendency to hallucinate and concoct information in the head. Therefore, for a complete success, it is necessary to increase the accuracy. In other words, if the artificial intelligence you want to buy an Air Jordan buys you a replica product, it will not be nice. Considering all these, we can say that we are at the beginning of the road. Such systems with feedback can also enable artificial intelligence to learn on its own in a subject that it is not trained. According to the researchers, this journey will end in developing artificial general intelligence (AGI).
anyone can try
You can try the whole thing yourself if you want. Currently, the AutoGPT product AgentGPT can be actively tried for free. You can try it right here on the website. You just need to set a name and just a purpose. All remaining commands are iterative and available to you as their feedback via GPT-4. I asked the artificial intelligence, which I call DH-GPT, the command “Make humanity better”. Then the tool “Analyze and identify the most common societal problems that prevent humanity from getting better.” He explained how humanity would become better by creating a total of eight new additional missions, along with a secondary mission containing the command.
31418
