AutoGPT is so hot that it can complete tasks independently without human intervention, GitHub 27,000 stars

Andrej Karpathy of OpenAI has vigorously promoted that AutoGPT is the next frontier of prompt engineering.

Recently, there seems to be a new trend in the AI ​​​​world: autonomous artificial intelligence .
This is not groundless, and a recent study called AutoGPT has begun to enter the public eye. Andrej Karpathy, Tesla's former AI director and just returned to OpenAI, also promoted it vigorously, and praised it on Twitter: "AutoGPT is the next frontier of prompt engineering." Not only that, but some people claim that ChatGPT is outdated, and AutoGPT

is A newcomer to this field.

Once the project was launched, it won 27K + stars in just a few days, which also verified the popularity of the project.

GitHub address: github.com/torantulino...
The question is, what exactly is AutoGPT? It is an experimental open-source application that demonstrates the capabilities of the GPT-4  language model . The program is powered by GPT-4 and can autonomously achieve any goal set by the user.

Specifically, AutoGPT is equivalent to giving a GPT-based model a memory and a body. With it, you can give an AI agent a task, have it autonomously come up with a plan, and then execute the plan. It also features internet access, long- and short-term memory management, GPT-4 instances for text generation , and GPT-3.5 for file storage and summarization. AutoGPT can be used in many ways, such as analyzing the market and proposing trading strategies, providing customer service, marketing and other tasks that require continuous updating.
As netizens said, AutoGPT is setting off a storm on the Internet, it is everywhere. Soon, a netizen started experimenting. The user asked AutoGPT to build a website, and AutoGPT succeeded in less than 3 minutes. During this period, AutoGPT used React and Tailwind CSS, all by itself, without human intervention. It seems that programmers really don't need to code anymore.

The user later added that his goal was simply to create a website with React. The request is: create a form, add a title "Made with autogpt", and change the background to blue. AutoGPT successfully built the website. The user also said that if more prompts are given to AutoGPT, the performance will be better.
Source: twitter.com/SullyOmarr/…
Let’s look at another example next. Pretending to run a shoe company, the order to AutoGPT is to conduct market research on waterproof shoes, and then ask it to give top5 companies and report the advantages and disadvantages of competitors: First, AutoGPT directly goes to Google to search, and then finds a comprehensive evaluation of waterproof

shoes top 5 companies. Once it finds the relevant links, AutoGPT asks itself questions such as "what are the pros and cons of each pair of shoes, what are the pros and cons of each top 5 waterproof shoes, top 5 waterproof shoes for men", etc.
After that, AutoGPT continues to analyze various other websites, combined with Google search, and updates the query until it is satisfied with the results. During this time, AutoGPT is able to determine which reviews may be biased towards fakes, so it has to verify the reviewers.

During execution, AutoGPT even spawned its own sub-agents to perform the task of analyzing the website and finding solutions to problems, all on its own.
As a result, AutoGPT has produced a very detailed report of the top 5 waterproof footwear companies, with the strengths and weaknesses of each company, in addition to a concise conclusion. The whole journey took only 8 minutes and cost 10 cents. There is no optimization at all during this period.

How does this AutoGPT, which can complete tasks independently, work? Let's see.
AutoGPT: Build your own AI assistant in 30 minutes
As the AI ​​agent that has swept the Internet,AutoGPT can be set up in under 30 minutes. You can have your own AI to assist in completing tasks and improve work efficiency.
This powerful AI tool is capable of performing a variety of tasks autonomously, and is characterized by ease of setup and launch. Before getting started, you need to set up Git, install Python, download Docker Desktop, and get an OpenAI API key.
Clone the repository
First clone the AutoGPT repository from GitHub.

Navigate to the newly created folder Auto-GPT with the following command.

Configure the environment
In the Auto-GPT folder, find the .env.template file and insert the OpenAI API key. Then copy the file and rename it to .env.

Install Python packages
Run the following command to install the required Python packages.

Running Docker
To run Docker Desktop, you don't need to download any containers, just make sure the program is activated.

Run AutoGPT

Execute the following command to run AutoGPT.

Setting Goals **
AutoGPT is a powerful tool, but it's not perfect. To avoid problems, it is best to start with a simple target, test the output, and adjust the target to your needs, such as ResearchGPT above.
However, if you want to unleash the full potential of AutoGPT, you need GPT-4 API access. GPT-3.5 may not provide the desired depth to agents or responses.
AgentGPT: Deploying Autonomous AI Agents Directly in the Browser
Recently, another developer has launched a new exploration attempt on AutoGPT, creating a
project that can assemble, configure and deploy autonomous AI agents in the browser - AgentGPT. ** One of the main contributors to the project is Amazon software engineer Asim Shrestha, who has obtained 2.2k Stars on GitHub.

AgentGPT allows you to name your custom AI and have it perform whatever you want. A custom AI thinks about a task to be done, executes it, and learns from the results in an attempt to achieve a goal. The following is a demo example: HustleGPT, set the goal to create a startup with only $100 in funding.

Another example is PaperclipGPT, which sets the goal to make as many paper clips as possible.

However, users also need to enter their OpenAI API key when using the tool. AgentGPT is currently in beta and is working on long-term memory, web browsing, and interaction between websites and users.
How much room for imagination of GPT, we continue to wait and see.
Reference link: medium.com/@tsaveratto…

Guess you like

Origin juejin.im/post/7221089899281580091