Developers are building ways to automate ChatGPT prompts that encourage the tool to perform autonomous connected tasks, which will alleviate some of the limitations users experience when using it.
For example, developer Toran Bruce Richards launched his open-source application Auto-GPT on GitHub, a popular web-based platform where developers can store code, work with others and track code changes. It helps the developer community perform code reviews, manage different versions and track issues that need fixing.
Auto-GPT is generating a lot of buzz because it automates prompts for GPT-4, which is the latest version of OpenAI’s powerful ChatGPT AI. Many users call Auto-GPT a “superpowered” version of ChatGPT. With the Auto-GPT application, users can enter a list of tasks they want to be completed, and it talks with ChatGPT to create prompts and finish the different interconnected tasks. The application spawns its own sub-tasks that make contextual sense and help it complete stated tasks.
Using Auto-GPT requires some coding knowledge, as users need to connect with the OpenAI application programming interface (API) and create a token-based payment arrangement. Users must pay a small cost for tokens, which represent a certain amount of text that the ChatGPT tool can generate at an agreed-upon cost.
Other automated ChatGPT programs are popping up because of overwhelming demand. These include JARVIS from Microsoft Corp. and BabyAGI from software engineer Yohei Nakajima.
To stay updated with top startup news & investments, sign up for Benzinga’s Startup Investing & Equity Crowdfunding Newsletter
Countless Use Cases
Instead of working with the GPT-4 interface manually, the Auto-GPT application helps users save time and optimize efficiency by automating multistep projects. It gives GPT-4 autonomy to showcase its incredible prowess, making it perfect for generating product reviews, analyzing investments and building multiple types of content.
The Auto-GPT application is also self-evaluating — it will prompt ChatGPT to verify data. If it notes some data is likely inaccurate, it will relaunch a prompt as a subtask to pull from better data sources. It also can read and write files and has a memory, so it can catalog completed tasks and understand when it needs to take additional action and create more information.
These capabilities make it ideal for unlimited uses. Companies can build complex to-do lists with Auto-GPT, asking it to perform layers of tasks that not only export information, but also complete actions like sharing with colleagues and performing market research.
For example, to use Auto-GPT sales prospecting, a user could set it to identify 100 top prospect companies for a marketing-related product. It could find the names of the top marketing officer at each of these firms, write a sales pitch email, send the email and manage the responses, including setting meetings through a calendar management application.
Another use case is customer service, where the Auto-GPT tool can produce answers to simple and complex customer service inquiries. Users can ask Auto-GPT to solve customer problems by providing them with text and directing it to take actions that will resolve an issue to completion. It becomes an autonomous and automated customer service tool that’s considerably more powerful than any currently available chatbot.
The growth of Auto-GPT and similar tools points to a need for interoperability and collaboration within AI. For example, GenesisAI is reducing the hold the major tech companies have on AI by offering AI tools through a web platform. It makes AI services more affordable and efficient for businesses of any size, meaning any business can implement AI into their existing business model.
As AI becomes a more prominent, startups like GenesisAI and innovations like AutoGPT and BabyAGI will only become more common. Despite the sweeping effects ChatGPT and AI has had on the corporate America, the technology is still in its relative infancy.
See more on startup investing from Benzinga.