There’s a lot of people arguing over the utility of things like ChatGTP.
Like most technologies, it’s not surprising that the usage of a complex tool like this has different levels.
Like other tools, such as word processors or spreadsheets, they’re easy to use to do simple things, and then as your skill and knowledge increases about the tools, you can create more and more complex results from those basic tools. ChatGPT is just like this.
And, learning to use it, one should progress through these levels for greater and greater utility.
It’s also good to understand that any critique, or pundit talking about ChatGPT will be at a certain maturity level in understanding the technology. It’s important for you to know what level they are at before you accept their information so you can understand their position more clearly.
For instance, if someone is a level one user of ChatGTP and they’re commenting on its reliability, or its ability to handle complex things, they really can’t justify their interpretation, other by their limited understanding and their biases. Use this model to watch for this situation carefully.
If you haven't tried AutoGTP... You should try it. See link below.
If you’re still trying to jailbreak ChatGTP or asking simple questions, find out more about AutoGPT .
People with, programming skills, logic skills, good imaginations, can create automations that can use the internet and other data sources to complete work in an automatic and dynamic fashion.
Think of each session as creating an small information machine with inputs, data seeking, planning, and formatting of output into useful formats, that can be other information machines.
These are the stages of Conversational AI so far. It goes from the simplest form of using things like ChatGTP, to the most complex forms.
The complex forms have more of a chance to be correct, and extremely useful to organizations in the generation of new software, and other new digital objects, in a scalable, repeatable fashion.
All the results will belong to the creator of the automation, the program that runs the different conversations and queries with the large language model, and data sources.
This is like a maturity model for usage of ChatGTP. Where are you on the maturity levels?
---
Level 1: Ask a one sentence question and expect a useful short answer. This is the Q&A pattern.
Level 2: Ask a more open question with more structure, and expect a longer generation of an answer. This is "ask a long question, and get a long answer pattern"
Level 3: Set roles, and have a conversation with AI to get to a set of answers that are required in the session. This will include tone, format, mental models to create a chain of thought, to get a change of results that are useful. This is the chain of thought pattern.
Level 4: Give the AI multiple roles to work with to give multiple viewpoints, checking facts, checking ethics, checking bias, among other checks to show a complex set of conversational results for the human to consider. It is possible to have the AI debate to for and against positions to help clarify once thinking. This is the role playing agent pattern.
Level 5: Using a construct for developing a program that has inputs from the real world, and data from the internet and other data sets, the complete programmatic automation, which includes looping, self-planning, and expansion. It can seek extra information and tools to complete work with an advanced level of automation. AutoGPT is an example of this automation. It is possible to create autonomous agents using this technique. This is the factory pattern for AI. (It is very advanced, when recursion is used.)
The GAI Maturity model was developed by Grey Swan Guild, co-founder at @RobTyrie.
—
More on autoGPT from @bernardmarr
https://www.forbes.com/sites/bernardmarr/2023/04/24/auto-gpt-may-be-the-strong-ai-tool-that-surpasses-chatgpt/
