No-Code and Low-Code Platforms Might Disappear! Because of ChatGPT
A model named ChatGPT that engages in conversational interaction has been trained by OpenAI. ChatGPT can respond to follow-up inquiries, acknowledge mistakes, refute unfounded assumptions, and reject improper requests thanks to the dialogue style. The twin model of InstructGPT, trained to follow instructions precisely and respond in depth, is ChatGPT. Additionally, Stack Overflow is considering allowing users to submit on its platform materials created with technologies for artificial intelligence comparable to ChatGPT. It will soon outline a new policy that is clear on all of these issues. Does it imply that ChatGPT could cause No-Code and Low-Code Platforms to disappear.
According to a blog post by Stack Overflow, the current rate of correct answers on ChatGPT is too low and might seriously hurt both the website and users who are looking for the proper answer. One worry raised in the blog article is that many of the individuals submitting the responses obtained from the AI chatbot lack actual subject-matter expertise.
How does it function?
The system is created to deliver information and respond to queries through a conversational interface and is trained by AI and machine learning.
A sizable sample of text pulled from the internet is used to train the AI.
According to OpenAI, the new AI was developed with an emphasis on usability. According to the research organisation, “the dialogue style enables ChatGPT to respond to follow-up queries, acknowledge its errors, contest false premises, and reject unsuitable requests.”
No-Code and Low-Code Platforms: What Are They?
The visual software development environments known as “no-code” and “low-code” platforms enable both professional developers and amateur programmers to link and connect application components to build mobile or web apps. These platforms and the development techniques they provide are frequently used interchangeably.
Professional developers may swiftly build apps with the help of No-Code and Low-Code platforms because they are not required to write code line by line. They also make it possible for people who are not software professionals to create and test apps, such as business analysts, office managers, small business owners, and others. These folks don’t need to know much about machine code, traditional programming languages, or the design effort that went into the platform’s programmable components in order to create apps.
Limitations
Sometimes ChatGPT provides answers that are correct but are actually erroneous or illogical. Fixing this problem will be difficult because (1) there is currently no source of truth during RL training, (2) making the model more cautious will make it reject questions that it can answer correctly, and (3) supervised training will lead the model astray because the ideal response will depend on the model’s knowledge rather than the demonstrator’s knowledge.
The input phrase can be changed, and ChatGPT is sensitive to repeated attempts at the same question. For instance, the model may pretend to not know the answer if the question is phrased one way, but if the question is phrased another way, the model may respond properly.
The model frequently employs unnecessary words and phrases, such as repeating that it is a language model developed by OpenAI. These problems are caused by over-optimization problems and biases in the training data (trainers prefer lengthier replies that appear more thorough).
When the user provides an uncertain query, the model should ideally ask clarifying questions. Instead, OpenAI’s current models typically make assumptions about what the user meant.
Despite OpenAI’s efforts to help the model reject unsuitable requests, it occasionally follows harmful instructions or behaves biasedly. The Moderation API is being used by OpenAI to flag or remove specific categories of harmful content, but the company anticipates some false negatives and positives in the short term.
