What’s All the Fuss About Chat GPT 5?

To answer that question, we first need to break down what Chat GPT is and how it has evolved.

Chat GPT is an LLM (Large Language Model). An LLM is a type of AI trained on vast amounts of text data to understand the natural cadence of the human language, and generate human-like responses. That’s why you’re able to go on Chat GPT, and instead of entering lines of complex code, you can simply type what you’re thinking, and it will generate a response accordingly.

Some common LLMs you’ll find out in the wild are:

• Chat GPT: The most common and highly versatile.
• Perplexity: Primarily used for online research.
• Gemini: Google’s response to Chat GPT. Integrated into Google Search, and other tools.
• LLaMa: Meta’s AI, applied in customer service, healthcare, translation, education, and more.

How Chat GPT Has Evolved in the Past

• GPT-1: The first model of Chat GPT (GPT-1), was released in 2018. This was Open AI’s proof of concept. Though it had very limited capabilities, GPT-1 proved that an LLM could work, and paved the road for future AI research and development. Its only function at the time was text generation, based on understanding the context of user input. 
• GPT-2: Launched in 2019, this was quite literally Chat GPT 2.0. Its training data was expanded to improve coherency and contextual understanding. While this version of Chat GPT was much better at generating text per the user’s demand, ethical concerns were brought up regarding the generation of fake news or misleading content. 
• GPT-3: Considered the “Game Changer” in natural language processing. Released in 2020, its capabilities went from generating text responses, to performing complex tasks like content creating and assisting in research. 
• GPT-4: This raised the bar for Chat-GPT. Launched in 2023, this update was focused on creativity and processing. Updated with enhanced context understanding, it was able to engage in conversations in a less repetitive and redundant manner. With the introduction of GPT-4 also came multimodal, meaning it had the ability to process images and other documents. Future interactions of GPT-4 (like GPT-4o) expanding on its multimodal capabilities, becoming more accurate in understanding, and gaining the ability to write code. 

What are its Capabilities Now?

With the rollout of GPT-5 came exciting new implementations that users have been anticipating. Study mode is a new tutoring tool that came out with the update. This allows users to submit homework, assignments, and textbook material to help people prep for tests, or simply learning a new skill.

In addition to study mode, GPT-5 also upgraded their dev tools. Making Chat-GPT a better partner when it comes to app prototyping, UI design, and debugging. It’s essentially an expansion on the multimodal capabilities of GPT-4.

However, what’s got people really riled up is GPT-5’s new thinking mode and smart switching. Thinking mode forces Chat GPT to allocate more resources to processing your query. This is ideal for complex planning, strategy, and problem solving. This pairs nicely with smart switching, which allows GPT-5 to switch automatically between fast responding mode and its new thinking mode. 

The Controversy Behind GPT 5

The new features brought with GPT-5 are the next steps of integrating AI deeper into our lives. AI tutors can help us learn new skills. AI developers can help us build apps at a much faster rate. Thinking mode is designed to better serve our requests and queries. However, all this effort towards building and perfecting AI is scaring people. Many users state that GPT‑5’s thinking mode and smart switching provide too much. They fear that its ability to provide comprehensive, and well-structured solutions with very little prompting can erode users’ cognitive capacities. Specifically, the mental processes of planning, repetition, and problem-solving that are essential for learning and self-reliance. To describe this phenomenon, concerned users have coined the term “Agency Decay”, which describes replacing our own agency with that of an AI.

Forbes states that with the new model’s automatic switching takes away a key component in the problem-solving process, stating that ” One study found that AI usage had a negative correlation with critical thinking, with one participant claiming that “The more I use AI, the less I feel the need to problem-solve on my own. It’s like I’m losing the ability to critically think.”

What Can You Do?

There is a good reason for us to be scared. However, there are things we can do to maintain our independence and critical thinking skills. 

• Controlled Experiments: Select one significant weekly task to complete without using AI. Focus on maintaining your cognitive processes rather than optimizing for output.
• Thinking Partner: Don’t outsource your thinking to AI, instead use it as a wall to bounce ideas off of. Share your own analysis first, request critiques and alternative perspectives. Vice versa, critique its responses, and compare its work to yours.
• Cognitive Assessment: Regularly assess your own problem-solving skills. Can you still perform the analyses you routinely outsource to AI? If not, you should make practicing routine until your confidence and capabilities are restored.

Technology and AI are advancing at a rapid pace. It would be unwise to completely abstain from using LLMs at all, as it’s important to stay up to date on the latest news, and best practices. However, it’s critical to understand that over-reliance does exist! Only by combining our critical thinking with advanced technology, are we able to build a better and sustainable future.