Exploring the New “o1-preview” Model

Have you ever wondered what the world of AI might be like? Or are you already someone who has an idea and wants to learn more?

Whether you're aware of the train coming or deciding whether to hop on or not, especially in recent years, artificial intelligence has taken on a crucial role in many fields such as service, education, research, writing, social media, art, and science.

It’s no surprise that day by day, something new is added to the world of AI. And these additions are no joke—remember how ChatGPT shook the world when it first became public? Some people were terrified, some were excited, and that’s completely normal because we all felt the same.

What is “o1-preview”

One of these advancements has just become public: the “o1-preview” model, released by OpenAI in September. People interested in AI or those who use ChatGPT in real life already knew that they could summarise content, create new ideas, respond to various questions, learn something new, etc. But this new model is literally thinking before answering.

Of course, it can still summarize content, write code, answer questions, chat with you about anything, and do everything the previous models—like GPT-4o, GPT-4, and GPT-3.5—could do. But what sets it apart is its ability to create a complete software from scratch!

You didn’t mishear me; it literally creates entire software within seconds, which means the software and AI industry will skyrocket because even people with little to no knowledge can now create programs if they know how to prompt. It's also nearly flawless at fixing bugs and errors because it can remember even the longest prompts—something the previous models lacked.

Moreover, it can even meet the standards of Olympiads in maths, chemistry, physics, biology, and much more. Since one of the most improved metrics of this model is reasoning, it can now even compete in competitive programming contests alongside real programmers—and succeed! Some professionals even say it’s at a PhD level, capable of helping PhD students with their theses.

1 year of PhD Level Code in an Hour?

Personally, I found a video on YouTube from a person pursuing a PhD in physics. He explained that he had written a complex code related to space—about 1,000 lines—and prepared a thesis around it.

He demonstrated that he fed this new o1-preview model with the thesis, essentially teaching the model all about it. Amazingly, the model recreated the same code in just 200 lines within an hour!

As much as I feel bad for him, can you see what this technology can do? How can a model, a piece of software, complete a year's worth of PhD work in just one hour? I think now you understand the potential and the risks AI poses to our world.

By the way, here’s the link to that video—I encourage you to check it out after reading this blog: https://www.youtube.com/watch?v=M9YOO7N5jF8.

Why did I write this Blog

Now, let me tell you why I wrote this blog. Sometimes, due to our human nature, we may not welcome new technologies and innovations because people don't like change. However, according to most experts—and I agree with this view—this new AI train will have a much greater impact than even the introduction of the internet into our lives.

For this reason, already many people have started learning AI or how to use it to optimize their lives. High school students are now using GPT for their homework, companies are creating their own private models, university students are writing code with these models, content creators are producing posts, advertisers are asking how to manage budgets, even families who want to plan holidays are benefiting from LLMs (models like ChatGPT) to enhance their lives.

The fact that these models have a positive impact on everyone's life, produce extraordinary work, and save time shows how necessary and justifiable the AI hype is and that perhaps it will not go away for the rest of our lives. Now, even university students who are not studying computer engineering or computer science—like industrial engineering majors—have started turning towards AI.

Everyone is now aware of the coming train and hopes to contribute to it and earn a living by becoming AI developers, engineers, etc. And they are succeeding; currently, the annual average income of data science and AI-related jobs in America is over $130,000. Even newcomers, entry-level jobs, earn at least $80,000 to $90,000 a year in these fields.

That's why even many people, including housewives—people whom we might say "what relation could they have?"—have started jumping on the AI train. But there is still a need for many people; the world needs AI, and whether we like it or not, this technology will continue to develop. That's why we keep thinking, why don't we lead this technology, benefit from it, and earn our living from it?

If you are also among these people but cannot decide to start learning AI, let the PhD example I gave above inspire you. When you enter this career, you can develop such models and do much better, and while dealing with such fun jobs, you can earn really high amounts of money.

I hope this blog has contributed to you, and I hope you have gained an impression of what AI can do and how rapidly it is developing. If you were undecided about whether to jump on the train, and this blog helped you decide—whether positively or negatively—the purpose of the blog has been fulfilled.

Maybe you have even given up on learning by thinking "this is too much, I don't want to deal with it," and I would be happy about that too because at least you realize it's not for you, and that means I have helped you.

However, if the blog excited you and you are following what AI can do with enthusiasm and have decided to start learning, I am even happier for you. I hope that, together, we integrate AI into our world in the most ethical and appropriate way.

As we always say, we're excited to see you grow, learn, and work towards your dreams! Thank you for being part of our community. Take care, and let's keep moving forward!