Generative AI took us by storm.
The growth of ChatGPT since its launch has been nothing short of remarkable. Within just two months, it reached 100 million active users. It took mobile phone 7 years, and Instagram 2.5 years to get to as many users. The platform continues to attract hundreds of millions of visits weekly, demonstrating its widespread application across various industries.
OpenAI's revenues crossed over $2 Billion in 2024, largely from subscriptions, api usages and its enterprise solutions. The job market has adapted to this change as well, with a 21-fold increase in job listings mentioning AI-related skills since November 2022. It's not a joke that chief artificial intelligence officer is the most sought-after job nowadays. And yes, every organization will benefit from AI integration in business now that the bar is low and its effect immediate.
ChatGPT first became a phenomenon when I was working on my last startup. And to be honest, I was not elated that my cofounder was so captivated by ChatGPT that I had to keep him focused on our own tasks. Fast-forward couple months later, I decided that this technology is transformational, and that it needs my full attention. We both agreed that we should define our interests in the world of AI before we go any further. It was not an easy decision to make.
Personally, I've found ChatGPT to be a game-changer for coding, significantly boosting my productivity by handling mundane tasks. The benefits I gained for coding led me to ask the question of how such impacts could be replicated across other industries: Customer service, legal and administrative work, education, design and art, content generation, photo editing. It was endless. And of course, you face an issue when there's too many ideas to work on...
All organizations will benefit from using generative AI. It's a matter of how.
Today already exists many generative AI applications. The field is green, and we are only a year or two years into this corner of the industry. The playing field is very much even as well. It is not only the OpenAi's GPT models that are powerful enough to serve the current problems. There are other competitors like LLama from facebook, and Mistral AI that works just as well and more efficiently. Anyone with some technical ability can download, install and run Mistral 7B within an hour on their laptop. The llm models will continue to get better and become more democratized as GPUs (thanks to Nvidia) become even cheaper and more powerful.
So, I started edyt.ai in an attempt to build an AI integrated company from the ground up.
I thought about the process of building an AI company, and where the AI should be integrated. When working on something, we strategize, commit to a plan, act on the plan, and then assess the results. So I decided to integrate AI to strategy and planning for the first 3 weeks of this project. I wanted AI to be present where I begin my work. This felt straightforward to me.
And that's why we added a note-taking tool right next to the AI chat module. Psychologically, it feels like an AI integrated company, and objectively, I just didn't want to switch tabs to get answers from chatGPT, to increase my ability to stay focused to create and generate content. I'll give you an example of how I use it.
To write this blog, I ran about 2 long research GPTs (I'll explain this in a bit) before I start writing. And I asked two more questions as I'm writing the blog.
Here are the examples:
"I found so much efficiency for coding myself that I asked the question of how such impacts could be replicated across other industries. rewrite this"
"What are the industries that are being replaced by generative ai"
The questions are embarrassing but I find not having to switch tabs while writing immensely beneficial. I didn't have to lose any attention while writing. Yes, you can also just have two tabs open, have a notepad right next to it, and so on. And I am sure there are other companies who already does this. But that's no fun for a technologist like myself.
As for running 2 long research GPTs. Our research bot helps run a research about a certain topic. For example, my first prompt before writing this blog was:
"What are some cool products that are using ChatGPT today? What are the users saying about them? How fast is ChatGPT growing today?".
Research bot does not send this question to the ChatGPT, but it comes up with questions that best answer the user's question. So imagine running a query such as: How do we best answer this question: "What are ... today?". Create 5 questions that will lead to a satisfying answer.
We get answers like these:
- What are the most innovative products integrating ChatGPT currently available on the market?
- How are these ChatGPT-integrated products enhancing user experience or productivity?
- What feedback have users provided regarding their experience with ChatGPT-powered products?
- Can you provide examples of how ChatGPT's capabilities are being utilized in unique or unexpected ways within these products?
- What data or statistics illustrate the growth trajectory of ChatGPT since its launch?
After this, the bot looks up individual search results on the internet. It then summarizes the result and have it in a readable format by the user. The research bot having the ability to search over the internet helps me gather resources faster. While I was working on this idea, I ran queries such as "What are the best hosting solutions?", "What are the best domain name generator?", "How do I create a logo in midjourney" to allow myself to learn faster.
It works great as a brainstorming tool. It's difficult to get stuck when you keep on moving.
All in all, generative AI helps us create faster. This makes the technology inevitable. If your competitor is using it, you shouldn't get left behind. However, we also recognize that generative AI does not always provide a fully thought-out answer to complex, subjective questions. This is especially true for hard questions.
I recognize that the nature of content generation often includes unnecessary information. And that this is the same even for the best of humans. So I prefer to select content that resonates with me, move over to my notes and kickstart my process. This initial step often leads me in a new direction. And when the ball is rolling, it's hard to stop it. I believe that this is the best way to make use of any content out there. Research, digest, and make it your own. We are here to pick and choose the most beneficial information we can receive or give to someone else.
ChatGPT is excellent when we want to get started on a task. It helped me to get started with this idea. And helped me organize this project in 3 weeks. I can't say I know for sure where we are headed, but I will be the first consumer of all our creations. We will continue using everything that we build, and make integrations where we see most fit. So far, I'm really enjoying what I built for myself. It keeps me focused and productive.
The next AI feature revolves around personalization, committing, scheduling and preparation. Stay tuned, and thanks for reading!