Welcome!
Welcome to AI Agents: A blog exploring the fascinating world of Generative AI and AI Agents.
You can trace the main reason for the creation of this blog to one quote:
AI will not replace you, someone who knows how to use AI will!
In this blog, I’m planning to:
This blog is about sharing practical experiments and real-life use cases for AI Agents in the rapidly evolving world of Generative AI.
Here, I’m trying to explore applications and case studies for AI Agents, while doing my best to keep up with the most promising trends, technologies, and models.
I’m NOT going to address the societal and policy implications of AI Agents: this is way too vast a subject and there are better forums to discuss that.
About AI Agents
Pandora’s box was cracked open when OpenAI released Chat GPT back in November 2022.
AI Agents are not the same as the Artificial Intelligence tools that have been around for a couple of decades already.
AI Agents are a form of Artificial intelligence that is built for collaboration.
Some are designed to work alongside humans to enhance efficiency and productivity.
Some are designed to work with machines, computers, or other AI Agents to solve even more complex problems.
AI Agents are more than tools, they’re interactive partners that learn from and adapt to their environment.
AI Agents can analyze data and patterns, offer well-informed recommendations, and automate routine tasks.
Why All The Fuss?
The rise of AI Agents is not “just another technological improvement” that will fade when all the hype has calmed down.
I have been using OpenAI Chat GPT since December 2022.
I have built AI Agents using OpenAI, Google PaLM APIs, and Anthropic Claude. You can see and try some of these AI Agents.
I have experienced firsthand how AI Agents are already making me vastly more efficient and productive.
Since early 2023, I have used many AI Agents and tools based on Generative AI to perform jobs I was previously routinely outsourcing to external vendors.
Let’s Do The Maths…
The cost of a software engineer in the US is about 1,200 USD/day.
On average a SWE (software engineer) will create and check about 100 finalized, reviewed, and approved lines of code each day.
TODAY (i.e.: December, 3rd 2023), if I use OpenAI’s latest and most expensive model (GPT-4 Turbo) to generate code, I can generate 100 lines of code (approximately 1,000 input and 1,000 output tokens) for 0.04USD.
YES, the code may not be as good (yet) as the code generated by a software engineer, but for the cost of ONE software engineer, I can have THIRTY THOUSAND AI Agents creating, reviewing, and testing code for me.
For the same cost as a software engineer, I can organize those thirty thousand AI Agents in three teams:
- Ten thousand agents to create code.
- Ten thousand agents to review and optimize code. Early examples of that are the Code Explainer and the Code Optimizer AI Agents.
- Ten thousand agents to test and approve the code.
Collectively, these AI Agents will be capable of generating ONE MILLION lines of code for me, EVERY DAY, compared to the ONE HUNDRED lines of code that a trained software engineer can produce during that same day.
I built this website in about ten days, without using any low-code/no-code platform, as a side project (I have a day job), without the help of any designer, developer, or web specialist, just using AI Agents. For more on this topic, see the case study I wrote to describe how it was built.
You cannot overstate the potential of Generative AI. Be prepared for a tsunami of disruptive innovations!
Hop On!
Your thoughts and insights are what make this journey exciting.
Comment, share, and subscribe to receive the latest updates.
Let’s embark on this exploration of AI Agents together!
If you want more, including some predictions made about the impact of upcoming AI innovations, see the About section.
For more information about my background, go to my profile here.
Let’s connect on LinkedIn if you want to get in touch. Drop me a message there if you have any questions!
