Getting Started with AI

What is AI?

Unless you’ve been living under a rock the past 4 to 5 years, you have probably gotten some exposure to Artificial Intelligence (AI) in the form of fear mongering by CEOs, “ChatGTP” (yes I have unfortunately seen it spelled this way), or short form “slop” video on a social media website. My personal favorite is the boomer-minded Facebook rant of “the Ay Eye is takin’ all muh watuh!” (we’ll touch on this in another article) posted with an AI generated image (ironically) to try and make their point.

At the end of the day, AI is mathematics, probability, statistics, linear algebra (the GPUs like that one), and a ton of data. When ChatGPT (that’s better) strings together a coherent sentence, it is essentially making an educated guess about which word should follow the next one and then doing that over and over again billions of times, very fast.

A Brief History of AI

I am not going to go fully into this but I recommend you watch the above video from IBM Technology. The TL;DR is that a lot of these concepts are 50+ years old but wasn’t as “sexy” back in the day due to the lack of computational power available to prove it out at the time.

Is AI Going to take my job?

Maybe. I can’t say for sure. I tend to lean towards “no” generally. If you’re in a “physical” job like an electrician, plumber, truck driver, or roofer even, you’re in a safer position than others. If you’re in a white collar “boring” office job, you probably should at a minimum be concerned. If said job is a bunch of repeatable tasks completed in succession or some sort of straightforward process, I’d probably be extremely concerned that an AI agent is going to take over that workflow and replace me. I personally believe that it’s important to NOT be complacent but to be continuously improving in any way possible (this doesn’t just apply to AI tools for me but for all aspects of my life).

The Human-in-the-loop is Imperative

The Human-in-the-loop (HITL) is the concept that a human is directly involved in interacting with a process, and/or providing feedback. If you’ve played with any of these tools before, you’ll immediately recognize this concept. With ChatGPT or Claude, it’s not a one way street where the AI is just spitting text back and just providing the answer, it needs YOU to drive it.

The goal with learning and using these tools is to become the Ultimate HITL. If you have specialized domain knowledge and become the UHITL, you only become even more powerful. For example, an accountant with AI is 1000x more powerful and faster than an accountant without it. This is why I am extremely bullish on these tools, it’s unlocked a whole new level of skill set for me that I didn’t believe was possible. This is why I am not worried about losing my job to an AI because I have put myself on the path to becoming the UHITL and I genuinely hope that you do the same!

The Big Three

There are a multitude of AI companies that have been started over the past 5 years alone but these are “the big three” in my book:

OpenAI

OpenAI is the non-profit company founded by Sam Altman and Greg Brockman (and Elon Musk but there’s some drama there). The recent AI wave started with them in 2022 with the release of ChatGPT running on the GPT-3.5 model. This tool dramatically changed the game with what was possible prior and has only gotten better from there. GPT-5.5 is currently their best performing model and they’ve expanded to a wide array of other tools:

• Sora (video generation)
• Codex (coding agent - I have really enjoyed using this one lately)
• DALL-E (image generation)
• GPT-Images-2 (image model - generated the image at the top of this article by the way)

Anthropic

Antrhopic was started in 2021 by former OpenAI researcher Dario Amodei. If that name sounds familiar, that’s because he’s probably the ultimate fear monger haha. Every other week it seems like there’s a new video clip of him coming out saying that “it’s all over and AI is going to absorb it”.

Despite all of that, Anthropic has had a ton of success. Opus 4.7 and their other models are outstanding on their own but their products like Cowork and Code especially have changed the game for using AI beyond the chat window. Below are some of their offerings:

• Claude Code (coding agent - this was the first agent in general I had ever used, I have a soft spot for it)
• Claude Cowork (general agent - absurdly good tool)
• Claude Design (UI/UX tool/agent)
• Claude Mythos (Cybersecurity agent - this one scares a lot of people and rightfully so It isn’t open to the public at this time)

Google DeepMind

Google has been no stranger to innovating in technology and their Deepmind division is no different. They are led by Demis Hassabis who won a nobel prize in 2024 for protein structure prediction with AI.

Gemini is their chat interface as well as the name of their model. They are currently on Gemini 3.5. (I personally don’t use Gemini as much as ChatGPT and Claude). Deepmind is winning over on the other two on images, video, and music generation (music generation will be it’s own article). Below are some of their offerings:

• Lyria (Music generation model)
• Nano Banana (Image generation)
• Veo 3 (video generation)
• Google AI Studio (the full suite of their offerings)

Below is a 30 second Complextro track I “generated” with Gemini.

Where do you start?

If everything I have shown here is a bit overwhelming, don’t fret. Here’s some guidance on getting started.

• Understand the Chat Interface and Prompting - Pretty straightforward. Use ChatGPT, Claude, or Gemini within the Chat interface itself to achieve a task that you would normally do without AI. Learn how to “prompt” properly. Have one model optimize a prompt for another model (generate a model with ChatGPT and then use it in Gemini or vice-versa). Throw documents or other files into these models to get information out or have it generate it’s own documents for you.
• Start playing around with more specialized tools - I’d start with Claude Cowork to take #1 to the next level. If you’d rather get to more entertaining AI uses, start playing around with image and video generation models on Gemini, OpenAI or Midjourney (I may do a “Getting Started” for special domains/tools and plan on logging my own experience with these).
• Apply any domain knowledge that you may have to these tools - If you like recipes, have Claude generate a new taco recipe and save it to a pdf. If you write rock music and want to play around with audio models, have Gemini write a rock song as a starting point. If you have a spreadsheet and need it cleaned up, throw it into Claude Cowork and have it build a dashboard or do data analysis…(You get the point).

The Genie is NOT Going Back In the Lamp

On a final note, one thing I continually keep hearing is that the “AI bubble” is going to burst and that all of this is just going to “go away”. That Genie is not going back into the lamp. If OpenAI or Anthropic declares bankruptcy tomorrow, another will company will form to take its place. The world will not move backwards to a “pre-AI” state and regress. The technology is simply too good.

You can either make peace with that fact and become the UHITL or don’t and become left behind. The choice is yours.

A Final Word of Encouragement

If you have any questions on any of this stuff or would like any more guidance, please feel free to contact me at the bottom of the About page of this site or on any of the social media platforms (X, Facebook, Instagram). I’d be happy to help!