2025 Wrapped: AI Agents, Fantasy Football, and 31 Posts Later

It’s that time of year again where we all pretend to be surprised by our Spotify Wrapped stats even though we knew we listened to that one song 847 times (it was a Taylor Swift song). Well, I figured why not do the same thing for my blog? So grab your coffee (or bourbon, no judgment (unless its bad bourbon)), and let’s dive into what went down on this blog in 2025.

The Numbers Don’t Lie

31 posts. Thirty. One. Posts.

That’s more than double my 2024 output. I went from a “once-a-month-maybe” blogger to someone who apparently had a lot to say. We’re talking about ~12,753 words of technical content, Python scripts, AI experiments, and weekly fantasy football updates that probably should have their own newsletter at this point.

If you averaged it out, that’s a new post every 11.7 days. Not bad for someone who also has a full-time job, is launching a SaaS product, and is apparently trying to train an AI to beat his friends at fantasy football.

The Monthly Breakdown

Here’s how 2025 shaped up:

  • September: 7 posts (my absolute peak – more on this later)
  • December: 6 posts (strong finish!)
  • October & November: 5 posts each (keeping the momentum)
  • February: 3 posts (infrastructure month)
  • January & August: 2 posts each
  • July: 1 lonely post (I was probably on vacation)

September was wild. Like, “did Aaron even sleep that month?” wild. Seven posts. SEVEN. And if you’re wondering why, well…

The Fantasy Football Chronicles

Let’s address the elephant in the room: 21 out of 31 posts were about Fantasy Football and AI.

Yes. You read that right. Two-thirds of my 2025 content was me documenting my journey to see if an AI could beat humans at fantasy football. Was it an obsession? Maybe. Was it worth it? Absolutely.

Starting with “An AI Fantasy Football Draft Assistant” in August, I embarked on what became a 17-week experiment combining AWS Lambda, DynamoDB, Claude, and way too much time analyzing NFL matchups. Week after week, I documented the AI’s decisions, the wins, the losses, and that one time Josh Jacobs was questionable and the Packers just… didn’t play him. (FTP go Lions)

The Results? We made the playoffs. Finished 4th place. Beat half the league. The AI teammate actually worked.

Will I do it again in 2026? You bet your AWS bill I will. But next time, we’re building a custom model and adding proper injury management. Because if there’s one thing 2025 taught me, it’s that you can teach an AI to predict football games, but you can’t teach NFL coaches to not make questionable lineup decisions.

AI Was Everywhere (As It Should Be)

Fantasy football wasn’t the only place AI showed up. Out of 31 posts, 28 touched on AI or automation in some way:

The through line? Making AI actually useful instead of just flashy. I’m not here to build chatbots that tell you the weather. I’m here to build agents that save you money, reduce your troubleshooting time, and maybe win you a fantasy football championship.

The Infrastructure Deep Dive

When I wasn’t playing fantasy football GM with AI, I was deep in the AWS and Terraform trenches:

Python and Terraform remained my bread and butter, which makes sense when you’re trying to automate literally everything.

Fun Stats You Didn’t Ask For

What 2025 Taught Me

  1. Ship it: I launched products, wrote code, documented everything. Perfection is the enemy of done.
  2. AI is a teammate, not a replacement: The best use cases combine AI capabilities with human expertise.
  3. Document the journey: The weekly fantasy football posts weren’t just about results – they were about learning in public.
  4. Infrastructure matters: Whether it’s Terraform, AWS, or your blog’s export tool, good infrastructure saves you hours.

Looking Ahead to 2026

I’m not slowing down. Here’s what’s coming:

  • CloudWatch AI Agent: Just launched on December 30th. This is going to help so many teams reduce their troubleshooting time.
  • Fantasy Football 2.0: Custom model training, better injury management, live NFL standings integration. We’re going deeper.
  • More open source tools: If I built it and it’s useful, I’m sharing it.
  • Consistent content: Keeping this momentum going. More tutorials, more projects, more real-world solutions.

Thank You

To everyone who read these posts, tested my tools, provided feedback, or just silently judged my fantasy football lineup decisions from afar – thank you. This blog exists because people actually find value in what I’m building and sharing.

Special shout out to everyone who:

  • Tested drawiototerraform.com
  • Used my GitHub repos
  • Sent me messages about projects
  • Challenged my AI’s fantasy football decisions

The Actual Spotify Wrapped Numbers

Oh right, this was supposed to be like Spotify Wrapped. Here you go:

๐ŸŽต Top Artist: AWS (it’s not music but it might as well be)
๐Ÿ“Š Minutes Listened: ~12,753 words = roughly 64 minutes of reading content
๐ŸŽฏ Top Genre: “Infrastructure as Code with a side of Fantasy Sports”
๐Ÿ† Your 2025 Achievement: Wrote more than 3x your 2024 output
๐Ÿ”ฎ 2026 Prediction: You’ll write about AI doing something weird again

Oh, and I tried to ride 6000 miles on my bike in 2025. Because apparently I don’t have enough hobbies.


Here’s to 2025. Here’s to shipping code. Here’s to AI teammates. Here’s to 31 posts and counting.

See you in 2026. Let’s build some cool stuff.

– Aaron

P.S. – If you made it this far, you’re either really bored or you actually like my content. Either way, appreciate you. Drop a comment, send me a message, or just keep reading. I’ll keep writing.

P.P.S. – Yes, the AI finished 4th in fantasy football. No, I’m not salty about it. Okay, maybe a little.

Comments

Leave a Reply