AI Ramblings
An unimportant post about AI (no em-dashes!)
Published on Tue Jun 23 2026
While AI continues to seep into every crevice of the world, it also continues to make some pretty funny mistakes. It can be hilarious to look at some of the most incoherent things created by all these new tools. At some times, maybe even a bit scary.
The Scary
It seems like every time I open up one of few social media apps on my phone I’m bombarded with posts generated with AI on pretty much every topic under the sun. In the moment reading these, it’s kind of funny. Who in the world would look at this and say this is a sincere message? Who talks like this? Every bit of this feels like satire! Then, I realize that not everybody is going to be able to easily point these out as slop. Not everybody is going to be able to catch what makes it obvious, me included. I’ve had to look at photos/videos multiple times before realizing they’re fakes.
Aside from the content, what also scares me is the repercussions this boom might be having. While I’ve heard and read about what the datacenters powering AI have been doing to the surrounding areas since it started becoming a household topic, I’ve not seen a lot of satisfactory coverage of it. I recently started watching More Perfect Union on YouTube. They recently posted a video on datacenters in Georgia, which really shows the water situation for folks in one particular area around the 5-minute mark. It’s a truly horrific thing to think about, and certainly not the kind of thing I’d like my children to grow up experiencing. I’ve already had my own fun with water pollution as a child, and I’d like to not repeat that.
The Funny
I’ll admit, I’ve laughed at a lot of the crazy unrealistic absurd memes I’ve found on the internet. Truly turning my brain off to disconnect from the horrors of the real world to enjoy absurdist humor does great things for my mental health. These things always make me wonder what goes through someone’s mind that helps them come up with the ideas that they then make memes of. The rise in AI video-generating tools only gave those folks more power to create the most insane pieces of media consumed by humanity. If I had no idea what was going on when the same people created still images, I’m at a complete loss for words when it comes to AI-generated absurdist videos. Some of these things are honestly unexplainable - trying to explain them would break my brain even further. I think that’s what’s so fun about absurdism.
Beyond the times I let myself disconnect, a lot of this content really just disappoints me. Especially when we start talking about AI-generated scam videos that come up on my social media apps every once in a while. These are another case where I can only think, “this is so obviously fake, how would anybody fall for this when that clearly isn’t that person’s voice?” Unfortunately, all it takes is a handful of people to fall for these things to continue making it worth the time.
On the LLM side of the aisle, I see agents making plenty of hallucinations daily, in and out of work. Despite their best efforts, I still see the occasional non-existent function call, or clear misunderstanding of a given bug. Sometimes, these can be hilarious. Other times are just frustrating. Letting GitHub Copilot spin on a markdown document/comment generating endless strings of nonsense is one of those things that can take a particularly frustrating bug or feature and make it a little more bearable.
Due to all the issues above, I’ve adopted a “trust, but verify” approach when interacting with tools like these. They’re quite good at speeding up research, pulling in all the information I care about quicker than I can even research a single option. The one downside is still the hallucinations, or the complete “carelessness” it produces results with that don’t even match the original request. While researching a new pair of headphones, I enlisted the help of ChatGPT to gather information on a handful of brands I’d heard about, and the use cases compared to what I would need. The most important factor to me was wireless capabilities, and the most promising one in the list was wired only! Had I not verified what I needed, I’d have had the pleasure of dealing with the headache returning them.
A Meta Example
One of the main inspirations behind this post was a mishap related to Google’s AI Overview feature - it mistakenly attributes several developer’s LinkedIn accounts to this site. What makes it funnier is that none of them are right.
Is My Intelligence Artificial?
You may have noticed a little discrepancy in this post. First, I talk about how I really don’t like what the AI boom is doing to the environment, and then I follow that up by talking about enjoying intentionally bad AI-generated videos. So, what gives?
I’ll tell you. I feel guilty using these tools and consuming media created by them. But, I also see value in tools like ChatGPT/Claude/Copilot/etc. I think that these could exist in a world where we also don’t pollute drinking water, raise power bills, and all of that. Maybe if we spent less time creating slop, invested more time in memory/cycle efficiency, and relied more on local-first models things could be better. But, that’s not really where the money is at. Sad.
