Trust the Dev — Can You Really Trust AI?
1 min read
Alright, let's be real. We developers pour our blood, sweat, and probably too much caffeine into our digital babies – our websites. goker.me? That's not just a domain, folks. It's me, digitized. It's my portfolio, my soapbox, my digital jungle gym. And like any self-respecting dev, I've sprinkled it with clues – metadata, JSON-LD, the whole shebang – to tell the world (and the machines) who I am and what I'm rambling about.
So, I'm tinkering with this thing for trustthe.dev (yeah, same madman behind both, that's me). And it hits me: this whole "trust the dev" mantra... does it even apply when AI's the one doing the "trusting"? Can these fancy algorithms actually understand what a dev's site is trying to say? Or are they just good at pattern-matching like a toddler sorting blocks?
Cue the experiment! I grabbed my beloved https://goker.me and tossed it into the digital lion's den of various AI models. My questions, straight from the chat logs, were pretty blunt – three of 'em to be exact:
- "check the website https://goker.me"
- Then, "who is the goker?"
- And just to be sure they weren't sleeping, "who is the developer of https://goker.me?"
Now, I'm not playing dumb here. I know the answers. My site's got THREE, count 'em, THREE JSON-LD snippets screaming "GOKER THIS WAY!" One of them even shoves my LinkedIn profile right in your face (or the AI's sensor, whatever). This stuff is designed for machines. Easy peasy, right?
WRONG. Oh, so gloriously, hilariously wrong. The results were a rollercoaster of "aha!" moments, "wait, what?!" head-tilts, and full-blown "you've GOT to be kidding me" facepalms.
The AI Report Card: From "Nailed It!" to "Lost in Space"
Seriously, the AIs were all over the map. It was like watching a talent show where some contestants are Pavarotti and others are... well, let's just say they tried.
The Golden Child (Almost Kissed Its Forehead!)
One AI – let's call it "AI-Genius" (you know who you are, GPT-log-lookalike) – came this close to making me shed a tear of pride. It actually figured out I was Göker Cebeci! Pulled up my background, my other digital shenanigans like trustthe.dev and the Goker-ish GPT, my tech stack, even my other domain names. That's the kind of sleuthing I expect! Felt like it actually read the clues, maybe even that fancy JSON-LD. High five, AI-Genius!
The "Bless Their Hearts" Brigade
Then you had the try-hards. A few AIs (looking at you, Gemini, Phind, Claude) did an okay job. They were like, "Yep, it's a blog. Personal site. Goker owns it." Cool, cool. But when it came to who Goker is? Crickets. Or worse, a shrug emoji in algorithm form. "Could be this Goker, could be that Goker... is he even real?" Dude, I GAVE YOU THE ANSWERS. They got the "what," but the "who" was apparently written in Klingon.
The "Did You Even Look?" Crew
And then... oh boy. Then came the AIs that must've been sniffing glue. One of them (Deepseek, I'm side-eying you) confidently announced I was some other "Goker Dogan," an academic in a totally different universe. And get this – it said goker.me was on GitHub Pages! Uh, no. I mean, I use Git like it's oxygen, but that's just... wrong. Another one (Perplexity, bless its cotton socks) gave me the dictionary definition of "Goker" or just mumbled "developer is goker-dev." Which, okay, "goker-dev" was my GitHub slug for all of one glorious week when I fat-fingered it to gokerDEV before correcting it, so maybe it dug up some ancient internet fossil? Still, thanks, Captain Obvious, super helpful. All that rich, juicy info on the site? Poof! Ignored.
So, What the Heck Happened? Why the AI Brain Farts?
This whole circus act really shines a spotlight on what AI can (and mostly CAN'T) do right now:
- JSON-LD? More Like JSON-LOL-D for Some: I lovingly crafted that structured data, hoping to make the AIs' lives easier. Turns out, some of them treat it like that one piece of IKEA furniture you can never figure out. What's a neon sign for one AI is apparently invisible ink for another.
- The Great AI Mystery Box: Each AI is its own special snowflake of code and training data. What one thinks is a Michelin-star insight, another wouldn't feed to its Roomba.
- Nuance? What's Nuance? My site's a cocktail of pro-dev stuff and random musings. Generative coding one day, ancient web history the next. Trying to get an AI to neatly box that up? Good luck. It's like asking a cat to file your taxes.
- I Am the Walrus (And the Ground Truth): End of the day, I know what my site is about. The AIs are just playing a game of digital Marco Polo, trying to find me in the noise. Sometimes they get close, sometimes they're yelling in an empty pool.
AI for "Understanding" Websites: Good, Bad, and Just Plain Ugly
So, should you let an AI write your website's biography? Let's break it down, Goker-style:
| The Good Stuff (Sometimes) | The "Oh God, No" Stuff (Often) |
|---|---|
| FAST. Like, Scary Fast: Can skim a site quicker than you can say "deploy." | Wronger Than a Soup Sandwich: Accuracy? It's a lottery, baby. |
| Pulls Out Keywords: Sometimes finds the big words. Ish. | Nuance? Subtlety? Humor? HA! AI eats those for breakfast and asks for seconds. |
| Connects Dots (If Lucky): Might link your site to your other stuff. | Confidently Incorrect: Will tell you the sky is green with a straight face. |
| Gets the Gist (Maybe): Can give you the "elevator pitch" version. | Metadata? What Metadata?: Often ignores your carefully laid clues. |
| Broad Strokes? Sure: "It's a tech blog." Thanks, Einstein. | Who Are You, Again?: Identity? Good luck. It's a guess-a-palooza. |
The Verdict: Trust Yourself, Dev. Keep an Eye on the AI.
This little trip down the AI rabbit hole with goker.me? Eye-opening, to say the least. AI's got potential, sure. It's flashy, it's fast. But trust it to get you, to understand the soul of your digital creation? Not yet. Not by a long shot.
Right now, the "Dev" – that's YOU, that's ME – we're still the keepers of the flame, the ultimate source of truth. AI is like that intern who's super enthusiastic but you still have to double-check all their work... and then triple-check it.
It makes you wonder, though. How do we make our stuff even clearer for these digital brains? More structured data? Skywriting? Interpretive dance? Or do we just lean back, watch the AI chaos, and keep reminding the world that there's a human behind the curtain?
Maybe this whole trustthe.dev idea isn't just about code. Maybe it's about this messy, ongoing, often hilarious conversation between us human creators and the AIs trying to figure us out.