I waited way too long for a package that was supposedly optimized by an algorithm

Watching the little icon move on the map for forty minutes

I spent a good chunk of my Tuesday evening watching a little dot move across a map on my phone. It was supposed to be the ‘peak efficiency’ era of delivery. The app kept pinging me with notifications, telling me the rider had been matched perfectly through some AI dispatch system. They keep saying it’s not about finding the closest person, but the most ‘suitable’ one. Honestly, looking at the guy weaving through traffic in the middle of a rainstorm, I couldn’t tell if he was optimized or just having a really bad shift. It took about 35 minutes longer than what the app originally estimated, which is funny because the promotional banners for these platforms always brag about cutting off three or four minutes from the average arrival time. It felt like a marketing script collided with the actual reality of a busy Seoul street.

Walking through the automated warehouse idea

I read somewhere that companies like Lotte are now using these massive centers with over 1,000 robots to manage everything from picking to packing. The idea is that these machines are 98% accurate. I tried to visualize it—a grid of rails with robots zipping around. It sounds like something out of a sci-fi movie, but when I’m actually standing there waiting for my own order, it doesn’t really translate to a better experience. I wonder if the robots ever get stuck behind each other or if they have their own version of a traffic jam that the customer never sees. You hear about these systems optimizing routes and handling invoices automatically to cut costs, but usually, that just means the interface gets more complicated while the actual delivery time stays about the same.

The gap between software logic and the road

There is this company called Vroong that talks a lot about how they use AI to optimize the entire delivery process. It sounds impressive on paper, and I’m sure it saves the company a lot of money in the long run. But as someone who just wants their groceries or a cold meal, the logic doesn’t always hold up. Sometimes I think these systems are designed to maximize the driver’s efficiency rather than the user’s convenience. I once paid about 4,000 won for a delivery fee during a peak hour, thinking the ‘optimized route’ would justify the price. The driver ended up circling the block twice because the GPS coordinates for my apartment complex are notoriously messy, even for modern AI systems. It’s like the software is playing a game of chess with the delivery route, but it doesn’t account for the fact that a building gate might be locked or that the intercom is broken.

Why I’m still skeptical about the tech-heavy promise

Whenever I see job postings for logistics optimization roles at places like Coupang, I wonder if the people applying actually use these services regularly. They talk about robotic process automation and software engineering, which is great, but it feels disconnected from the person sitting in an empty apartment wondering why their delivery tracker has been stuck at the same intersection for ten minutes. Maybe I’m just impatient, or maybe the technology is still just in that awkward phase where it’s good at data processing but not quite ready to handle the chaotic, messy reality of everyday city logistics. I’m not really sure if these platforms will ever truly feel ‘optimized’ or if we’re just getting better at hiding the waiting time behind cleaner UI designs and more polite push notifications. I’ll probably keep using the services, but I’ve stopped trusting the arrival times. It’s just easier that way.