223
Fears of employee displacement as Amazon brings robots into warehouses
(www.theguardian.com)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 20 Oct 2023
223 points (96.7% liked)
Technology
59583 readers
2308 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
People hear the word “robot” and assume that there’s some level of intelligence involved. Often, that isn’t the case. A robot is usually just a sophisticated machine following a painfully specific set of instructions.
If something unusual happens that an engineer hadn’t written a thousand lines of code to deal with, it could shut down the entire line.
I’m dealing with these specific issues right now in a distribution center, and it’s just with shelf moving robots that Amazon has had for 10-15 years already. It’s amazing how dumb they are and how poorly they are programmed to handle exceptions, and they aren’t even doing puts and picks.
Eventually someone will figure out how to make robots that can handle the more complicated tasks that humans currently do. I figured we were still a decade or 2 away from that point, but if anyone can figure it out quicker, it’s Amazon. I kind of hate the possibility that they might have already figured it out, but I’m very skeptical of a simple announcement.
The orange kiva bots? Amazon stole/purchased a robotics company for those as they wanted to develop them inhouse. A lot of the issues with robotics is cost and imperfect environments. Even dumb shit like changing the warehouse lighting can screw up sensors, guidance systems, and other automation if it wasn't designed that way. Customers want automation, but they don't want to pay for it. If we have to cheap out on sensors, cameras, drives, and other parts, it makes for less reliable systems depending on the application.
There is nothing more demoralizing than knowing a design is going to fail from the beginning because sales let the customer dictate the parts and design elements to cut costs. Then we get yelled at when it doesn't work perfectly or make rate. Or worse, it does/did work but the customer now uses the system in a way it wasn't designed and sabotage any good we might have done for them. Best is when a customer doesn't maintain a system, something that has to start on day 0, and then throws a tantrum when it breaks down all the time.
Robotics and automation isn't perfect. I have seen some great systems run with little to no downtime and shitty systems that operators have to constantly babysit. Us engineers try our best, but we have to use the tools we are given. I will say that technology overall has boomed over the last decade, but the parts and shipping situation since the pandemic started still hasn't been solved.
Not Kiva bots specifically, but similar concept, different vendors. Lots of new companies popped up in the years after Amazon bought them, but their protectionism definitely set the industry back at least 5 years. I don’t want to share any specific vendors, because that would cause the info that I’m sharing to violate my NDAs.
And yeah, you’re totally right about the environmental and reading constraints, but I’m talking more the internal problems that I’ve noticed in a few different vendors’ logic. I’m also an engineer so i totally get what you’re talking about, and the dumb robot problem that I’m talking about is simply insufficient engineering. Problems like sending a robot to pick up a rack to move it, then setting a path that is impassable despite the system having all the necessary inputs and correctly recording and storing all the data necessary to make a more informed decision and choose a better path. But instead the robot takes the impassable path, then when its censors notice there is a blockage, it stops but it’s too late to get another path at that point because of system constraints. If the logic had properly validated the path with the available data at the time of assignment rather than visual cues much later, it would cause much fewer issues. So that’s one example the sort of thing I’m referring to when I say they’re too dumb.