Slop's Achilles Heel: It Increases the Workload, Not Just Costs
Set aside its inability to scale well (slop companies are trillions of dollars "out of pocket"), it does not seem to save money for its users either
Some time ago we wrote about how automation "on steroids" wasn't automation at all but a spaghetti of words without intrinsic value falsely marketed as "productivity".
Many people openly complain how that spaghetti of words actually wastes time and reduces progress at work. It creates more work (for workers to cope with), it doesn't do the actual work. As for "bots" (now dubbed "agents"), it's all hype and nothing new. They repackage old stuff using newer buzzwords.
Ryan has told us that slop now causes errors in Walmart, where he used to work (even manage, briefly).
He said sarcastically, "Gee I thought the computer was always right."
"Fixing AI slop for minimum wage while they scream that you're not productive enough isn't my bag," Ryan joked.
There are issues on top of it, distracting the staff. "Many Walmart employees say this thing is bothering them on their ME@Walmart app," Ryan explained. "Sending them hundreds of alerts that do not even apply to their department."
Skinnerboxes are a separate but related matter. They reduce the ability to focus, to concentrate on work. They destroy workflows. They cause mistakes.
Well, slop companies have mass layoffs even in August (shutdowns too), so whatever they have dubbed "era" or "revolution" clearly does not pay off. They're losing their ability to pay salaries (i.e. to do payroll) and they just borrow more and more money to toss down the well (or data centres that do fruitless work, if any). This is just a more 'reputable' version of fake-coins. They pay the media to play along and participate in the hype. This won't last much longer because expectations set by crooked media don't match up against reality. The complicit 'journalists' who said slop was the future would soon be disgraced or fall from grace.
As noted in IRC, slop does not scale because if you get something done shoddily (for example, a prescription, where mistakes cannot be made) and it's wrong in any way whatsoever and then you ask people to correct it (or to double-check everything), then you're not saving time. In many cases, correcting it takes longer than doing it right in the first place.
I said that the Sam Walton company will learn it the hard way as it also pisses off customers (who are told the "computer" made very blatant errors). Humans don't make the same sorts of mistakes and customers know it. For instance, a computer might dispatch 100 items of something... a human would not. It's common sense.
To use an analogy, suppose you get given a text like a template of an E-mail to send, bearing your name and presumably - or preferably - saying something both useful and correct. To quote a post from Geminispace, dated yesterday: "As an actual human, and someone that has a little bit of pride in communicating, I don't know why copy and paste parrots exist. What is a copy and paste parrot, you ask? A copy and paste parrot takes information you've supplied, usually internal between staff members, and copy and pastes that information _VERBATIM_ into an email or some other communication; it's a clear indication that the information is _not understood_ by the parrot. I am sure that the parrot thinks the information is useful for the recipient, but it usually isn't due to jargon that exists in the field, the organization, and the speaking/writing style of the author. I know that not everyone is the same; we weren't raised the same, our educations aren't the same, our values aren't the same, etc. This is what makes us unique! My complaint is this - why would you simply copy and paste informal jargon to a client who most likely/more often than not, would not understand the information?"
We now have a similar problem with slop. Evri, for example, sends E-mail messages in error. They rely on bots and those bots only cause distress to customers. They say complete nonsense. Automating the process results in anger.
In the case of E-mails 'drafted' by slop, the sender has to check and rephrase everything to make sure it is both helpful and factual. This is what actually takes time. Likewise, in pharmacies, packing up some pills doesn't take long (mechanical task), what actually takes longer is assessment, consultation, sanity check etc. Managers don't get it because they never ever did such tasks. At the end, slop-infused workflows might take longer, corrode customers' experience, and will result in more toil (error correction takes take) than doing it right the first time. Similarly "shorteners" and "summaries" by slop can mislead; while it's always true that reading long texts takes longer (it might take longer to read the original, not an abbreviation thereof), you take more information out from it and it's as accurate as the author intended and put it.
On this core idea that correcting it takes longer than doing it right in the first place, an associate has remarked that this is not about productivity, "it's about keeping people under the broligarch's thumb and if they're overwhelmed and underpaid they don't have time to unionise etc."
So slop is a power grab. It improves nothing. Andy wrote a lot about this. For the broligarchs it's difficult to fully control billions of workers (except financially), but if they outsource some tasks to their slop engines, then they gain a little more control over those workers and can bias processes at the engine's level, using all sorts of levers. Mechanising society where the mechanisation is about control by broligarchs is never going to be about progress, as Andy's latest article explains.
There are some relevant links in today's batch of Daily Links, an associate has noted. We consider this topic worthy of future attention because it impacts many people and we are doing our best to articulate what's happening. █