If you have ever worked with Indians, you may have noticed that they are utter sycophants (in Minecraft). If needed, they also lie and deflect criticism. At the same time, they exhibit a staggering sense of entitlement. It only takes a weak manager for comical situations to emerge. For instance, I know of a case where an African manager with a huge chip on his shoulder gave an Indian engineer on his team one salary rise after another, but not because the latter was doing such great work. The reason was simply that this manager was conflict-avoidant and wanted this guy to shut up. Sensing a weakness, this brazen Indian just kept at it. It made for great entertainment.

If you think that the only reason Indians are so active in trying to scam people via the telephone because they are desperate, you may want to think again. My view is that there are so many Indian scammers because these people do have a different set of morals than you and me. Similarly, there are plenty of people who are poor but do not steal or trick people. On a related note, I am familiar with Indian scammers not only from videos online but also in real life. “Jack from Microsoft, here to help with Windows”, with a nigh-unintelligible English accent calls me from time to time. You can have a bit of fun with these pajeet scammers (scamjeets?). They just read a script and if you ask questions they do not expect, they just repeat the previous question, but only up to a point. Some of these Jacks also lose their temper quickly, in particular if you tell them to eat a dick and make an honest living.

The character defects of some Indians may explain the bullshit we see in tech. A prominent example is Amazon. Recently, the story broke that the company did not use A.I. to enable their “just walk out” shopping technology. Instead, they had 1,000 Indians go through videos, trying to figure out which items people actually put in their grocery cart. About 70% of all purchases were entered by Rajesh instead of some A.I. Of course, Amazon is deeply sorry about this and did not want to mislead you.

My reading is that the kind of deception Amazon engages in is simply a consequence of employing a significant number of Indians. In my experience, Indians will never say that they cannot do something. They furthermore frequently lie about deadlines, try to gaslight you if needed, pretend that you did not clearly explain what task needed to be done, and bootlick like no tomorrow in a vain attempt to ingratiate themselves with you. Amazon’s A.I. disaster is similarly explainable. All it takes is some VP of Engineering telling his staff that he wants fully automated A.I. for tracking customer shopping behavior. The Indians give promising status updates. Everything looks great on paper.

Six months pass. The status updates come in. Everybody is happy. Now the one-year mark is up. The VP wants to push this fantastic in-house solution out. At this point, he may be told that the solution is not quite ready but perhaps they could use some manual labor to cover the last 5% their A.I. just cannot deal with. At this point, the VP may face some heat from his SVP, so he agrees. The A.I. solution is rolled out. Seemingly, it all works perfectly. The costs for the 1,000 pajeets is labelled a “cloud computing expense”, and nobody raises an eyebrow. This sham goes on for years. The press is ecstatic. The VP is happy because he got a fat performance bonus, and the Indians on that team get pay raises and promotions. The pajeets are high-fiving each other with a shit-eating grin. They have no qualms about their actions because they delivered what they promised, and the VP just said that they all did a great job.

The problem with layers of bullshit, however, is that there is point where you will face problems that are so obvious that you cannot lie or deflect anymore. You can put 1,000 pajeets into an office building in Pune to watch video feeds in order to cover up that your shitty A.I. only has an accuracy of 30%, as opposed to 95%, as was claimed — these numbers are from the article. However, there is a limit to how much you can scale this approach. If Amazon wanted to increase the number of stores, the number of pajeets they use for video analysis purposes would also have to increase. Before you know it, you would need to use half a million pajeets to watch what some American shoppers put in their shopping car.

Reading between the lines, I wonder if the aforementioned Amazon A.I. project collapsed precisely because some VP wanted to expand it. After all, it was working so well in a small number of stores, so why not go nation-wide? At this point, he may have been informed that, “Sir, there may be some aspects of our A.I. that we should perhaps explain.” Obviously, this is speculation, but it is quite obvious that deception at such a scale is only possible if there are deep cultural problems, and based on my experience working with Indians, it is simply a plausible hypothesis that Amazon became the victim of the pajeetification of its workforce. I am sure that we will see a lot more of this in the future.