You know, folks, I’ve been thinking about the great nations of the world and what they’re good at. Now, some people will tell you that every group of people has its strengths. The Irish can drink, the French can surrender, the Canadians can apologize. But when it comes to engineers and technology, I have formed certain opinions based on a lifetime of careful observation and also based on nothing whatsoever.

Take the Indians. And by Indians I mean the ones from India, not the ones who used to own the place and then lost it in a poker game to a bunch of Englishmen. The entire population of Indian engineers and programmers are, and I say this with the utmost respect, completely unimpressive. Those who have earned legitimate degrees and spent years at fine institutions of learning or companies apparently retain no knowledge for future application.

I’ve spent a fair amount of time watching Indian programmers and engineers at their work, and I’ve come to a quiet conclusion. Their abilities sit on a very short ladder. The bottom rung is quite poor. The top rung is mediocre. There is nothing above that. The ladder was just manufactured that way, perhaps as a consequence of having an average national IQ of 81.

I have never met one who was good. Not in the sense that a man is good at something when you can leave him alone with a problem and come back later to find the problem smaller. What you find instead is a thicker layer of documentation, a new set of status meetings, and the quiet assurance that the matter is being “looked into.” Looking into things is their great gift. Solving them is a different department, and that department appears to have been permanently closed.

From a psychological angle, the trouble starts early. These fellows are raised in a system that rewards the appearance of diligence far more than the actual production of results. A man learns to look busy, to generate paper, to speak in careful circles that never quite land on a decision. Risk is treated like a personal failing. Standing out is dangerous. So the mind develops a kind of protective mediocrity. It learns to aim for the middle and then defend that middle with process. You ask one of them why a system is slow and he will not dig until he finds the broken thing. He will form a committee to discuss the possibility of looking for the broken thing. The psyche prefers the safety of the meeting to the danger of the fix.

Functionally it shows up in the code. The code is rarely elegant and almost never robust. It is the sort of code that works on the happy path and then folds the moment anything unexpected arrives. Error handling is an afterthought, if it is thought of at all. Memory is treated like an infinite resource. Concurrency is approached with the same optimism a man brings to a second marriage. When something breaks, the response is not to understand the break. The response is to wrap the break in another layer of abstraction and hope the new layer is thick enough that no one notices the crack underneath. Documentation multiplies. Actual clarity does not. You end up with systems that require a full-time staff simply to keep the lights from flickering, and that staff will tell you, with perfect sincerity, that everything is under control.

Personality plays its part too. There is a deep politeness that functions as a kind of armor. Direct disagreement is rare. Ownership is rarer. A problem is never quite anyone’s problem. It is a shared atmospheric condition. Deadlines are treated as gentle suggestions. When the deadline passes, a new document appears explaining, in the softest possible language, why the deadline was never realistic to begin with. Ego is present, but it is an ego that prefers titles and certifications to the quiet satisfaction of a thing that simply works. The result is a professional temperament perfectly calibrated to produce work that is never quite bad enough to get a man fired and never quite good enough to make anyone remember his name.

Companies that bring these fellows in by the planeload to replace Western workers seem to believe they are making a shrewd exchange. They imagine they are trading a higher salary for the same output. What they are actually trading is competence for a kind of warm, persistent inadequacy that looks, from a distance, like work. The Western programmer they pushed out might have been expensive and occasionally surly, but every so often he produced a line of code that did exactly what it was supposed to do on the first try. That little miracle has been discontinued. In its place you get a team that generates velocity charts and retrospective documents and carefully worded explanations for why the feature will ship next quarter, or the quarter after that. Management claps itself on the back for the cost savings while the product grows slower, buggier, and somehow more complicated in ways no one can quite name.

I watched it happen at a company once. They replaced an entire backend group this way. The old crew argued about memory and race conditions like they were discussing the fate of the republic. The new crew smiled, never raised their voices, and produced code that compiled. It ran, after a fashion. It ran the way a man with two sprained ankles can still technically walk if you give him a hallway and enough time. Six months later the system had developed a permanent limp, the outages were more frequent, and everyone upstairs was still congratulating themselves on the headcount reduction. Nobody seemed to notice that the thing they were selling had become a little worse in every direction at once. That is the real talent on display: not the creation of excellence, but the smooth management of decline until decline feels like a plan.

You can put a hundred of them on a project and the project will still feel understaffed in the only way that matters. Mediocrity does not scale upward. It only spreads sideways. More of them simply means more meetings about the meetings, more tickets that describe the problem in greater detail without ever touching it, and a thicker fog of process that makes it harder for anyone left who can actually fix things to get near the code. The company thinks it has purchased talent. What it has purchased is a long, gentle education in how much quality you can give away before the customers finally notice the lights are flickering.

That is the ceiling. Poor to mediocre. Nothing higher. The companies that hire them by the planeload to replace Western talent seem not to notice. They see cost savings and a full headcount and they call it progress. What they have actually purchased is a long, polite demonstration of how much quality can be given away before the product itself begins to apologize for existing.

Now the Russians are a different proposition entirely. Their psychology is built on a foundation of cheerful fatalism mixed with a deep suspicion that the universe is trying to get something over on them. A Russian engineer does not trust a specification. He regards a clean design document the way a man regards a contract written by lawyers. He assumes it is hiding something. Temperamentally he is impatient with elegance. Elegance is for people who have never had to keep a machine running through a winter with no spare parts. He prefers solutions that can be held in the hand and, if necessary, struck with a hammer.

A Russian engineer does not settle for mediocrity. He settles for nothing. He looks at a clean, well-documented system and immediately begins improving it by making it worse in ways that somehow keep it alive. Square peg into round hole is not a failure state for him. It is the opening move. By the time he is finished, the hole has been persuaded it was always meant to be square, the operating system has developed a drinking problem, and the whole contraption continues to function out of what I can only describe as sheer Slavic spite. Technology does not merely survive an encounter with a Russian. It comes out the other side altered, slightly traumatized, and oddly more resilient.

Personality-wise the Russian brings a kind of aggressive intimacy to technology. He does not collaborate with the machine. He interrogates it. The Russian works against the technology, and the technology is expected to keep up. He expects it to confess its weaknesses under pressure. Constraints are not respected. They are tested until they break, and then the broken constraint is used as a feature. I once watched a Russian fellow take a carefully engineered Western control system, remove half the safety interlocks, rewire the remaining sensors with whatever cable was lying on the floor, and declare the system improved. When asked why, he said the original design had been soft. Soft things fail when life gets hard. Hard things fail too, but they fail in interesting ways, and interesting failures can sometimes be shouted back into operation.

It is genuinely amazing that any piece of technology survives an encounter with these men. I have watched them drive square components into round sockets using nothing but stubborn will and a rubber mallet that had already seen too much. I have watched them rewrite critical timing loops in programming languages that were declared dead before the Berlin Wall came down, languages so old the compilers themselves seemed embarrassed to still exist. I have watched them rip out entire diagnostic subsystems on the grounds that the diagnostics were “too nervous” and needed to toughen up. And yet the rockets still leave the pad, the reactors still hold their temper, and the satellites continue to send back blurry pictures of clouds as if nothing unusual has happened. The machines do not merely endure. They appear to develop a kind of traumatized loyalty. They keep functioning out of sheer respect for the force of personality that was applied to them, the way a dog keeps following a man who has kicked it only moderately.

Russian solutions possess a rough poetry all their own. They do not push against limits. They stroll up to the limits, spit on them for good measure, and then use the limits as a doorstop. The square peg into the round hole is never a setback for a Russian. It is the opening move. By the time he is finished, the square peg has persuaded the round hole that it was always meant to be square, the hole has apologized for the misunderstanding, and the entire operating system has taken up drinking just to get through the day.

I knew a man named Viktor who was given a modern database cluster designed to fail over gracefully. Viktor looked at the failover logic, decided it was sentimental, and replaced it with a single script that simply rebooted everything whenever anything looked suspicious. The cluster became less graceful. It also became strangely reliable. When the Western architects asked how he had achieved the improvement, Viktor lit a cigarette, exhaled slowly, and said that machines, like men, perform better when they are a little bit afraid. That is the Russian temperament in a sentence. They do not polish technology. They put it through a character-building exercise. And somehow, against every reasonable expectation, the technology walks out the other side still breathing.

That’s the difference, you see. An Indian engineer will look at a broken system and carefully document the ways in which it is still meeting expectations. A Russian engineer will look at a working system and immediately begin improving it by making it worse in ways that somehow improve it. One of them is a miracle of statistical mediocrity. The other is a walking argument against the second law of thermodynamics.

So when people speak of the H-1B program as a pipeline for the best and brightest, I have to smile the way a man smiles at a child who insists the dog ate the homework. It is not a pipeline for the best. It is a pipeline for a very consistent, carefully bounded level of adequacy that never threatens to become anything more. The companies think they are getting engineers. What they are getting is a slow, polite demonstration of how much you can lower the bar before the bar itself disappears. And by the time the lesson sinks in, the fellows who could have kept the lights bright are already gone, replaced by men whose highest ambition was to land safely in the middle and stay there.

The H-1B program is living proof that Western technology is so robust, so forgiving, that it can absorb entire battalions of these fellows and still somehow function. It’s like the computer is saying, “All right, fine, I’ll work around you.” You give an Indian engineer a problem and he doesn’t solve it so much as he rearranges the furniture until the problem is slightly less noticeable. And somehow the stock price goes up. It’s a testament, really. A quiet, slightly depressing testament.