Obligatory Seriousness on H1B
A personal angle. Plus some data.
When it comes to immigration, I almost don’t care. That’s because I have ultimately become one of those upper middle class people for whom an expense account has been an everyday reality. I’ve done a fair amount of business overseas, and I’ve worked with a good number of people who come to the US from overseas. Also being in the IT business, I’m sure that I work with a great deal more Indians than the average Joe. However what I know, anecdotally, is all I know and I haven’t bothered to calculate a political opinion. That may be changing, depending upon how much good data I can get my hands on. After all, that’s my business as a data engineer - helping management understand what it’s supposed to know without resorting to seat-of-the-pants calculations.
In the meantime that primarily means that I can’t tell you how my general assessment of the mediocrity of the Indian IT workers I’ve known is owing to the fact of my own mediocrity or otherwise. The simple fact of the matter is that I’ve known and worked with some extraordinary individuals - some of the brightest I’ve known - who are Indian. The best boss I ever had was born in India, but grew up here in the US. One of my favorite colleagues ever was Desi (and I’ve never written that word before, it’s pronounced deshi). But also one of the most ethically poison people I’ve ever had to work with was Indian as well. So let me get that story out of the way.
Indenture at Redmond
My first position as a vice-president put me in Redmond working at Microsoft. My company had several projects in process there and I was tasked to open up a new office for the company. Our company was producing the first implementations of MSFT’s next generation of enterprise software. Think of it as the jungle of products before what eventually became Microsoft Dynamics and then PowerBI. MS was just in the process of acquiring Great Plains and something called FRx. It was a very instructive time for me as I had every expectation of becoming reasonably wealthy, which didn’t happen. But while that was a distinct possibility I had a sit-down meeting with the owner of a consulting organization similar to mine which specialized in whatever you would call the next softest form of human trafficking above indenture. Let’s call them GoaBlueStar Consulting. They had smart people from India, the usual cities, Bangalore, Hyderabad & Chennai and brought them over to work at Microsoft. He put them up at the kind of place where I stayed - Oakwood suites and various lower rent apartment complexes that rent to the hundreds of tech workers in the Redmond area. He billed Microsoft $75/hour. He paid them $35/hour. They lived away from their families and he was working on me to crack the whip, more or less. “When all this business closes, we’re going to have a big party in Goa.” Long story short, I was ready to swallow the vomit until I found out exactly how horrid the Microsoft product was. That product remained horrid for several years after I got away of that shitshow. It reduced my respect for Microsoft tremendously, but that’s another long story that probably does not apply to the company in its current form, or maybe it still does.
I will say however that the one thing that remains constant from then to now is that Microsoft is an empire of influence, and what it does affects the way hundreds of millions of white collar employees live and work. Everybody pretty much starts with Microsoft as a computer user and relatively few get beyond it. Never forget that Microsoft created MSNBC and that’s about the level of quality and massive influence you can expect. What that has to do with Indian H-1B is tangential to the indenture story.
The barriers to get an H-1B are, from what I can see, fairly tremendous. If you’ve worked on getting grant money for a PhD, I imagine the process would be similarly complex, frustrating, time-consuming and all that. So I make a significant distinction between those people who are on that train of merit and the average ordinary Indian dude or dudette who makes the journey to the states. I should also make it clear that I have been first hand witness to the Indian caste system at work here in the US, which is to say that there is a significant consistency in Indians’ regard for each other between there and here. People don’t suddenly become something else when they come to California.
Some White Dude Named Bob
I worked in Las Colinas, TX years before the Microsoft debacle at a company called I2 Technologies. They were, at the time, the market leader in supply chain tech. So of course I had to learn about supply chain and some things about their competitor, Manugistics and the little upstart Red Pepper. Supply chain is fascinating. At any rate, I was more than a little bit surprised to find that this company was something like 90% Indian. And you really don’t know or see stuff until you have that kind of immersion. There were quite sharp people on my project, a young woman was the brains, and her husband worked there too. During my first or second day I needed to find my project manager, whom I had not yet met. I only knew his name was Bob. My colleague asked around for his phone number, who? Mike’s project manager. Some white dude named Bob. Oh yeah, he’s on the 3rd floor. I had to laugh.
I2 had the highest percentage of PhDs I’ve ever worked with at a company. I ate pizza in the cafeteria and one of the employees taught me channel trading in the stock market. He had a math PhD and explained to me that a company the size of the [idiot company that dropped the eBusiness division] place I worked for was not large enough to merit its own analysts on Wall Street. Our shares were traded in a block with larger companies like Oracle and Siebel, in those days before ETFs. So when our financials were going to be good it wouldn’t matter if Oracle and Siebel were going down. Our stock would go down and hit near the bottom of the channel. But when we reported to the street, we would get a bump, still we were constrained by the price of the other holdings in the tech portfolio of the major brokerages. It was a valuable lesson that made me a lot of money, thanks to an Indian guy I met just a few times.
As an American, ‘of color’ sometimes I had a weird affinity. I’m sure I didn’t assume much about Indians, but it’s true that I tended to like those like the two or three I was close with who were generally from Delhi. In the way I prefer New Yorkers over folks from Pittsburg, I picked up certain signals. This made itself clear when I was working at Nissan on one of my biggest projects.
Lunchbox Mentality
So there above are a couple exceptions. Here’s the mainstream meat. There are several stereotypes about Indian IT workers, and like all stereotypes, there are people who rage against them and those who don’t bother trying to outrun their constraints. The primary stereotype has everything to do with headhunters. Nobody who does contract work, like I did, likes them. “The Indians run the business”, and it’s something that never makes the mainstream news. I once made the mistake of putting my phone number on my resume and I regretted it for months. The amount of email I get with boilerplate language is daunting. So let me speak from the gut about these stereotypes.
Indian headhunters are a pain in the ass. They are skimming middlemen. They’re everywhere. This is the cattle-call entrance to the IT business. Stupid American corporations don’t know how to hire technical people. They have outsourced the entirety of their technical expertise to 3rd parties. That keeps the skimming and the scamming alive.
Tata, Wipro, Infosys are the biggest body shops on the planet. They swarm American companies with their ‘talent’. They are chock full of C students.
Accenture, Cognizant, IBM Global Services are the second biggest body shops. They are chock full of B students.
Robert Half is better. Adecco is better.
Silicon Valley is a techbro network, and they don’t use those agencies. They use Y Combinator and their own access to VCs to find new talent. They find A students.
All of these contract employees have a lunchbox mentality. You can’t avoid it. I’m just here to do a job, I have no real loyalty. I might be balancing on the steel beam 200 feet above the street for this cool picture, but I cannot wait to go home. All I care about is what you help me put into my lunchbox. To hell with your company’s culture.
Second String Gigantism
If you need a way to understand this, think about the company Avenade. I learned about their existence when I was in Redmond. It turns out that most every corporation uses the ‘security system’ that was built into Microsoft Active Directory (MSAD). MSAD is how you are identified, by email, on your corporate intranet. So if you’re a corporate Windows user and you use a ‘shared drive’ or Outlook or that thing called Sharepoint, which evolved into Microsoft Teams, you might be familiar with it. Bear with me.
On the other hand, if you use GMail by Google, you have no need whatsoever for MSAD. It has all of the features of MSAD and it scales way beyond 50,000 users in one large corporation. There are about 1.8 billion GMail users.
Avenade began their global business by doing one thing and one thing only. Configuring mail servers running MSAD. In 2007 you would call Avenade and have their consultants come onsite and fix MSAD which was mostly broken. Today we have open security protocols like OAuth2 which is why you can sign into any [new] website on the planet with your email account, no matter what your email account is.
See what I’m getting at? There are millions of users stuck in the Microsoft ecosystem which used a lot of second rate technology and outsourced a lot of IT work to second rate companies. How much IT work do you need to use GMail? None. And GMail is free.
If you understand this dynamic, then you recognize how much of business and enterprise computing is stuck and captured by outdated, expensive, second-rate technology. But with the H-1B issue the question is how much of that second string has become so large and ubiquitous in the American labor market.
I hate to use analogies, but IT is America’s cocaine. If we weren’t addicted to it, there wouldn’t be cartels overseas profiting at our expense.
IT < Software Engineering
Open source technology has generated high quality software. The analogy of journalism vs literature is apt. In an information junkie world, all we need is our daily hit of NPR or whatever it is that Gannett makes. But if you develop a taste for higher quality information, you look for literature. Literary criticism requires broad and deep reading and writing too. So you want to know what the sharpest most articulate writers say about their own literary diets. That’s one of the reasons you, dear reader, are at Substack.
The same critical regard for what runs on my computer applies. There’s a limit to the amount of crappy software I can tolerate. The critical question now is how the cloud and AI architectures are evolving what IT is. It’s absolutely crazy to think that the best programmers, software engineers are going to step aside and let some AI do tricks. Just like you wouldn’t expect Steven Soderberg in film or Cormac McCarthy in literature or Wynton Marsalis in jazz or any such genius step aside and let bots do their job.
So ultimately the moral question of H-1B comes down to whether or not we are importing genius or just another crowd of indentured servants. Are we getting brilliant software or just cheap labor in the IT mines? How much of that labor market is going to get kicked in the teeth by Sam Altman’s AI gang?
Before you answer that question, ask yourself why these IT imports are not Japanese, Taiwanese, Russians or Swedes. What is so special about India? I don’t presume to know the answer fully, but I have my suspicions. They take me back to the stereotypes. Now I can tell the Nissan story.
Balaji
Somewhere back in the early 2000s I got an offer to work on a giant project at Nissan. It was, like much of my work back then, an ERP integration project. Which basically means you take your corporate accounting system {Oracle Financials, SAP FI, Great Plains, Peoplesoft, et al} and you graft it onto your data warehouse and make a financial datawarehouse capable of handling analytics for all of the financial analysts and managers in your company. I came onto the project late because the people who hired (IBM Global Services) didn’t think about a particular piece of the puzzle, which I happened to know. BTW, the guy who hired me on the technical side was the most irascible chain-smoking bastard I’ve ever met. He was an Indian who quit the project early to go work building porn sites. I’m sure he’s rich now.
Anyway, there was solid mix of talent which included my Muslim and Mennonite friends I spoke of in another essay. My project manager was the second-best ever, an ex-sergeant from the Taiwanese Army. The technical lead, was a dude from Orange County who did a brisk side business in the pre-owned luxury watch market. Most of us were independent contractors that were subcontracted to IBM Global Services. I came in through Adecco. There were perhaps 30 people on the project of which only a half dozen were Nissan employees. We finished the project, more or less on time. So I only had knowledge transfer (KT) left on my plate. This would be the part of the project where the IBM Team and independents would skedaddle and the guys from Tata would take over the day to day support of the new system in production.

Mike (above) would take over to be the lead support guy. He was a full-time Nissan staffer. We began a series of KT conference call sessions two hours a day with the offshore Tata team. Their leader was Balaji. He and I immediately connected. I could tell him something once and he would get it. The half dozen others on the call and three that were here onsite were ‘typical Tata’ and need extra time and simplification to get it completely right.
But what was amazing was the social dynamic. Every one of the Tata guys spoke Balaji’s name with reverence. It was as if they got extra cookies for just saying his name. I didn’t make anything of it until one day there was a disagreement about whatever. One of the local Tata guys tells me that since they are from the South (of India) they get no respect from the North. And I get a quick bath in the salty water of the Indian caste system. These particular guys were from Tamil Nadu and were darker complected. I paid attention to this one guy, whose name I forget and noticed that he and others in particular took the bus to work and ate sack lunches. They literally had lunchboxes. Who takes the bus in LA? People who can’t even afford a beater.
This all occurred long before my stint at I2, but the seed was planted. So for short time I began asking people which state they came from, but there is entirely too much to know about India and never enough time in my life to digest it. I’m just pleased that I’ve known enough good people from there and don’t bother trying to accommodate all of those classes, cultures, religions and languages. That doesn’t change the fact that some significant fraction of them over here are doing scut work in IT. That too is an ethical problem of sorts. When you’re going back to India or sending remittances, you know very well that you’re a day laborer.
Show Me The Data
But the over-representation of Indians is no mere stereotype or coincidence. Since we in the chatting class are burbling over with opinion and invective, I stumbled over the following Twitter thread. https://x.com/RobertMSterling/status/1873174358535110953
85,000 jobs per year is no joke. That is a stimulus package.
Now you can see the companies I talked about. This is mostly all about IT, not software engineering. You can’t buy Tata software, you buy Tata services. You can also see this isn’t rocket science money either. I say it’s depressing the salaries of people who can do multiples of these things, speaking for my regal self. Again, I’m a little pissed today because for the first time in more than 15 years, I don’t have an immediate connection and I have to go through the cattle-call entrance. But I’m working that out. Check out Tessellations if you want to get into the technical weeds.
These are not genius Silicon Valley tech bro salaries. These are, starting with the orange band, 2-3 years experience in basic technologies like HTML. This is how I perceive the current digital technology labor market. You get more and more people into the labor pool with less than 7 years of experience and they are forced to do narrowly crafted jobs, pigeonholing them into a technology that gets outdone by something new in that same time period. Then you have a flood of eager newbs who cram interview question books. There are literally thousands of people out there AI generating resumes to be read by AIs through the volume-based headhunter industry. It’s not a pleasant thing. And nobody uses Star Trek metaphors any longer.
Robert Sterling who made the presentation of all this data with no pointer to exactly where he got it, so that I could make a package here, is properly critical, and responds to the knee-jerk.
I agree with you 100% that they are almost entirely productive, law-abiding people. I’ve enjoyed working with many of them. My point is not about the character of individual H-1B visa holders, though—it’s about the effect the program has overall on the wages of US workers.
Yes it’s that, but also within the broader context of what is a proper American, we can’t blindly import any other country’s caste system into our major institutions. Moreover we cannot allow our ideologies poison what merit remains in our major institutions.
In the end, I don’t mind picking on Tata. As bad as many large boring mediocre American companies are, the Indian companies are just as chock full of midwits with ugly politics that are an entire drag on society. It’s rather a shame that so many institutions have lost the sparkle of excellent management teams and thus rely on HR coercion which bends the knee to political activism in fear of bad PR in our populist times. This is the axis of ugliness that I know quite well. So I guess I am obliged to be serious.
PS. I’ll be looking for the raw data. The US DOL site is hopelessly complex.
PPS. Do we care about making steel in America?








ChatGPT sez:
Over the past decade, the Asian Indian population in the United States has experienced significant growth. According to the U.S. Census Bureau, the number of individuals identifying as Asian Indian alone increased from approximately 3.2 million in 2013 to about 4.4 million in 2020. 
This represents an increase of approximately 1.2 million people, or a growth rate of around 37.5% over the seven-year period. While specific data for 2023 is not available, estimates suggest that the Asian Indian population has continued to grow, reaching approximately 5 million by 2024. 
This upward trend reflects the broader growth of the Indian American community, which has become the largest single-origin Asian group in the United States, surpassing Chinese Americans by about 270,000 residents as of 2020. 
The increase in the Asian Indian population can be attributed to factors such as immigration, higher birth rates, and the community’s strong presence in sectors like technology, medicine, and academia.
It’s important to note that these figures are based on available data up to 2020, and while estimates provide insight into recent trends, precise numbers for 2023 may vary.
---
I'm starting to believe that nobody gets rejected for H-1B. Can this be?
There has always been a shortage of top talent in programming. Top talent isn't merely faster in coding; they are able to divine elegant solutions to complex problems, resulting in greater simplicity, maintainability, adaptability, scalability. No number of Grade B bodies can do what one Grade A body can do (which is to see simplicity in the midst of complexity).
Management doesn't tell Grade A talent what to do. All it can do is tell Grade A talent what problems it wants to solve, and leave it to the talent to determine strategy. This is a giant cultural problem in the corporate world, an inversion of top-down authority, where low-level nobodies make decisions and top-level executives, including bean counters, rely on trust. That's an uncomfortably uncertain void in typical business operations.
At the scale of large corporations, particularly in the software industry, the execs (*not* programmers) see 20% of their payroll consumed by programmers, and that group is no less subject to a normal distribution of talent as any other. So despite the special role in top talent, it still only accounts for <15% of the programmers, and the averages remain, well, the averages. As management reviews its financials, it can't help but consider how many heads are in programming, the average cost per head, how high that number is, and what can be done to drive it down. Because that's what good management typically does: try to drive down costs.
The truth is that top programming talent is somewhat like top Hollywood talent in terms of the outsized effect that talent has on overall performance (read: "financial results"). Movie producers know the A-listers are crazily expensive, but understand that such talent is a unique but economically sensible aspect of the business. There are actors, bodies, and there is "the talent." They're all actors, but two distinct classes of people with two distinct places in the financial models of the movie-making enterprise.
People intuitively grasp the significance of A-list Hollywood talent. But applying such a model to programmers is much less obvious. We're talking about programmers now, not movie stars. We're talking about modern day laborers. Cost per hour is as significant a managerial force now in software engineering departments as it ever was in managing any major class of labor.
Keep in mind the nature of large software development departments. Those who can program do program, while those who can't program manage programmers. What, therefore, could be a discernible contribution of a mid-level manager of programmers? Cost per head often becomes their priority of the day.
Good CEOs understand this software talent problem, and try to somewhat provision for it. But shareholders and most board members rarely look beyond the top and bottom lines, and presume nothing but cost in between that should best be driven down. Like the problems of a centralized command economy, as a practical problem of complexity itself, the view from the top necessarily ignores important details down the chain because there are simply too many important details for even the greatest brains to completely consider. When speaking of Grade A programmers, we're talking about a fraction of a fraction of employee laborers, and that's too far down in the details to be worthy of much consideration.
Completely aside.. Why do we import so many programmers from India? Answer: It's a source of a relatively lower cost English-speaking population. Language conformance...that is the largest distinguishing factor, in my non-expert opinion.
It is striking, as always, how similar your experience is to mine.