IT Meritocracy. Part 8: Don’t Beg Google for a Job.
Take Its Open-Source Tech to Build Your Own Future.
Object-oriented purity, little exception handling frameworks, and neat database access wrappers bring great satisfaction during the learning phase. Algorithmic exercise proficiency and hackathon wins make one proud too.
No offense, but that’s (for the lack of a better term) childish or developmental meritocracy: potty training if you will. Sure, a few reputable employers (Amazon, Google, etc.) love potty-trained (algorithmically proficient) kids they can groom according to their agenda. I don’t want to stir the leaders vs. followers debate. I admire the current consumer tech leaders. But despite my very open mind, I’ve always had a problem with grooming.
In any case an ambitious (and thus underpaid) developer is only worth grown-up compensation upon, well, graduating from the potty-training and elementary reading daycare.
Let’s not idealize Google or Amazon. Let’s not rationalize their hiring criteria to prove it right or wrong. Yes, it hurts one’s ego to be rejected by any of the top tech employers: legit (compared to dysfunctional corporate IT) and contributed a lot to our civilization. But for the most part they are irrelevant for your professional and financial future.
If you haven’t watched the cult Office Space movie, please find it online. It was shot in 1999, and I can attest from my 1996–2006 experience in the US alone, changing projects every second year on average (contractor’s life), that absolutely every client/employer/interviewer company I’ve come across (out of a few hundred) was more or less an Initech.
Most sucked organizationally and culturally, while a few “caring” midsize shops prided themselves on being one of “10 best places to work”. You’ve seen those website badges, haven’t you? I wonder who issues those awards. The same mythical agency, that grants “top consumer satisfaction” ones? Everyone knows what a “best place to work” really means: a foosball table in the hallway, free sandwiches/sodas, and, woo-hoo, the latest outrageous startup “perk” — a beer keg in the kitchen. All to compensate for the 5–25% lower than average salary. How many kegs of beer that buys?
Kids, who change schools and us, contractors, learn to adjust to any boss and environment. The absence of the aforementioned “perks” and even bad bosses don’t make one’s life miserable. However, when it comes to pay cuts, no can do. At least for me. Show me the money.
Google delivers. I can care less about its cafeteria, sleeping pods, and other quite annoying in your face “work-life balance” BS. The money is (finally) there: inflation-adjusted pre-outsourcing compensation, which should have been $300–400K everywhere — if you are wondering about “what if” (outsourcing never happened) scenarios. Of course, paid to significantly fewer real pros instead of everybody.
Which is a 100% technical problem — automating software development to force the code monkeys out completely vs. the current stalemate: the 80/20 ratio of the top 20% of the team writing 80% of the code, with the mediocre 80% struggling to write the remaining 20, eventually rewritten by the top pros anyway as a result of bug fixing.
Not everyone is paid $300–400K at Google and Facebook. Not everything is rosy there work-wise. They are public corporations meaning they are governed by bleak “shareholder” committees, subject to the usual corporate America BS. But for the most part, they do care and offer their engineers career future. Grow professionally, do more, and earn more. Compared to the theme of this series: cycling, they hire pros, race, and win, meaning from the technical standpoint, they pedal in a higher gear, albeit in a few relatively narrow consumer domains.
And that is exactly why Googles of the industry don’t want you. Or me.
See, there are no other promising consumer tech companies. Most of the startups in that field gets acquired by “Googles” relatively early. The chances are, you are not in the consumer automation (B2C, whatever) field — and good. Otherwise you’d have either gotten “poached” by a Google individually or as a company acquisition. Not to mention the whole consumer tech sector combined (blogs, messengers, eCommerce, marketing/search, etc.) hardly covers 1% of the worldwide programming jobs. I am not going to guess where you work, underpaid and abused to take an interest in this series, but if you do something consumer-related and hasn’t been discovered by Google or Facebook, simply apply there, problem solved.
All less fortunate geek brothers and sisters, suffocating in outsourced corporate IT, read on. Google doesn’t want to touch unglamorous custom business software: hundreds of trick interconnected data entry screens, complex workflows, role-based security, reports, etc. with a 10ft pole. Backed by mind-blowing data science analyzing clicks and likes, Facebook is still a vanilla PHP blog. Amazon has its focus too. Will they ever compete with Salesforce and Oracle? The answer is obvious.
It reminds me of prehistoric delivery vans here, in the US. People love to speculate about drone delivery, but no one cares about the 1970s UPS and Fedex fleet. No, AI is not going to write software anytime soon. Nor everything will be “mobile”. Any other “technology trends”? Keep ignoring the majority of the financial, insurance, and other ERP software: straight from the 70s, like the delivery vans.
Those systems and their archaic makers: SAP, Oracle, IBM, etc. plus all of the customizers and integrators from Deloitte and Accenture down to the most boutique solution provider shops are invisible for the general population (and the Silicon Valley). Does their stagnation matter? Not, while those prehistoric steam engines keep slowly chugging. They are going to run out of steam and stop, aren’t they? How long a software system can be zip-tied and duct-taped? The real Y2K is coming, while everyone is talking about flying cars and smart eyeglasses.
Here’s the current situation. A lot of great open-source tech has come from B2C Google, Amazon, and Facebook over the last decade and a half, while the B2B aka enterprise software industry (IBM and Oracle) entered hibernation upon the successful Y2K “aversion”.
Can Google, Amazon, and Facebook frameworks and databases be adapted for enterprise software development to finally bury the mainframe sh-t for good?
Absolutely. Along with all amateur circa-2002 enterprise Java spaghetti, typed by millions of code monkeys since the early 2000s. Will Google itself do it? No.
B2B is not more demanding compared to B2C. It’s different.
Both B2C and B2B are pro sports. The former is best compared to cycling sprints. The latter is more like long-distance endurance races. Enterprise software development requires a significantly longer attention span, not commonly found among Google geeks: eternal scholars and other “competitive programmers” groomed into perfect Googlers. Sure, they work on complex tasks and come up with brilliant solutions. Can they handle medical billing automation or equally convoluted accounting laws and regulations? I don’t want to pit B2C agains B2B. They are just very different.
B2B is not for everybody. Which makes it a lucrative niche for me — seeing a vast empty market, created by the two decades of enterprise technology stagnation. It’s stagnated for a reason. Not many developers love automating business processes. No one is forcing you. Just beware, B2C no longer needs programmers. Which you’ve probably already heard from various “industry trend” analysts, extending it to all software, since B2B is invisible for them like brown UPS vans for the general population. No “programmers” needed. Only “scientists”. Before you dismiss it as bull coming from semi-technical (at best) industry “observers”, it does hold a bit of truth.
Every eCommerce website, social network and mobile app has been “programmed”. All simple consumer products have been already developed. Not surprisingly it is all about behind the scene science now. Anyone can develop a vanilla messenger or blog overnight, like Zuckerberg. Smart content offering and friend suggestions are the key among other things. Predictive analytics for ad placement, NLP for natural voice conversations, trick 3D graphics and physics engines for games… They all use “algorithms”, hence Google’s scientific and academic bias.
But that’s just a handful of companies at the top. First, while Google and Facebook may have perfected their products (I’ll reserve my opinion about quite amateur PHP blog the latter remains to this day), which only need “science”, there are countless other B2C niches. And second, there is unpopular in the Silicon Valley B2B, that Google won’t touch with a 10ft pole, while Oracle and Salesforce cannot invent anything radically new, clinging to their 40 and 20 year old technology. It needs fresh engineering before any science.
Do college exam style “algorithmic” interviews present a problem, attracting tiger mom bred eternal scholars? Who am I to judge? Google seems to have a well-thought grooming plan to shape the fresh grads into ideal scientifically-minded employees.
To be fair, Googles also hire established scientists and top engineers — through a different, “poaching” channel w/o algorithmic tests. It sucks, that you have no official cycling race wins to be “poached” for, and thus have to enter algorithmic “time trials”. How else would you prove your pro racer claims? By submitting an abbreviation-ridden resume staffing bodyshops flooded the system with?
Google is a private company — free to hire whoever it wants. Who the f-ck cares?
Whether, besieged by tiger moms, Google has finally hired enough of their eternal scholar offspring, promoted to decision makers and dictating its even more “academic” hiring practices… Or bored out of their minds Larry and Sergei, inspired by Elon, spending his windfall on childhood dreams of fast cars and rockets, felt similar philanthropic urge to give back to the civilization by supporting science, Google is no longer an engineering company. It has turned into a giant extension of CalTech and MIT.
It won’t hire non-scientists like you or me. Let them be scientific. Let them invent new programming languages. It’s their money. It just doesn’t make prove anything outside their walls. Contrary to the popular urge of “inventing” a yet another Java-like language, the world doesn’t need more programming language inventors. It desperately needs top-skilled: creative and experience programming language users.
Google and facebook will survive without you. There is enough data scientists there to steer search and analyze likes via ML. Marketing wizardry is not essential for the world survival. Getting rid of COBOL is — to finally offer smaller customers w/o IT departments the same robust business process automation, that takes tens of man-years to develop conventionally e.g. with Salesforce.
The technology is already here. Don’t convince Google and Amazon recruiters, that you’d be proficient with their internal software development tools because consumer product development requires the same programming skills, as enterprise systems. Take those open-source tools, apply to your (business automation) domain, and prosper.
I am also grateful for competing with prehistoric Oracle and IBM instead of agile and deadly Google or Amazon.
Self-Driving or No Driving?
One can compare consumer software to pretty tech-stuffed cars. The best image for the enterprise software is, yes, a brown 1970s UPS van. I am aware of the Tesla truck. It is still a conventional truck.
Being electric, or touched by celebrity “genius” doesn’t automatically turn a conventional machine into an invention of the century. Just like being a (data or another) scientist doesn’t lead to inventions.
Engineers invent too, and often it is the fresh engineering look at the problem rather applying conventional science, leads to building things that’s never been built before. I don’t understand industry’s attempts to diminish engineering ingenuity in favor of 1990s and 2000s mythical (detached from programming) B2B “software architecture”, and today’s B2C “science”. No inventions are made if architects and scientists are simply “doing their job”.
Can the people with deep pockets at the helm of the industry finally accept the fact that software products and services are neither “architectural deliverables” (roadmap, standards and guidelines, server diagrams, and other useless documents), nor AI dissertations. They are code that works — doing something most (conventional) scientists and architects consider impossible.
Here comes my truck example.
Before we start formulating the conventional goal: trucks need to be electric to end the dependence on fossil fuels, etc. let’s ask a simple question. Why cargo vehicles use the same freeways, as personal transportation? Their driving patterns are different from passenger cars’. Car routes are unpredictable. People drive wherever they want whenever they want using all available streets and roads. Cargo business on the other hand is well-planned. Trucks don’t make sudden detours or stops. They drive from point A to point B. Or along their defined route with scheduled stops.
An interesting exercise, isn’t it? First, cargo vehicles shouldn’t share general-purpose roads designed for arbitrary car travel. Whether above- (drones) or under-ground (tunnels), that transportation system should be cargo-only. Will people drive/fly their trucks — assisted by self-driving tech? Nonsense. Self-driving, car communication, and accident avoidance are no longer needed.
There shouldn’t be individual vehicles either. Just completely automatic platforms or containers, self-hooking to each other to travel in the same direction and disengaging to make a stop or go somewhere else — all w/o slowing down the chain (train of those platforms if you will). Of course the infrastructure itself should be designed differently to support such non-stop flows. A circular system of connected beltways comes to mind.
Sorry for being a vague “technologist” like Musk. I am just illustrating my point. Why such system hasn’t been built before? Two reasons: a) the most essential piece: control software wasn’t possible with existing technology and b) everyone, including the celebrity visionary Musk is either clinging to the conventional (separate vehicle, self-driving, etc.) concept for backwards compatibility with the existing infrastructure, or want some cool science-backed concept like “hyperloop”.
A fully automated radial cargo transportation grid would still use wheeled platforms. Not cool enough. No scientific or algorithmic sizzle. Just pure engineering. That’s how good B2B is different from hyped B2C. Will it help our civilization more than the entire Tesla lineup, including the latest models: the roadster and the truck? Rhetorical question.
Will replacing IBM, SAP, Oracle, and even Salesforce with 21st century tech help the world more than the next version of Google Glass or “smarter” Facebook ads?
New ML-backed blogs and messengers are great, but there is a lot to do in the enterprise software, stagnated since Y2K.
I live in the greater LA area and though I don’t drive much nowadays, am intimately familiar with the infamous LA traffic. Doesn’t software history look a lot like the largely lost LA transportation battle? First there were mostly trucks (mainframe and mini computers), since only the rich could afford personal transportation. Then came the Ford’s assembly line and Model T (the PC). The latter in turn heavily influenced cargo transportation as well: corporate software started to use the browser securely, though initially Internet focused on public consumer websites. Finally Tesla built a state of the art, yet conventional driver-operated truck. Isn’t it what Salesforce did it 1999 to Oracle, where its founder came from? Meticulously optimizing and polishing age-old Oracle Forms.
No old player will come up with something radically different e.g. a shared service of those automated cargo platforms, that combines packages from different customers and automates the delivery through the entire city. First, they don’t want to cannibalize the sales of their existing technology — the reason IBM let Oracle steal the relational database idea 40 years ago. Second, no one can build a completely new infrastructure alongside the old one w/o disrupting the latter.
What if someone new is not concerned with the business continuity and backwards compatibility — offering the radically new service to the new set of customers never dreamt of it. Would these customers be open to move to a new futuristic city built on top of the new underground cargo infrastructure? Why not? I am sure, no backwards compatibility issues was a big part of Facebook’s win — annihilating MySpace, Google retiring Yahoo, and Salesforce beating Oracle.
I am not venting about the current state of IT. Merely pointing out how much easier and orders of magnitudes less expensive it is to build the great new infrastructure w/o backwards compatibility and other worries, on top of cutting tons of other workaround efforts of “integrating” and tweaking the decaying old crap.
Armed with the new generation of power tools, one pro developer can do the work of 100 Oracle customizers/integrators and 10 Salesforce ones.
Think about all of the (AI-assisted?) self-driving complexity. Changing lanes, analyzing and predicting other car movements, etc. All because your smart car is sharing the backwards-compatible infrastructure with the dumb ones and you insist on operating the car in two: manual and automatic modes. Is the new infrastructure for fully automatic platforms on virtual rails an impossible feat? They don’t need any self-driving tech, because the whole infrastructure is built for driverless vehicles.
Get into the minimalism mindset. Cut off one piece and you’ll find out, that you can cut two more. Suddenly the radical new technology doesn’t seem so daunting, does it?
Am I theorizing about it? Can the example above be classified as theoretical or even “applied” “science”? To impress some recruiter or interviewer rejecting all non-scientists? With fancy math and physics formulas, trick algorithms, molecular structures of newly synthesized materials? No, it’s pure engineering: assembling things of already scientifically researched pieces — in a completely different (compared to conventional “forward thinking” celebrities) way. That’s what good engineering is about.
I’m an engineer, not a “technologist” celebrity. I build things. And my partner Jason sells them.
Don’t Expect Any Help from the System.
Sure, there are accelerators like 500 Startups. There are angel sharks, that want minimum of 60% for $100K. Was it a completely empty space or Musk’s charisma in 1995? Young and green Elon was able to find solid investors that made him $11M (out of $300M) on his undeniable POS (no, that’s not the Point of Sale, but a more widely used abbreviation) Zip2, that never saw the light of day. And no, he didn’t invent PayPal on his own after that.
Where are his investors now? Churning vaporware of their established aka ”serial” entrepreneur buddies? Today Musk, with all of his entrepreneurial drive and charisma, would have no choice, but to make Zip2 work. He and everyone else would have no help from the system whatsoever — forced to operate outside it.
Not having a choice, I am happy currently doing that, but does it need to be this way? Well, with no one rushing to help, let’s build our own system — centered around engineering.
I am not talking about some “lamp” (Russian expression referring to the warm sound of old lamp amplifiers) geek community writing admin front-ends for databases, DevOps automation, and cute little frameworks, GitHub is full of. How about those 100-people projects reduced to five-expert teams? Or 20-people projects meticulously executed by a single developer.
Still not convinced? Let me give you a hint. You think you can do a much better job, than your dysfunctional employer, right? What are you waiting for? Unless you work at one of the vaporware startups, the system you are currently developing or fixing, sells, doesn’t it? Your employer makes enough to keep that 100 people (you can single-handedly replace) on its payroll. Here’s your niche validation.
Post your resume online and interview at different companies. Job-hop to learn different fields. Find your soulmate salesperson cofounder like I did. Yes, it’s very different from conventional expectations of employment and pay raises for life. The wise say “it’s only hard to make the first million”. If you really want that million, the niche discovery process is rather straightforward.
Almost forgot... If anyone has any doubt about practicing, what I preach, we will pay Google wages.