I have effectively been exiled from Big Tech. In this post I’m going to give you my perspective on how that happened and my feelings about it. If you aren’t interested in hearing this tale, I’ll go ahead and summarize it for you up front: The values I grew to appreciate and build my work and hence my tech career of the last few decades around, are no longer in vogue. I refuse to apologize for this and I refuse to shift my position, hence the growing impasse.
Lets clear something up, right up front: This is not an official status. There is no central entity rating, ranking and black-balling tech professionals here. There is no established set of rules or system in place that prevents me from plying my trade, at least not directly. The danger of the position I now occupy is far more insidious than that. While contemplating this post, the one reference that keeps popping up in my mind is the classic essay written by Eric S. Raymond (whom I will have the joy of sharing a room with during Southeast Linux Fest in a few weeks) entitled “The Cathedral and the Bazaar”:
Linux overturned much of what I thought I knew. I had been preaching the Unix gospel of small tools, rapid prototyping and evolutionary programming for years. But I also believed there was a certain critical complexity above which a more centralized, a priori approach was required. I believed that the most important software (operating systems and really large tools like the Emacs programming editor) needed to be built like cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation, with no beta to be released before its time.
Linus Torvalds’s style of development—release early and often, delegate everything you can, be open to the point of promiscuity—came as a surprise. No quiet, reverent cathedral-building here—rather, the Linux community seemed to resemble a great babbling bazaar of differing agendas and approaches (aptly symbolized by the Linux archive sites, who’d take submissions from anyone) out of which a coherent and stable system could seemingly emerge only by a succession of miracles.
The fact that this bazaar style seemed to work, and work well, came as a distinct shock. As I learned my way around, I worked hard not just at individual projects, but also at trying to understand why the Linux world not only didn’t fly apart in confusion but seemed to go from strength to strength at a speed barely imaginable to cathedral-builders.
If you take nothing else away from reading this post, then I sincerely hope that it is the realization that tech, like so many other human systems, tends to evolve and devolve in cycles that can be observed at a higher level but generally only after the fact. It is exceedingly hard and generally foolhardy to believe that you can define the cycle while being within the cycle. However that is the thing I am going to attempt to do here.
So what is the cycle? Well as ESR aptly describes, the replacement of proprietary Unix by its now open source replacement, Linux represented his personal awakening in terms of the cycle we have seen play out over and over again within tech. That is the cycle of centralization followed by decentralization. While tech companies are gaining economic power and influence they tend to embrace the decentralized views of the day. They do this by appealing to people who exist outside of whatever the current mainstream of tech is defined as. Once they acquire a requisite level of economic power their views, actions and products tend to shift towards bolstering a more centralized approach to tech. This is because once they have acquired this power their mission shifts to preserving their grasp on that power. Convincing us that we need them and no substitutes will suffice is how they get there.
This happened with IBM mainframes decades ago. This happened and is still playing out with Microsoft Windows. We are seeing this cycle also play out in our current smartphone duopoly. Most importantly this cycle is driving all of the biggest big tech trends of the last few decades. This list includes the “Cloud” and of course “AI”.
There is an old phrase that I used to hear when I was first starting out. I still hear it sometimes but its a relatively rare reference as most of the people old enough to remember the time when this was true (not me, I’m actually a bit too young to make that claim) have long since retired:
Nobody Gets Fired For Buying IBM
The basic thinking here is that if everybody is buying IBM, then its the least risky choice one can make. Sure you could adopt tech produced by a competitor, but if the project goes south, there is a large chance in the aftermath, you’ll get saddled with the blame because guess what? You didn’t buy IBM! Oh the humanity!
This same mindset exists today and it has gotten far far worse. In our current cycle, we’ve seen this play out with the cloud and we are now seeing it play out with AI. All the movers and the shakers in tech have embraced the cloud. Nobody even blinks an eye when it comes to deploying their brand new service / product on Azure or AWS. These same movers and shakers have also embraced AI. Why are they doing this? Well to hear them tell it, its because these things are inevitable.
But are they? Once upon a time proprietary Unix was inevitable. With the exception of MacOS its basically dead tech now. Once upon a time Windows was inevitable. A few decades later, it’s basically in a tailspin and the tyranny and incompetence of Microsoft is directly to blame for it. Nobody ever got fired for buying those things… until they did.
So what’s wrong with the cloud? Well long story short as an organization, especially a tech organization that embraces the cloud, you are basically outsourcing key portions of your autonomy and agency to possible future competitors. Even in the event that Google, Amazon and Microsoft don’t want to directly compete with you, guess what? They don’t have to. Because they sold this entire industry on the idea that nobody but them is qualified to manage basic tech infrastructure, they have created a system where they directly profit for nearly everything everybody does. If we build everything on top of their tech, they don’t have to innovate. They can now profit off everybody else’s hard work while acting only as a middle man.
They no longer need to compete in a scenario where they establish themselves as a baseline dependency of virtually everything being created. This is by design. Microsoft doesn’t care if you use Windows or not, because the real money is in convincing enterprises to make use of Azure. Amazon doesn’t care if you shop at their site because the real money is in convincing enterprises to make use of AWS. Of course whenever any of these platforms experiences a major outage, massive portions of the internet suddenly become unavailable. Productivity can grind to a halt for hours or days at a time. It’s a shame more tech professionals don’t consider those incidents to be harbingers of a possible future in which the cloud becomes even less reliable.
So with all that being said the first step to becoming a big tech exile is to reject this notion that you must build dependencies on services offered by big tech in order to compete much less succeed in tech.
The second step is to reject AI as a subsequent follow up to this. Why would one reject AI? Well I’ll throw out a list of reasons real quick as a primer, how about that?
-
AI is trained on data stolen from everybody. They claim the right to copyright their models but those models were built on the idea that they had the right to universally ignore your right to copyright your creations. They raped every video, image, piece of art, blog post, news article, piece of classic literature, music and code (both properly licensed and not properly licensed) and then sought to sell us access to the output generated from their theft.
-
AI is a nightmare for our natural resources. Data Centers are being created for the sole purpose of bolstering the infrastructure required to implement this tech. These Data Centers suck down huge amounts of power, waste water, pollute the air and emit noise pollution (both in forms you can hear and in forms cannot hear).
-
AI is defective by design. The appropriate term for AI is actually “probabilistic computing”. Up until now most of the work in tech revolved around the creation of deterministic constructs that would ingest input in a structured format and produce output in a structured format. There were rules that were applied and enforced on both the input and output sides of the equation. Probabilistic computing dispenses with all of this.
However none of those reasons outline the primary reason classic tech practitioners ought to be rejecting AI, instead the primary reason is far worse. Just like we outsourced our infrastructure to big tech via the mechanism we now refer to as “The Cloud”, AI is a ploy to encourage us to outsource our actual thinking to big tech. When somebody asks a question at work, it is not uncommon to hear the following response:
Have you tried asking Co-Pilot?
Frankly I cannot think of a more offensive response than that one. But it is useful in that it belies the baseline truth of the trend: People don’t want to think about things. Much like those who find themselves deeply entrenched within any kind of cult, whether it be political or religious in nature, the primary defense of the trend falls back to what may in fact be humanity’s favorite logical fallacy, “Appeal to Authority”:
Insisting that a claim is true simply because a valid authority or expert on the issue said it was true, without any other supporting evidence offered.
The incontrovertible truth is that each one of these tech trends acts to centralize power in the hands of small group of tech oligarchs who now wield incredible wealth, influence and hence power. They have become virtually unaccountable. However there is absolutely nothing stopping all of us from opting out of this. If enough of us begin to opt out of these business models, these companies will be forced to adapt by providing us the kinds of services and products we are willing to pay for.
My willingness to advocate that position while being part of an industry that absolutely benefits from this continued malfeasance and society’s willingness to turn a blind eye to it, is what has ultimately led to my exile. My problem solving skills are very much on point. My ability to code competent and stable solutions is still very present. But those are no longer the top line skills that employers are seeking. That want to quiz you on design patterns. They want to have you solve cute gotcha l33t code problems live. They want to make sure that you can effectively go through whatever the current accepted practices are around ritually trying to draw any kind of non-hallucinatory output out of whatever LLM you have been saddled with for “coding”.
Ultimately my focus on actual fundamentals that used to matter, still matter, but just aren’t consciously observed by a lot of people in the hiring process has made me into a pariah.
When I was a young, bright-eyed and bushy tailed software engineer I had a collection of quips. Most of those quips did not stand the test of time. An example of one of those quips that did not age well was:
I work with computers because I like them more than people.
The primary problem with that one is that it is absolutely not true. As I came to eventually realize, every part of building software really is about the people. If you aren’t focused on the people, you are completely missing the point. On the flip side, here is another quip that has absolutely stood the test of time and has become more relevant than ever for me:
If you don’t understand a thing, you can’t control a thing.
That one underscores the biggest failure present in our current cargo cult thinking around AIs / LLMs. The people embracing this tech like it was the second coming of Christ have absolutely no idea how it actually works and they don’t care. A not insignificant number of the enthusiasts believe that it can reason, think and act as an independent entity which can be reliably consulted, tapped and trusted. Many of them also believe that this tech will result in AGI (Artificial General Intelligence) being achieved.
Nothing could be further from the truth, and truth is no longer in vogue.
So yeah if you want to be exiled from Big Tech all you really have to do is start calling bullshit. Tech used to be a semi-welcoming place for those of us who were willing to do it. Spitting in the face of trends was never a huge problem because there was generally never quite this much on the line in terms of money and trust. In addition most of the crazier trends of the past didn’t have this kind of reach or last this long. But times have changed. When the AI bubble pops, the reputation of Big Tech is going to pop right along with it. I can’t say I’ll be sorry to see that happen!
There are of course other opinions I have that spit in the face of other ideas considered to be sacrosanct in this industry, but I think that we can save all of that for another post on another day. As it stands I believe I’ve done enough damage for the time being.
May fate treat all of you kindly until the next time we meet.