Too Big to Bother

Working in technology is weird. My work and home realities can seem like different planes of existence. Some of this is just the nature of the job. Distributed systems design and cloud computing management simply don’t have that many analogs in my family life. But it’s not just the domain. It’s the socially constructed reality: the shared consensus view of what the world is like.

On the one hand, I live in the real world, where I spend time with my family and do normal house stuff. I run errands and go to the store. I try to fix things around the house. I cook. I clean. I spend time with friends and we discuss these ordinary minutiae. We complain about how expensive things are getting and how bad the roads are. We talk about sports or the weather. Our lives do not appear to be changing that dramatically.

On the the other hand, I go to work, where everyone talks as though there is some giant change taking place that’s going to transform everything. A sizable chunk of folks are just obsessed with following everything that comes out of the AI world. Mandates are coming down that we need to be using AI in everything we do, because if we don’t we’ll be left in the dust. The inevitability of the AI transformation is assumed to be so obvious that it overrides any quibbles a mere mortal might have about it. We are at an Inflection Point, I am told.

Outside of my wife using ChatGPT to help with meal planning and research for house projects, none of this has seemed to leak into our personal lives in any meaningful positive way. My friends, by and large, do not care. My parents and in-laws need a bit of coaching to avoid trusting the machine, and to be on extra caution for AI-generated scams. And, that’s about it! This revolutionary technology seems to have done little more than change the way products are marketed.

In fact, this is a technology that, outside of white collar work culture, is generally unpopular. A majority believe the risks outweigh the benefits. People are increasingly skeptical of the control Big Tech has with their platforms and the influence they have gained over politics the world over. Communities are fighting AI datacenters being built, over environmental and utility-related concerns. Folks are increasingly annoyed at AI features being shoehorned into seemingly every product they use or consume. Yet it all continues, unabated. Every datacenter blockage is redirected to a buildout somewhere else, and every product failure is rationalized in a way that avoids blaming their ill-conceived and ham-fisted chatbot integrations.

I feel genuinely guilty to be a part of tech industry these days. There was a time when I felt some pride in telling people what I do. It was a respectable job. Now I’m less bold. Granted, I don’t expect to get blamed for these problems, but I do feel ashamed. I should have known earlier, I tell myself. I should have realized this was a crooked industry sooner. Didn’t they make a show about that?

Of course, when I think about it, it’s a bit silly to feel guilty? Most of us were duped. And what could I have done anyways? Technology was so genuinely cool for so much of our lives, it was easy to overlook all the ways it was elevating some truly nasty people to incredible power. And I got to build some of it!

There was a time when I really wanted to go out to Silicon Valley and join the whole movement. Around 2010 I was reading a lot of Paul Graham’s essays and thinking they made a lot of sense. Fortunately for me, I was too much of an anxiety-riddled man-child at the time to muster the executive function it would have taken to move across the country. I say this is fortunate because, chances are, if I had made that move, I would have burned out hard and fast.

But because I never made that jump, until my current job, I only ever managed to find work in marginally functional software shops, and never really saw how cutthroat the more dynamic parts of the industry were. So I experienced a much slower burn that gave me a front-row seat from which I could observe the management dysfunction that plagues most organizations. And so I concluded that this was the big problem with the technology business: it’s run by a bunch of goddamn PHBs and MBAs, stooges and sociopaths, sharks and management consultants.¹ If they’d let us focus on what’s important and stop interfering with useful work, then we’d be doing awesome things!

Let me illustrate: at one job, I was the a backend engineer embedded in an AI image recognition team tasked with deploying fraud detection software. Fraud was on the critical path, since we gated the overall amount of automation based our team’s specificity. Put simply, if we couldn’t reliably catch Fraud attempts, then we couldn’t automate incoming request processing, and had to route them to the back office.

So given the importance of our work, we of course had every manager and stakeholder conceivable involved in all our meetings, which tended to take up half of our time. Priorities changed on a weekly basis, and it was expected that we keep at least 1-2 workstreams in flight per person, guaranteeing maximum context switching and stress for everyone involved. We used some kind of “Spotify model” org structure that left everyone with effectively 3-4 bosses, each with competing agendas of their own.

I only managed to ship a handful of actual projects in 10 months there. It was hell, and everyone knew it. My colleagues were miserable. I remember a team member calling me at one point and crying because she was so exhausted from the constant and pointless pressure. She was a North African immigrant on a work visa in Europe, so even though she was stressed out of her mind, she felt completely trapped. But there was nothing we could do.

I genuinely wondered if it would be possible to unionize. Hell, I wasn’t even trying to get more more money; I just wanted to get workers to organize so we could kick management out of our meetings for long enough to do the work we were asked to do! But when a team is operating across 7 countries on 3 continents with multiple folks on contract or conditional work visas, it’s not even clear how workers could organize together.

The nature of our work, and the way we were working, had totally atomized us. Every layer of matrix reporting and worker status and complexity had served to further distance us from our colleagues and erode our ability to do anything about our situation.

There wasn’t anything left for me to do but move on.

I’ve since worked for more effective organizations. Nevertheless, it’s pretty easy to spot glaring inefficiencies virtually anywhere you work, if you’re inclined to look for them. But it’s surprisingly harder to find a consistent correlation between the presence or absence of these problems and broader company success. I mean, you can find companies that people generally say are well-run and are consistently successful. But you can also find tons of well-oiled failure machines, or inexplicably successful monstrosities. And I think many of the folks that operate at the highest levels of commerce implicitly understand this.

No one really knows how to run a business in all circumstances, or how to pick the right people to run them. So the folks who succeed at the highest level are either extremely lucky, or they’ve learned to focus merely on the bits they can control and treat the rest as a coin toss. Investors mostly influence capital structures and very high-level strategies. CEOs mostly influence what gets measured, what gets funded, and how a company is structured. The rest is details, and you can’t produce the kinds of results investors care about with any consistency if you focus on them.

Of course, most businesses do need to have some core fundamental strength to ever grow in the first place, but this is extremely hard to predict. The startup ecosystem exists specifically to attract talent so these dice can be rolled enough times until these things generate enough results to be recognizable to investors as something that fits into one of their playbooks. The most well-known of these playbooks are the scaling ones, where in the company burns cash to gain market share until it has near-monopoly power by virtue of some moat that fends off competition and allows it to start extracting ever-increasing rents from its captive customers.

This is desirable for investors because at this point, they don’t really have to worry too much about whether the company is consistently good at its own operations. Once you have a captive market there are a ton of ways you can hold onto it. Because you don’t have to worry as much about the boring minutiae of running and effective business, you can spend more time focused on ways to grow shareholder value, which is the only metric your investors really care about anyways. And since you’ve ascended high above the plane of business operations, you can survey the land for acquisition targets, direct capital towards trendy business goals, engage in corporate finance shenanigans to juice the numbers, ensure you’re well-guarded against systemic legal and labor risks, and lobby and hobnob your way to regulatory capture in multiple jurisdictions.

Once an organization has most of these kinds of things in place, there’s not a whole lot anyone can do about it, other than perhaps other organizations that operate in a similar vein and at a comparable scale. To put it another way: once you’re successful enough (usually defined in terms of market capitalization), you only have to be more useful than your competition.

When I was still despairing at my job working with the Fraud team, the most desperate idea I came up with was to go directly to the board. My assumption was that the board members, at some level, actually did want to run a good business. It’s possible some of them care about this. After all, this company had stagnated before it had achieved the kind of scale where it could stop worrying about being good. But it’s unlikely any board members would have been able to interpret what a well-run business looks like from their level. It’s just not their expertise.

And this is the realization I want to drive home: the people making capital allocation decisions, and really, setting the terms of who gets funded and how, do not understand what it actually valuable or effective. They have a number of playbooks that suggest how to use technology, skilled labor, and financial maneuvers to turn a small pile of money into a bigger pile of money often enough that they don’t get fired. And if it doesn’t work, the innovators among them will exploit whatever loophole they can to exit a bad investment in a way that leaves someone else holding the bag.

Needless to say, this is bad for everyone else. The myth of Silicon Valley capital has been that they change the world through risky investments in unproven technology, using their keenly developed senses to pick out great ideas and people before the rest of the world could see it, and give them the cash and support they need to succeed. It’s a more extreme and localized version of the essential social construct of capitalism: we tolerate people becoming rich assholes because the opportunity to make a fortune motivates ambitious people to do socially useful things. That is, the side effect of folks getting rich is innovation, production, and value. Everyone wins! Just some more than others.

But what happens when folks no longer have to be good to win? What if Venture Capital, now flush with cash and Twitter clout, is just a caste that dispenses largess to its chosen winners, and any interest in unique technology or challenging people has long since vanished? What if shareholder value has become so manipulable and gameable that it has ceased to bear relation to any underlying ground reality. SpaceX, a company with negative cash flow and no clear path to profitability, is currently trading at an obscene 100 price-to-sales ratio, mostly based on vibes around its AI operations and its CEO.

If companies don’t have to be good—just dominant, lucky, large, and memetic—then society is getting a raw deal from all this. It’s all well and good for private investors to take on huge risks in pursuit of brave new technologies. But the way this has played out, the risks are being foisted upon average people. Passive retirement investments will be coerced into buying speculative equities. Utility costs from increased power usage are being passed along to every day consumers. Datacenters buildouts are routinely cutting corners and causing harm to the communities in which they are rushed into service. Environmental damage affects everyone.

People don’t want AI, at least not in this form, where it’s clumsily shoehorned into every product whether you need it or not. Workers will adopt AI features as they become more genuinely useful. The insane top-down pressure to adopt AI and burn as many tokens just seems like a weird conspiracy your boss got unwittingly drafted into to juice AI company numbers right before they go public.

Watching all this play out in my own industry has radicalized me. I don’t know how else to put it. I can’t sit by and do nothing. I used to just be a guy that wanted to start his own company because I thought I could at least do better than some of these morons. I could at least figure out how to make software shops consistently ship working software. But now I don’t know if that kind of thing matters. It seems like the point was never really running a company well anyways. It was spinning the narrative so you could make the stock go up by any means necessary. And I don’t know if I can be that craven.

I still want to start my own company. Probably a consultancy. I have to opt-out of being a regular cog. I’m confident there’s plenty of work available for those willing to do a thoughtful, considered job. Some things do actually need to work, it turns out. This is quiet rebellion, aiming to leech off the inefficiency of the beast and foster a nice life for a select few.

But I don’t know how I can actually fight this kind of thing on a broader scale.

Electoral politics feels deeply broken for resisting industries with this much clout, money, and influence. Most governments seem deeply allured by the siren song of technology spending and the juicy potential mass surveillance capabilities the technology might realistically provide. And frankly politicians seem to be at least as deeply insulated than CEOs and investors from the consequences of their actions in the world at large. I won’t discount this entirely but it just feels like a broken way to try to bring change.

I’m not sure what technology workers (or any workers for that matter) should do about this to resist. I wish I had a better idea. I know there are a lot of people that suggest we should unionize, and I’m sympathetic to that notion, even if I have my doubts. A lot has changed about work since the heyday of unions, particularly with regard to how much less local work is for both employee and employer. Even if you are required to be on-site at your job every day, it’s likely you work with people in other locations and probably countries? Can they unionize? Do you know their local labor laws? How do you organize in this environment?

For software professionals specifically, I’m not sure we share enough common ground to even properly organize. Programmers are an idiosyncratic, pedantic bunch. Have you ever seen the arguments that happen when you try to get a bunch of coders to agree on something as relatively trivial as code formatting? It’s a holy war. Tabs versus spaces is a real argument that has gone on for at least 50 years.

More pointedly, trade organizations work when they have common goals. What would a software engineering union agree on? Quality controls? Training standards? Ways of working? Those all seem like huge arguments that a bunch of socially challenged…erm…neurodiverse individuals would have trouble settling. And many would balk at the notion that they need to enter in some kind of negotiation to innovate. I think the only thing we could agree on is that we’d like to keep our jobs.

It’s not even clear that would work though. I’m sure there are plenty of young folks that would gladly scab. Think the kind of kids that are anxiously working 996 schedules hoping to get a lottery ticket out of the “permanent underclass.” Or senior folks could just quit and work on contract. All of this works to undermine collective actions.

Union busting is also far more advanced than it once was. Don’t forget how good some investors and execs are at corporate structure manipulation. Are you a CEO who is frustrated that your code shop unionized? Just give a perpetual source-included exclusive license to another company, pay out the cash to your investors, and take a job with the licensing company, leaving your former company as a husk to be consumed at a fire sale.

Like I said, I like the idea of unionizing, but in practice I’m not sure it make sense for technology. Still, the core principle of organizing in response to the atomizing forces of industry is a good one. We see this in the success some communities are having resisting datacenter buildouts. In those cases, a community can come together and leverage existing legal structures to make it more difficult and expensive for companies to continue their plans. Granted, this kinda organization isn’t new—it’s just NIMBYism—but it represents and actual hurdle to this crazy bubble.

What the datacenter NIMBYs do show us is that resistance is actually possible. But it needs to be oriented around communities of common interest. Massive AI datacenters galvanize this by being tangible manifestations of the demonstrable harm this industry is causing. And maybe that’s the uniting force: when we agree about the harm being done, it’s the starting point for discussion of what to do about it.

So perhaps the principle guiding organized resistance to technocapital abuses should not be by workplace, or even trade specialization, but by a shared view of what is harming us, and how. Maybe not everyone agrees on the exact nature of those harms, but if we could at least nucleate around a couple of core grievances, we might have the beginnings of an organization.

Big Tech and Silicon Valley have succeeded in creating a socially constructed reality where AI is the next big thing and nothing should be let in the way of its inevitable progress. What is the alternative reality we can posit in opposition?