Hello premium subscribers! Today I have the first guest post I've ever commissioned (read: paid) on Where's Your Ed At - Nik Suresh, one of the greatest living business and tech writers, best-known for his piece I Will Fucking Piledrive You If You Mention AI Again, probably my favourite piece of the AI era.
I want to be clear that I take any guest writing on here very seriously, and do not intend to do this regularly. The quality bar is very high, which is why I started with Nik. I cannot express enough how much I love his work. Brainwash An Executive Today is amazing, as is his teardown of Contra Ptacek's All My AI Skeptic Friends Are Nuts. Nik is a software engineer, the executive director of an IT consultancy, and in general someone who actually understands software and the industries built around selling it.
You can check out his work here and check out his team here.
Ed asked me to write about why leaders around the world are constantly buying software they don’t need. He probably had a few high-profile companies in mind, like Snowflake. Put aside whether Snowflake is a good product – most people don’t know what a database is, so why on earth does a specialized and very expensive database have a market cap of $71B?
That’s a fair question – and being both a software engineer and the managing director at a tech consultancy, I can talk about what’s happening on the ground. And yes, people are buying software that they don’t need.
I wish that was the extent of our problems.
Pointless software purchases are a comparatively minor symptom of the seething rot and stunning incompetence at the core of most companies’ technical operations. Things are bad to a degree that sounds unbelievable to people that don’t have the background to witness or understand it firsthand.
Here is my thesis:
Most enterprise SaaS purchases are simply a distraction – total wishful thinking – for leaders that hope waving a credit card is going to absolve them of the need to understand and manage the true crisis in software engineering. Buying software has many desirable characteristics – everyone else is doing it, it can stall having to deliver results for years, and allows leaders to adopt a thin veneer of innovation. In reality, they’re settling for totally conservative failure. The real crisis, the one they’re ignoring, is only resolved by deep systems thinking, emotional awareness, and an actual understanding of the domain they operate in.
And that crisis, succinctly stated, is thus: our institutions are filled to burst with incompetents cosplaying as software engineers, forked-tongue vermin-consultants hawking lies to the desperate, and leaders who consider think reading Malcolm Gladwell makes you a profound intellectual (if you don’t understand why this is a problem, please report to my office for immediate disciplinary action).
I’m going to try and explain things as they’re actually like at normal companies. Welcome to my Hell, and hold your screams until the end.
# I. The Industry is Sick in a Way That Can’t Be Solved by SaaS Spend
The typical team at a large organization – in a truly random office, of the sort that buys products like Salesforce but will otherwise never be in the news – might literally deliver nothing of value for years at a time. I know, I know, how can people be doing nothing for years? A day? Sure, everyone has an off-day. Weeks? Maybe. But years? Someone’s going to notice eventually, right?
Most industries have long-since been seized by a variety of tedious managerialism that’s utterly divorced from actually accomplishing any work, but the abstract nature of programming allows for software teams to do nothing to a degree that stretches credulity. Code can be reported as 90% done in meetings for years; there’s no physical artifact that non-programmers can use to verify it. There’s no wall with half the bricks laid, just lines of incomprehensible text which someone assures you constitutes Value.
This is a real, well-known phenomenon amongst software engineers, but no one believes us when we bring it up, because surely there’s no way profit-obsessed capitalists are spending millions of dollars on teams with no visible output.
I know it sounds wild. I feel like I’ve been taking crazy pills for years. But enjoy some anecdotes:
My first tech job was “data scientist,” a software engineering subspecialty focused on advanced statistical methods (or “AI” if you are inclined towards grifting). When a data scientist successfully applies statistical methods to solve a business problem, it’s called producing a “model.” My team produced no models in two years, but nonetheless received an innovation award from leadership, and they kept paying me six figures. I know small squads of data scientists with salaries totaling millions that haven’t deployed working models for twice that long.
During my next job, at an entirely unrelated organization, I was tasked with finishing a website that had been “almost done” for a few years, whose main purpose was for a team to do some data entry. This is something that takes about a competent team two weeks – my current team regularly does more complicated things in that time. I finished in good time and handed it to the IT department to host, a task that should take a day if done very efficiently, or perhaps three months if you were dealing with extreme bureaucracy and a history of bad technical decisions. It’s been five years and the organization has to deploy the finished product. I later discovered that the company had spent four years trying before I joined. It’s just a website! When the internet was coming up, people famously hired teenagers to do this!
I’m not even going to get into my third and fourth jobs, except to say they involved some truly spectacular displays of technical brilliance, such as discovering a team burning hundreds of thousands of dollars on Snowflake because they didn’t take thirty seconds to double-check any settings. I suspect that Snowflake’s annual revenue would drop by more than 20% if every team in the world spent five minutes (actually five minutes, it was that easy) to make the change I did – editing a single number in the settings that has no negative side-effects for the typical business – but they’re also staffed by people that don’t read or study, so there’s no way to reach them.
I warn every single friend who enters the software industry that unless they land a role with the top 1% of software engineering organizations, they are about to witness true madness. Without fail, they report back in six months with something along the lines of “I thought you were exaggerating.” In a private conversation about a year ago, an employee that left a well-known unicorn start-up confided:
“After leaving that company, I couldn’t believe that the rest of the world works this way.”
There are places where this doesn’t happen, but this madness is overwhelmingly the experience at companies that purchase huge enterprise products like Salesforce – the relationship between astonishing inefficiency and buying these products is so strong that it’s a core part of how my current team handles sales. We don’t waste time trying to sell to companies that use this stuff – it’s usually too late to save them – and I spend a lot of time tracking down companies in the process of being pitched this stuff by competing vendors.
In 2023, software engineer Emmanuel Maggiore wrote:
“When Twitter fired half of its employees in 2022, and most tech giants followed suit, I wasn’t surprised. In fact, I think little will change for those companies. After being employed in the tech sector for years, I have come to the conclusion that most people in tech don’t work. I don’t mean we don’t work hard; I mean we almost don’t work at all. Nada. Zilch. And when we do get to do some work, it often brings low added value to the company and its customers. All of this while being paid an amount of money some people wouldn’t even dream of.”
This will be totally unrecognizable to about half the software people in the world – those working at companies like Netflix which are famous for their software engineering cultures, or some of those working at startups where there isn’t enough slack to obfuscate. For everyone else, what I’ve described is Tuesday.
Also as a note to my lovely fans who email me about the RSS feed "including the whole premium newsletter" - I give generous previews! The rest of the (premium) article follows.