thoughtwisps One commit at a time

Hello and welcome to thoughtwisps! This is a personal collection of notes and thoughts on software engineering, machine learning and the technology industry and community. For my professional website, please see race-conditions. Thank you for visiting!

dark quiet

I wrote this note almost exactly two year ago - on the 25th of December, 2017 when I was visiting Finland. The weather was crisp, the cold prickly and capricious on the skin. But instead of being clothed in its usualy gloomy grey garb, the Christmas sky was clear. I walked to the river that runs through a small, sleepy suburb, comfortably cradled between one of the highways leading to cental Helsinki and the large woodland areas of the Helsinki Central Park. On the footbridge that crosses the river, I looked up and saw the brilliance in the clear sky above me.

the bridge is in darkness

an angular ice cataract covers the river

only here the dark quiet reveals the stars

afterbloom

Written sometime between 2013 and 2014. Edited in 2019.

I still remember those April evenings when the air was hot and heavy and sweet with the afterbloom of cherry trees.

We were in the library working our way to the next best thing. I felt as though I was on the eve, my toes touching the water. And yet I yearned to suspend myself in this coccoon of heat and sweetness.

We wondered if these mellow days would ever succumb in the harsh glare of the outside world, its edges still held at bay by time, but already visible in the future. For honestly, what did we, at Bryn Mawr, little scholars in a world of wonder, know about the life beyond.

always be optimising

It started out as ‘always be shipping’, the little mantra we had in our tech department. We loved it. And used it liberally to the probable annoyance of everyone around us. Repeating it, living in it, simultaneously made us look like those chic, young, faux-moneyed Londoners, always working, always busy, drinking our fashionable almond chai lattes while running from lunchtime power cycling to a slew of work meetings and gave us an excuse to stay glued to our screens until late 8.10pm, after which you could officially expense any dinner on the company.

“Nando’s or Shake Shack?”

I rarely stayed for the Nando’s or the Shake Shack, but I remember observing this life from up close and enjoying the dregs of limelight that seemed to emanate from this reckless, relentless desire to produce more and more software, so that, at the end of the year, we might be the ones leaving our comp meeting happy and blessed with the bonus.

From where we sat we could see the glass-walled corner office, with a gorgeous view of East London. And I’m sure, we all took turns into little narcissistic power trips, imagining ourselves in it.

It wasn’t easy being an outsider in this world where (almost) everyone was a white man, second or third generation Oxbridge, second or third generation fancy, posh grammar school, with a busy social schedule of golf, tennis and dinner with some of father’s old Oxford friends, and little by little I drifted further away.

I took up long distance walking. Usually solo, but a few frightening experiences prompted me to sign up for a walk with a group. Walking along a canal tow path, a fellow walker started a conversation with me. He told me, he had left a City career to work as a career change coach of sorts, helping high flying executives who had had a come-to-Jesus moment change to a more fulfilling path. I suppose, as part of a professional curiosity, he started asking what makes my career fulfilling.

I didn’t have a response. The only thing I came up with was that stupid phrase “always be shipping”. Always be shipping, I said. That’s it. I added, that it was easier to run on the treadmill than step off it and think about actually moving forward.

I think a lot about that conversation and the “always be shipping” thing, which was borne out of blind ambition and hubris, than anything more noble. Very few things in that place were noble. Unless, of course, you consider acquiring a fifth pool or a sixth holiday home in Southern France noble. I suppose for some people it is.


I’ve since left this business of building systems that directly make money for a minor career change to building systems that obliquely make money. On the surface it’s dressed up in the fancy buzzwords of AI, business intelligence and whatnot. Although this industry has sufferred much the same fate in the court of public opinion as did the bankers in the post-2008 turmoil, it still continues to clothe itself in the garb of world saviour. And part of it’s rescue plan, is building these artificial intelligence algorithms that will one day diagnose you better than a doctor, police you 24/7 and customise the news and media you get exposed to, to make sure you are maximially engaged in the system. A happy dystopia, in other words.

A peculiarity about building these systems, and the reason I came to talk about them via the memory about “always be shipping”, is a step of optimisation that is required in the process of constructing them. We define a function that captures what we want to optimise and then we run an iterative mathematical procedure until we find values that optimise the function.

In other words, I’ve swapped out “always be shipping” for “always be optimising”.

This idea of “always be optimising”, which occurs in the process of constructing AI (please, Dear Santa, for Christmas I want nothing more than for the tech marketing spin machine to stop calling these glorified if-else statements artificial intelligence), is a reflection of the general tendency of human existence in late capitalism: optimise every waking moment to be more productive.

The media spends a lot of time speculating about the dangers of runaway AI, but we already have it in our midst: corporations that run on the mantra “always be optimising”.


It’s somewhat easier to sleep away from the restless frenzy of the big city. Here in the far North, in the middle of nowhere, there is nothing else except errant lights of airplanes landing on the nearby airport.

I miss the city very dearly, but for the very selfish reason that I fear that now, that I have removed myself from an interesting surrounding, I myself am no longer interesting. I will no longer draw curious questions in conversations or invitations to gatherings. Perhaps that is why I clung so dearly to a life that, at the end of the day, was not terribly fulfilling.

Even though I am now far removed from the relentless pace of the city, I find myself no farther from the same mentalities. I still work in technology, I’m still always be shipping or always be optimising. However, the fog of technoutopia, is thinner here. There are no meetups, no swag bags. There are fewer stickered laptops in work places and at the security screenings at the airport. It’s easier to see through the fog.

And it’s the things I see through the fog, that now, keep me awake at night. Every day I go to work and go about my “always be optimising”, someone else takes the work product and uses it to optimise things far more sinister than the benign, world-saving usecases we always envision for our work. Today, I make a few tweaks to optimise an algorithm to predict whether someone is at high risk for having a heart attack. Doctors use it to classify patients into high-risk and low risk categories to provide better care and monitoring.

Tomorrow, the same optimisation benefits a health insurance company using the algorithm to deny coverage to applicants with a risk for heart attack (fictional example, but I’m sure you get the gist). Because in a world where most consumer technology is built by companies, what matters is optimising shareholder value/ venture capitalist return on investment and for that optimisation function all that matters is who can pay the most.

an end in itself

It’s almost midnight here. The clocks have given us an hour’s mercy and induced mild horological confusion. Not that it matters to my brain anyway. Saturated with images of Primark hauls and the latest scandalous beauty community eyeshadow palette review, my brain has had some trouble going to sleep at the right time.

I probably should be brushing my teeth and flossing right now, but instead I’m running the hoover to un-gross the latest victim of my millenian weekend laziness: the kitchen floor while keeping half an eye on the banana bread in the oven (an attempt to rescue yet another victim of my laziness: overripe bananas that have been sitting on the kitchen counter for far too long).

These monotonous things which mostly run on autopilot and don’t require much thought are dangerous, because they leave my brain with space to ramble into spaces which I’ve neatly sealed into the attic of Randomly Accessed Memories and so to keep it occupied, I’m feeding it on a steady diet of backroung noise.

The selection is almost unlimited. Right now, my preferred background noise choices are 15679098727576th rerun of all three seasons of the Bold Type, an infinite playlist of Primark hauls or a selection of the best in the beautyguru drama episodes. None of them feel right for this quiet evening of last-minute chores.

So I go to my newest obsession: watching interviews with Jia Tolentino so I can get new insight into my millenial anxieties (her writing is very good).

The first thing that I find on YouTube is an interview with Tolentino and Doreen St. Felix from the Books Are Magic YouTube channel. Towards the end of the interview, an audience member asks Tolentino a question. The mic is very low so it’s hard to make out what exactly the audience member is asking, but based on Tolentino’s answer it must be something about “how to make it in a precarious industry constantly teetering on the edge of automation”. Tolentino prefaces her response with a disclaimer. Something of the “I don’t know if this is good advice” sort. At the end of the day, no advice is ever good advice, probably. The grains of salt, the survivorship biases, the hidden advantages and privileges are hiding behind well-intentioned advice.

You can't control anything about this industry. 
You can't control if you're going to get paid well.
You can't control whether you'll get paid.
You can't control whether people are going to read you.
You can't control what they think when they do.
What you can control is the amount of pleasure you can generate for yourself within your work. 
If you can make your writing fun and hard for yourself in such a way that even if nothing comes out of it,
it will be worthwhile to have done it. It will probably be the thing that makes you write well.
And that will probably be the only thing that can steer you in this industry.
I just think your writing has to be an end in itself."

In many ways, this “don’t do it for the external rewards, do it as an end in itself” sounds like something that belongs in the same sentence as unpaid internships and unbridled hustle culture in which we are posting and liking and swiping and tweeting in the hopes that the next little bit of ourselves that we project into the Machine will the one that will catapult us into the influencer heavens. But this would be an ungenerous reading of Tolentino’s advice. Instead I’m reading her advice as an optimist’s choose-your-battles-carefully resignation, an admission that lots of things about success are ultimately a crapshoot and beyond one’s individual control. And in this system, the only thing that is in one’s control is the thing itself, the writing.

There are lots of analogies here, lots of parallels I’d like to make with coding, with my own time in the tech industry, but I’m going to save them for another time.

thoughtwisps newsletter

hello friends and readers, first of all, thank you for being here and thank you for reading my blogo-essays. The bitverse has no shortage of material that is competing for your eyeballs, so I am grateful for the time you’ve let my words inhabit your imagination -even if just for the timespan of 280 characters.

Lately, as the days have been growing shorter and the pumpkin-spice latte hype stronger, I’ve given some serious consideration to the tasks that matter, the things I should be dedicating my time and attention too.

Of course, in doing such exercises, one must practice a certain amount of self-delusion (yes, I have been reading a lot of interviews with Jia Tolentino lately) to convince oneself that this writing is important enough to be written and to be read.

So, having taken my daily serum of self-delusion (which comes packaged in a cute dropper bottle and is probably available as a CBD infused version, because honestly, what things aren’t these days), I’m writing this to say, that maybe now, that I have gone through the tech pipeline and been disillusioned, I should go back to the drawing board and remember what was important to me before I became obsessed with outrunning this late-capitalist treadmill (you can’t outrun it, you’ll only get burnt out and drop off).

One of the things that I used to delight in was allowing the present reality to dissolve in worlds conjured by words. I’m hoping to spend more time working with word-worlds and sharing this with you. Thus, and here comes the whole point of this post, if you would like to subscribe to the thoughtwisps newsletter, which will feature writing about tech, AI ethics, machine learning, beautytech, literature etc, please follow this TinyLetter link.

Thank you! <3