Last week I was tired. This week I’m exhausted. There’s a lot going on.

MSc update

Gang don’t – don’t try to do an MSc alongside a full-time job, especially when that MSc is in a proper academic subject like securing computers. And the Internet. And the various other Internet-connected things we’ve stupidly built. The people who started thinking about securing computers were the sorts of people who thought real deep about math and then built thinking engines. It’s so damn complicated and it’s so mathematical and it’s so wrong. It’s as if someone sat down and thought all the way through what a perfect meal would contain and never once considered that it would interact with a human.

Don’t get me wrong – it’s fascinating. But in my rush to perform perfectly, to continue being the guy who can do everything, I may have gone slightly beyond my means. If I could I would rewind and maybe just take this one module at a time. But I’m halfway through the session, after which the workload drops by half. And I will get an evening or two back and I am so excited. I’m going to take people on dates. I’m going to sit down of an evening and not fret that I should be studying.

My notes continue to take shape, a messy constellation of linked concepts. Not unlike the Internet; not unlike people. We design systems that reflect ourselves. As above, so below.

And this incoherent mess is how I feel at the moment, with so much pulling me all over the place. In a few weeks I’m going to become the sole maintainer of my team’s platform. Within 18 months I need a promotion, or my move to what’s parochially called “The Regions” (I’ve been here only six months and yet I bristle like the best of them: as if London isn’t a region!) will result in a chunky pay cut. The relationships around me evolve in ways I wasn’t expecting (but that I always hoped for!) and they must be managed. Salem has become a design critic, and is manifesting his unhappiness with peach curtains by spraying them with urine.

And my attempts to automate a problem – a piece of work I have been struggling with for seven years! – are starting to come apart at the seams as I fight with the same problem I accused my academics of up above: resenting that my beautiful mathematical model is being sullied by humans and their needs. It is the most fascinating, the most crunchy of problems. It is exhausting.

And despite this! Despite my sullenness and my cave-dwelling; my lack of energy; my frankly embarrassing obsession with a problem nobody else cares about (would that I had got into Dungeons and Dragons and could put forth Opinions on the OGL!); my need to end a sentence with a joke that punctures the tension built into the rising tempo; my self-referential bullshit –

Despite this I am surrounded by love. Love that is unconditional. Love that is still there when I don’t have the energy to match it. Love that overwhelms me and keeps me afloat. This is what it’s all about pals.

Work update

At work I continue to play with AWS. It’s so exciting and so powerful, and yet so incredibly badly documented. We also made a wrong decision about 2 years ago, and we’re kind of stuck with it for the interim, or at least until I spend two days re-writing the thing by hand. Nonetheless even working with that less-than-preferable approach, I’ve figured out a really neat way to get some code to run on a remote server in a way that’s deterministic and predictable, and that we can test.

I’m hoping to convince my organisation to pay for me to do a bit of training in this, because the more I use it the better I understand how well-architected cloud solutions like this can really cut costs. I think in one area I can probably decrease it by about 75%. It also means we’re locked into the vendor, but right now I can live with that.

My corporate objective is crunching along. I’ve got access to hundreds of data points, and now I need to figure out how to make them match properly. The model I’ve developed works on a small scale, but unfortunately at a larger scale things stop working. I think it means we’ll need to flex some rules, and that really disappoints me. I was so confident that if I could just find the right…configuration, the right algorithm, this intractible problem could be solved. But it can’t. At least, it can’t at the moment. There are a few tweaks I can still make. I’d like to start by flexing what’s disqualifying, and additionally figuring out what to do with people who have really strict constraints. There’s more work to do there, and it remains really interesting.

I’ve got an interview for a promotion on Monday. Wish me luck.