Photo Credit: Cubix.co
The company I work for has a “platform powered by ML/AI“ AKA two of the hottest buzzwords in the Valley. Just tacking that bit on will secure at least a couple rounds of VC funding. Others include AR/VR, IoT and blockchain. So you can bet your ass I found a way to weave that hype into a post. HMU for investment opportunities.
As previously mentioned, I am decidedly “not about boyfriends rn”. But, because I like attention and intercourse, I am definitely still dating. However, I think there’s a bit more to it than that.
Someone (okay, a fuck buddy) once asked me,
If you’re not dating to even maybe meet a future boyfriend, isn’t that just wasting everyone’s time?—J4 (Four?! holy shit, I must really dig this first name or something)
First off, pretty ballsy of him to say that and jeopardize our very clean arrangement— I would enthusiastically endorse the multiple orgasms I experienced as solidly not a waste of time. Second, so long as the two parties can be truly honest with what they’re looking for, and (this is a huge caveat, disclaimer below) respect each other as adults, it’s a great way to learn.
Let me flex for a minute with some jargon I’ve picked up. There were those three months I spent working on a consulting project for an ML startup while in b-school, so that basically makes me an SME.
Hear me out: Dating can be likened to certain ML models—specifically, unsupervised learning. According to Mathworks, unsupervised learning is “a type of machine learning algorithm used to draw inferences from datasets consisting of input data without labeled responses“. The most common unsupervised learning method is cluster analysis, which is used for exploratory data analysis to find hidden patterns or grouping in data.
Swap humans for machines and you lose the “M” in ML, but that’s dating in a nutshell. We’re left to form our own inferences from our data set, and everyone is kind of making shit up as they go. For instance, ghosting after a second date with no explanation counts as “data without labeled responses” and that dude probably falls in the “fuccboi” cluster. We create archetypes of different “types” of guys (or girls), and follow sexual scripts for what to expect after a third date.
But even more basic than ML algos is just sound understanding of statistics. Statistics 101 tells us to understand whether our findings are statistically significant . We should be wary of sampling error in our analyses. How can we be sure the variables included in our models are truly explanatory, and what can we chalk up to insufficient sample size or variability in the underlying dataset?
In layman terms, I never had a ‘slut phase’ in college. I didn’t date in high school. And I think I have some weird avoidant/anxious attachment issues with each of my parents that’ve spilled over into my platonic and romantic relationships. Whatever the case, I’ve not yet developed a sufficient sample size to gather enough evidence to develop any solid conclusions for how this stuff should work.
On the upside, I’ve had to start asking myself some questions that frankly, I wouldn’t have the self-awareness to even address earlier on in life. What do I find attractive? What do I value in someone else? How do I want to be treated? How should I treat others? And is there a gap between that ideal and my actual behavior? If so, why? This requires terrifying levels of honesty.
So when people ask, “What are you looking for?” I try and be upfront that I am looking for “something casual” (saying I’m “building ML models” doesn’t sound as smooth). Of course, I will always have to contend with the core tenet of good analysts everywhere: Garbage in, garbage out.
Your models are only as good as the data you feed into them. And part of what makes this process messy is that I can only really control my side of the equation.
But I can say that whether you’re a one night stand or we’re planning something a month out, I’ll likely give you the respect most people deserve.
Disclaimer: Renowned blog on the subject, “Towards Data Science” notes that “acquiring training data for our machine learning work can be expensive (in man-hours, licensing fees, equipment run time, etc.).” This point holds true for my proposed approach to dating; gathering data is costly. I might be methodical, but I’ve certainly had my own weird hangups and anxieties, and my feelings hurt in the process. In addition to the whole casual thing, I’ve also started telling people, “I just want someone to be nice to me.”
Aspiring data scientists, I will definitely go on a date with you. In exchange, please make a training data set available via simulation.