At the sort of reverse withdrawal stage when one is actively looking for absurdities in one’s ongoing action in order to justify inaction instead, ergo at the very moment I was composing this entry, I was thinking about how maybe writing is now obsolete anyway because algorithms like GPT2 exist and are only getting better. So I fed some prompt to the algorithm here and indeed there are times when it spits out better gibberish than I ever could so for a moment I felt vindicated I found the perfect excuse for inaction.
Doing so led to several other thoughts, however, providing an opposite excuse to continue writing. Like why did the algorithm need a prompt and would it be able to write something unprompted and if so what would it write? Of course it could and it would pick something random from its training data.
Now the initial reaction to this is to compare what the algorithm is doing to how we think. Humans seem to also just respond to prompts, albeit often physiological. When thoughts randomly pop up from the synthesis of our training data i.e. our experience, we appear as if coming up with something original. Down this path, one might say either GPT2 is thinking or maybe it is the case that we think we’re thinking but like GPT2 we really aren’t.
But that’s a repulsive thought. GPT2 is more plausibly just writing for the sake of having written something while humans are writing as a self-reflection. But then as I write that how sure am I GPT2 doesn’t have a self, or that I do? Pretty soon this whole train of thought devolves into asking what a self is.
At this point, I have to confess yesterday I wrote the following: “One doesn’t know the shallowness of one’s perceived depth unless measured through writing. I am perfectly aware how self-serving this is in the context of this exercise, but thoughts are not thoughts unless written.”
I have no intention to even attempt answering what a self is or what a thought is because then I would be out of my depth pretty quickly. So instead the question is, was the reason I wrote all of the above just so I could use that aphorism for which I apparently developed a day-old attachment? Not initially. The first two paragraphs above directly contradicting what I wrote yesterday attest to that.
But then after realizing I have those two sentences and that I would like to include them in this entry I started writing about how I think I’m thinking instead, because in that context the artifact would then become relevant, and now I’m writing about how I think about that, to provide a runway for this whole writeup to land on.
Maybe that’s what separates humans from GPT2. We don’t just think, we also think about our thoughts, and sometimes about how we come up with them, including instances when we think writing down our thoughts make us think better by virtue of seeing our thoughts more clearly. But then maybe that’s also not entirely non-transferrable to an algorithm.
This post has gotten too long. So much for beginning with coming up for an excuse not to write. Enough for today.
Some things I noticed 01/05/2020:
- mema, the Gen Z Filipino slang short for memaisulat, memagawa, memasabi seems to be in the same ilk as KSP, but more politically correct, mocking just the action as opposed to the whole person of the perpetrator
- materials for niche educational topics on Youtube are getting better, such as this playlist containing a video recommended to me by the algorithm
- tried the Calm app; so much meta-commentary I can’t even; but on the surface I found the scenes feature much better than the unnatural white-noise generator I was using so there’s that