Other Substacks I read write celebratory articles when they get to 10^n subscribers, but I can’t help but be different. I am celebrating 96 subscribers! Thank you all. Your readership means a lot - it is encouragement to keep writing and keep thinking, to keep trying things out outside of the enclosed space of my own skull.
If you’ve just joined us, or have been reading since the beginning, it is time for some (unasked) questions and answers.
Q: Is this Substack ever going to have a paywall?
A: Tell you what - the day I get to 10,946 subscribers (the 21st Fibonacci number!) then I will think about putting in a paywall. At the moment, while I hope you all value what I write, I cannot think of what I could say that would be worth putting behind a paywall. If I figure out a way to consistently beat the returns on the S&P500 index, then maybe. Perhaps, if the comments on this Substack ever get unwieldy, I could paywall subscriber-only posts for discussions.
*Examines the ghost town that is the comment section over the past year* That will be a while. So in short, when I get to 10,946 subscribers, I’ll think about it.
Q: Why don’t you write about (contemporary issue) X?
A: I’m glad you asked, conveniently timed questioner I made up! I do plan on writing about a contemporary issue very soon, in fact within a day or two after I post this.
That issue is going to be machine learning, artificial intelligence, and computation in general. Many people are doing so now, particularly on Substack, but the study of philosophy has given me the courage to say “they’re all doing it wrong…”
Q: But I read this Substack for the explanation of the obscure in terms of the esoteric.
A: That’s not in the form of a question, imaginary voice in my mind, but your point is well taken.
Someone long ago once described an ideal dinner party conversation. To paraphrase, they said “superior people talk about ideas, mediocre people talk about events, inferior people talk about other people.” He did not, from I can gather, mean that talkingly exclusively about those topics denoted superiority or inferiority. Perhaps it might be better paraphrased as “superior people [primarily[ talk about ideas, mediocre people [primarily] talk about events, and inferior people [mostly] talk about other people.” When you put it that way, I know whom I would rather be seated next to at a dinner party.
What does this have to do with the new, somewhat contemporary focus of this blog. Consider three ways to talk about AI:
(Superior) The analogies between weather forecasting and the nature of large language models like GPT-3 are thought provoking
(Mediocre) Open AI got a new infusion of funding from Microsoft. Does this mean that the open nature of their ML research is in doubt?
(Inferior) Person X said that GPT-3 exaggerates harmful stereotypes, but Person Y said that they are wrong, and that its potential to affirm the biases of [people we don’t like and think are morons] is more worrying. In the current intellectual climate, Person X has more clout, so I am on their side, which makes us both right.
This Substack is going to aim to be as much as possible to supply the first rather than the second or third type of thoughtful commentary. I am not at all interested in who said about whom when. Nor, I hope, are most of you.
Q: Why doesn’t this post have a wild, AI-generated piece of art on top, or an obscure but still obliquely apropos quote?
A: The reason for both is that I am figuring out how to use Midjourney effectively. Look forward to (somewhat) less oblique and (somewhat) more representative ML-generated artwork in future posts.
The oblique quotes will continue to be human-generated and human-provided. I just couldn’t think of one to use before I wanted to post this one.
Behold:
"For a period in the mid to late 1940s, Robert Oppenheimer took over playing iconic film detective Philip Marlowe while Humphrey Bogart took time out from his Hollywood career to pursue his passion for nuclear physics.”