Predominantly Erroneous (Exohedron nonsense blog)

"Speculative fiction, or as some people call it, finance"

 

I just realized that I don't actually remember when I first learned the Fourier transform. Some time in high school? But I don't actually remember it.

 

We can talk about Doylist explanations and Watsonian explanations for various things, but the real question is: what does Sherlock Holmes think is going on?

 

Kind of confused on the 6-month GPT moratorium. There are roughly three problems with the GPTs at the moment. The technical one is prompt injection, which isn't solvable for the same reason that automated moderation doesn't work: you're going to get people actively figuring out how to get around it, and nothing you do will actually make it hard.
A more important flaw is that people are stupid, and the only thing you can do about that in 6 months is to stick warning labels on everything. I suppose how you judge the potential efficacy of that solution correlates to how well you think warning labels work on anything else when there's money involved.
And then there's the jobs problem. This also can't be solved in 6 months, and I would expect that by the time an actual solution gets devised that isn't some form of Universal Basic Income, market evolution and generational shifts will have changed the set of things that are considered jobs enough that the problem gets buried under other things. See for instance the electronic calculator, and most typist positions, and a bunch of early interacts-with-computers jobs that have been replaced by various Microsoft products.

So is the 6 months just supposed to be a stopgap? At least avert the apocalypse for half a year, and then die anyway?

But hey, at least if I want it I can look at a list of people whose opinions on AI I can probably safely dismiss.

 

Realizing that some of the paperwork only serves to give an opportunity to fail to fill out paperwork properly, as everything that might be conveyed by the paperwork is already known to the recipient.

 

Really irritating that restarting doesn't fix the issue, but holding down the power button until the computer turns off and then turning it on again does. Isn't restarting supposed to simulate "turning it off and on again?"

 

One of the big themes of Lord of the Rings is that of estel, the unsupported hope, the hope that is purely faith rather than reason. Again and again throughout the three films, estel is rewarded, by the sudden appearance of allies or exits or even convenient foes. Lord of the Rings is not a fight against evil, but a fight against despair, in the face of not just overwhelming odds but the certainty of defeat.

Poe Dameron's storyline in Star Wars Episode VIII: The Last Jedi is also about hope and despair, about having faith and the consequences of losing it. I'm trying to figure out why it doesn't work for me nearly as well. I think part of it is that it's so blatant, the repeated line Hope is like the sun etc. I don't know, maybe it's just me but it really grates on me when characters repeat the same phrase over and over again. Callbacks are good but should be used sparingly. We heard that line five minutes ago; you don't need to repeat it. Say something new. But even without the repetition, it was so clearly about hope that the characters were talking about it. But also it was a passive hope, a hope where instead of charging forth into peril and darkness, one just waits, waits for rescue, waits for someone else to come. Which might be the harder test of faith but is also really boring to watch.

 

There's a certain level of productivity at which removing sysadmins won't actually make the situation worse.

[Edit]: I got dodgeball'd

 

When suddenly you hear a voice from nowhere saying "Congratulations! You are now exiting the timeloop. Have a nice day." and someone appears out of thin air looking like they've been locked in a metal crate for a year screaming "No! Let me try again! Let me try again!".

 

I should just write a love song. So far I've got "acolyte gives in to despair", "yandere with a sword", "topology puns disguised as break-up song", and now "whining about cosmic horror". Like, I can write a stupid, sappy love song, but when I try for anything else I get bitter nihilism. Which I suppose is entirely unsurprising given my aesthetics.

 

I still love international cooperation (this is still a complaint about timezones)

 

The problem is that, due to their appearances in logic puzzles, we are now primed for unexpected tigers. It is the expected tigers that come out of left field.

 

One of the best things about my neighbors having a Saint Bernard is that I get to see a huge dog. A slightly better thing about them having a Saint Bernard is I get to see them struggling to get this huge dog to move when it decides to just flop down in the middle of the lawn instead of continuing to walk.

 

Problem: tumblr is bad at time stamps, leading to people wildly misjudging when a post was made.
Solution: put a timestamp (or at least a date) into the body of the post itself.
Problem: I would be terribly tempted to lie about the date, thus completely invalidating the mechanism.

 

The problem of having too high a charisma stat and enough intelligence to get away with hallucinatory token-prediction, but not enough wisdom to realize that I am in fact hallucinating.
Please don't ask me for project-management or policy type decisions, I don't actually know any better than you do despite my apparent high-confidence.

 

Going to start telling people that when I say "crypto", I don't mean "cryptocurrency", I mean "cryptozoology".

 

Detective stuck in a time-loop trying to solve a murder mystery but because of his actions the murderer is different each time.

 

I think I'm finally coming to a better understanding philosophy as a discipline. If it should be treated as akin to a humanities discipline, wherein the texts are irreducible objects of study on their own and the goal is not to understand the "underlying ideas" but to understand the people who created these texts and the contexts in which those people worked in, then I understand the value of reading the original texts.
Mathematics, "natural philosophy", (mathematical) logic, get spun off from philosophy as idea-centric, wherein the objects of study are not the presentations but something independent, and the writer assumes a secondary position. If this is the wrong way to view philosophy, then of course I wouldn't understand it, because the idea-centric framework is the way that I understand things. I appreciate the notion of art, history, literature as disciplines, even if the practice baffles me. I suppose then I could come to appreciate the notion of philosophy as a discipline, even if the practice baffles me.

 

My dad forgot that I don't speak French. Fortunately, this wasn't during an attempt to communicate with me in French. He was just trying to get me to test the translator on his phone.

 

Suppose you've been volunteered for a job, only you weren't told in time to do a good job at it and you weren't really given any instructions. Now, it doesn't matter if the job is done particularly well; it's supposed to be a fun, community-building thing. But also it's a lot of effort-per-fun and you're not actually part of the targeted community, and the targeted community is giving you basically no support.

Question: Is it better to do a bad job, which nobody is really checking because it's supposed to be fun, or to not do the job at all, which would be noticeable but you can blame on the people who were supposed to be organizing and supporting this?

On the other hand, I'm not entirely sure even doing a seemingly-good-but-in-reality-bad job is actually possible at this point in time, so the choice might be between doing a noticeably bad job or not doing the job at all.

Unfortunately, I do like a lot of people in the community, and also the community contains several of my bosses.

[Edit] As expected, the quality of my work here was almost entirely irrelevant and most of it was just straight-up ignored, but the fact that I did anything at all ended up being somewhat meaningful.