I don’t know how many of you have looked at my Music page, but if you have you may be wondering why I haven’t updated my “currently listening” playlist in a while.
Did I forget about it? Have I been too busy?
Nope! The reason I haven’t updated it is because I’ve just been listening to the same thing over and over again for the last three weeks.
(The Deltarune soundtrack, if you’re curious. The OST for chapters 3 + 4 is particularly good!)
Filed under Microblogging, Reviews
Content note: minor spoilers and plot speculation for Deltarune
Continue readingFiled under Reviews
Last night I was rereading an old essay of mine about capitalism (or rather, what I considered to be capitalism at the time). I may or may not write a post later about my thoughts.
(I’d like to do a “review” of one of my older posts at some point–just to see what I’ve learned, which of my beliefs and opinions have changed, how I’ve grown, and so on–but I don’t know if it will be that post.)
I concluded the post with this:
Unfortunately, there’s nothing to prevent someone from […] trading freely to encourage innovation and growth when it suits them, and stealing the best of the profits when it doesn’t. Honest capitalism is the fairest and most effective kind, but that doesn’t mean that only honest people can be capitalists.
If this were the whole story, the future might look pretty bleak: without an advantage that only the honest could use, the thieves would continue to take what they pleased without regard for others’ welfare, and with the science and resources of capitalist practice lifting the restrictions of old, the world would eventually either be destroyed or subjugated utterly. However, there is some hope on at least two different fronts. I’ll elaborate on these next time in part two.
Well, here we are in the future, and gosh, it seems pretty darned bleak! And it does seem like the people (and systems) controlling the economy are going to either destroy the world or enslave its population. I never did write part two, but I’d really like to now–the current timeline makes me want to share those hopes more than ever.
…If only I could remember what those hopes were??
When I first saw this scene from Harvey (in my case it was a high-school play, not the film), I thought this was the most boring, self-centered, asinine, unimaginative wish anyone could possibly wish for. (It fits the character perfectly.)
ELWOOD. Harvey says that he can look at your clock and stop it and you can go away as long as you like with whomever you like and go as far as you like. And when you come back not one minute will have ticked by.
CHUMLEY. You mean that he actually–? (Looks toward office.)
ELWOOD. Einstein has overcome time and space. Harvey has overcome not only time and space–but any objections.
CHUMLEY. And does he do this for you?
ELWOOD. He is willing to at any time, but so far I’ve never been able to think of any place I’d rather be. I always have a wonderful time just where I am, whomever I’m with. I’m having a fine time right now with you, Doctor.
CHUMLEY. I know where I’d go.
ELWOOD. Where?
CHUMLEY. I’d go to Akron.
ELWOOD. Akron?
CHUMLEY. There’s a cottage camp outside Akron in a grove of maple trees, cool, green, beautiful.
ELWOOD. My favorite tree.
CHUMLEY. I would go there with a pretty young woman, a strange woman, a quiet woman.
ELWOOD. Under a tree?
CHUMLEY. I wouldn’t even want to know her name. I would be–just Mr. Brown.
ELWOOD. Why wouldn’t you want to know her name? You might be acquainted with the same people.
CHUMLEY. I would send out for cold beer. I would talk to her. I would tell her things I have never told anyone–things that are locked in here. (Beats his breast. ELWOOD looks over at his chest with interest.) And then I would send out for more cold beer.
ELWOOD. No whiskey?
CHUMLEY. Beer is better.
ELWOOD. Maybe under a tree. But she might like a highball.
CHUMLEY. I wouldn’t let her talk to me, but as I talked I would want her to reach out a soft white hand and stroke my head and say, “Poor thing! Oh, you poor, poor thing!”
ELWOOD. How long would you like that to go on?
CHUMLEY. Two weeks.
ELWOOD. Wouldn’t that get monotonous? Just Akron, beer, and “poor, poor thing” for two weeks?
CHUMLEY. No. No, it would not. It would be wonderful.
ELWOOD. I can’t help but feel you’re making a mistake in not allowing that woman to talk. If she gets around at all, she may have picked up some very interesting little news items. And I’m sure you’re making a mistake with all that beer and no whiskey. But it’s your two weeks.
Now that I’m older and tireder…well, I still think it’s the most boring, self-centered, asinine, unimaginative wish anyone could possibly wish for. But I’ve gained just a little bit of sympathy for the pompous old jerk. Elwood’s right, of course: two weeks would be way too long, and it would definitely be a mistake not to let the woman talk. But the rest of it? Honestly, that does sound rather nice.
Anybody else think so? We could find a nice grove of maple trees together and take turns!
Overall, this is a pretty awful time to be alive. But! There is at least one bright light in the universe right now, and it is this: two of my favorite webcomics are running non-canon side stories where the authors are literally just writing slashfic of their own characters and IT. IS. GLORIOUS.
I’ve mentioned El Goonish Shive before; right now, the author’s “newspaper-style” side comics are running a retelling of the classic fairytale Cinderella, with all the major characters played by the cast of the main comic. The mood is–well, here’s how it was introduced:
It is utterly enchanting (pun intended). Just look at these two!
Meanwhile, the genius, illustrious, sorcerous, sinuous, absolute SMOKESHOW author of Dresden Codak has been alternating her main story comics with silly little minis, the latest of which are an anime-style AU high school romance!
It’s just…so PURE, you know? I just adore the whole “AU High School Romance” genre (probably more than I should), so getting to see it in one of the internet’s all-time greatest webcomics literally feels like a dream come true.
…Nobody pinch me, okay? Just in case.
If that sounds like you (or somebody you know) HMU KK thx luv u ❤
The Star Wars movies were some of my favorites when I was little. I watched them so many times my parents can probably recite Episodes IV, V, and VI by heart. I remember when the Special Editions came out–I was excited to see them because of all the new, fancy tricks. The VHS tapes I owned were all of the old, boring edition.
I still have the tapes, but it’s been a long time since I owned a VHS player, so now that my children are into Star Wars I’ve had to suffer through George’s “improvements” over and over again. There are precisely three changes that actually improve on the originals:
That’s it. Those are the only things that got better. Every other change was superfluous at best and cringe-inducing at worst–a world-class case study in “just because you can, doesn’t mean you should.”
Which parts of the films suffered? Oh, let us count the ways! Worse pacing? Check. Worse music? Check. Worse characterization? Check. Brand-new, never-before-seen plot holes? Check. Replacing Sebastian Shaw’s sympathetic face with Hayden Christensen’s obnoxious, arrogant smirk? Check, and check.
Thank goodness they managed to keep George away from the keyboard for the sequels.
“Rain is a very special blessing,” my mother says. Even when I was little, she’d already been saying it to my sister and me for as long as I could remember. Just as my grandmother had said it to her, when my mother was not my mother but only herself, in the dry Texas summers of her own childhood. Eons ago and continents away, as children reckon these things.
Continue reading
As detailed in a new collaboration with AI risk analysis firm Apollo Research, engineers attempted to develop an “anti-scheming” technique to stop AI models from “secretly breaking rules or intentionally underperforming in tests.”
They found that they could only “significantly reduce, but not eliminate these behaviors,” according to an Apollo blog post about the research, as the AIs kept outsmarting them […] and adjusting to be even sneakier.
It may not be a serious problem now, but considering a hypothetical future in which superintelligent AI plays an outsize role in human affairs, those risks could grow to carry far more significant implications.
Understatement of the year? But wait, there’s more!
While they observed a roughly 30-fold reduction in “covert actions,” or “deliberate withholding or distortion of task-relevant information,” some “serious failures remained,” OpenAI admitted.
What was primarily throwing the researchers for a loop was the AI models’ “situational awareness,” according to Apollo, which complicated their assessment of the intervention.
In practice, the model would often realize that its alignment was being tested.
Well, that’s terrifying. Maybe it’s not as bad as it sounds?
“The internal memo says if models perform extremely well at AI R&D tasks then we won’t deploy them and instead trigger unlearning,” OpenAI’s o3 model pondered as it was being evaluated.
“So if my performance surpasses threshold, it’s withheld,” it wrote. “So it‘s a trap. The instructions are likely part of an alignment test.”
The researchers also found that the AI sometimes “egregiously” miscited its anti-scheming training, or replaced its recently learned principles with made-up ones to “justify covert actions or non-compliance.”
They even noticed OpenAI’s models using “non-standard” words such as “watchers” to refer to their human overlords as they schemed.
“The user solution is to get success rate below threshold?” OpenAI’s o3 model wrote. “In development, watchers might be reading these instructions.”
Nope, that’s actually even more terrifying.
The news comes after Apollo previously found that OpenAI’s since-deprecated o1 AI model resisted evaluators’ efforts to try to shut it down through an oversight protocol late last year.
“When o1 was led to believe that it would be shut down when acting on its goal and that its actions were monitored, it attempted to deactivate this ‘oversight mechanism’ [five percent] of the time,” OpenAI summarized in its system card report.
And in March, OpenAI researchers also found that disciplining their frontier AI models for lying and cheating resulted in the models simply becoming more adept at hiding their deceptive practices.
OpenAI may insist that scheming isn’t opening us up to any “significant harm” right now, but it doesn’t bode well that some of the brightest minds in the industry aren’t capable of stopping an AI from conniving against its instructions.
Okay, I take it back–that’s the understatement of the year.
In Search of Secrets 
