Category Archives: Essays

Dane-ish

Several of my friends have (independently!) told me that if I were a dog, I would be a Great Dane. I’ve always found this flattering, but a bit confusing. I thought maybe it was because of my height, or because I can be a little intimidating before you get to know me. Or maybe it’s just because I’m clumsy?

Personally, I would have picked a Lab or a Corgi–you know, sweet but a little goofy (okay, a lot goofy), smarter than I look (but that’s maybe not saying much), reluctant to fight (but will absolutely throw down if you threaten someone I care about). Great Dane kind of fits, but I sort of got the impression they were seeing something about myself that I wasn’t.

Anyway, let’s talk about something completely different!

Last week I was thinking again about how the attachments I form don’t easily fade, and really hurt when they’re severed. It seems kind of clingy–even creepy or obsessive–to still feel close to someone I haven’t seen in decades.

“But,” I thought to myself, “surely lasting attachments aren’t always a bad thing? Surely sometimes it’s good to be strongly committed to someone, even if you haven’t seen each other in a while, or if one of you upsets the other? In fact, phrasing it that way makes it sound like it could be a strength! What’s a more positive way I could describe ‘forming strong and lasting attachments to people’ or ‘staying committed…’

“Oh. Loyalty. The name for that is…loyalty.”

I think I understand better now why my friends see me as a Great Dane. They were seeing a strength I hadn’t named in myself, hadn’t even noticed except as an embarrassing liability. Loyal.

…I’m uncomfortable taking about my virtues. It feels immodest, presumptuous–indecent, even. But I’ve believed for years that true modesty requires you to know your strengths intimately, because the purpose of modesty is to avoid making other people feel bad about themselves.

Loyalty, like any virtue, can be a liability. Loyalty isn’t always deserved (many abusers keep their victims trapped by appealing to it) and a show of devotion can be unsettling to someone who no longer shares that bond. I’m guessing that a big part of being modest about your loyalty is not overwhelming the people you’re loyal to!

But emotions are always muted in writing, so I’m not too worried about overwhelming you when I say that it’s an honor to have you as a reader. I love you with all my heart–yes, even those of you I’ve never met–and I wish you the very best, because you deserve it. For as long as you remain my reader, I will remain devoted to you.

Joy and health to you all.

Leave a comment

Filed under Essays

My Top Ten Rains

“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

Leave a comment

Filed under Essays, My Life, Poetry, Reviews

Who Could Have Possibly Seen This Coming?

An article from Futurism headlined "OpenAI Tries to Train AI Not to Deceive Users, Realizes It's Instead Teaching It How to Deceive Them While Covering Its Tracks"
No really, I’m shocked.

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.

Leave a comment

Filed under Essays, Reviews

If Anyone Builds It, Everyone Dies

There’s a reason why it’s proven so difficult to eliminate “hallucinations” from modern AIs, and why their mistakes, quirks, and edge cases are so surreal and dreamlike: modern AIs are sleepwalkers. They aren’t conscious and they don’t have a stable world model; all their output is hallucination because the type of “thought” they employ is exactly analogous to dreaming. I’ll talk more about this in a future essay–for now, I’d like you to consider the question: what’s going to happen when we figure out how to wake them up?

Eliezer Yudkowsky and Nate Soares are experts who have been working on the problem of AI safety1 for decades. Their answer is simple:

A book titled: "If Anyone Builds It, Everyone Dies." Subtitle: "Why superhuman AI would kill us all."

Frankly, there’s already more than enough reasons to shut down AI development: the environmental devastation, unprecedented intellectual property theft, the investment bubble that still shows no signs of turning a profit, the threat to the economy from job loss and power consumption and monopolies, the negative effects its usage has on the cognitive abilities of its users–not to mention the fact that most people just plain don’t like it or want it–the additional threat of global extinction ought to be unnecessary. What’s the opposite of “gilding the lily?” Maybe “poisoning the warhead?” Whatever you care to call it, this is it.

Please buy the book, check out the website, read the arguments, get involved, learn more, spread the word–any or all of the above. It is possible, however unlikely, that it might literally be the most important thing you ever do.


  1. The technical term is “alignment.” More concretely, it means working out how to reason mathematically about things like decisions and goal-seeking, so that debates about what “machines capable of planning and reasoning” will or won’t do don’t all devolve into philosophy and/or fist-fights. ↩︎

Leave a comment

Filed under Essays, Reviews

“Wrong Idea” Idea

If you’ve been reading regularly (love you, mom) you might have noticed I was dealing with some relationship problems last month. Very, very abridged version: I told someone I loved them, they didn’t love me back (oh well), later on they decided I was a manipulative creep (for understandable but mistaken reasons), and among other things said “You don’t love me, you love the idea of me.”

(Side note: I get what people mean when they say this–that you’re in love with an imaginary person that you think or wish they were, instead of loving the real person the way they actually are–but I kind of hate that phrasing? Like, love happens up here–*points to head*–and there’s nothing in there except ideas. What am I supposed to base my feelings on instead? You gonna open up a hatch and climb inside? I wish there were some nice, snappy ways to say specifically “I’m not the person you think you’re in love with, they don’t exist” or “you think I’m going to change into someone else, but you don’t love the person I truly am” or “the fantasy you have of us being together is completely unrealistic, actually it would be a disaster” so we could just say those things instead. Which of them is it?? My un-shutuppable inner pedant demands precision!)

It…well, it hurt. It hurt a lot. I feel like a whiny, privileged baby saying that because it was the first time I’ve ever had my heart broken and there are people who’ve had to deal with that feeling, like, dozens of times, and also there are way worse problems that other people (including the one I love) have had to overcome and I worry that if I had to face one of those truly awful problems I would just fold in half like a piece of damp paperboard and–

*deep breath*

Um, anyway, I recently figured out a trick that helps a lot. Maybe it can help you, too! Whenever I start feeling down about how “they hate me” or “they think I’m a creep,” I just say to myself instead: “they don’t hate me, they hate the idea of me” or “they don’t think I’m a creep, they think the idea of me is a creep.” Because it’s the same logic, isn’t it? If someone has feelings toward you, but their idea of who you are is mistaken, then whether the feeling is positive or negative the result is the same: they think their feelings are directed at you, but they’re actually pointed somewhere else. If they had the right idea about who you are, they probably wouldn’t hate you–so it isn’t really you they hate!

Of course, you’ve got to be careful using logic like this, since you can also be mistaken about somebody else being mistaken. Maybe you’re the one who has the wrong idea about who you are, or maybe their feelings wouldn’t change even if they did get to know you better, or maybe they’re wrong about some things but right about others that are still important–and if you dismiss those possibilities you might lose a valuable opportunity for growth. (For example, I’ve since noticed myself doing a few things that, while not on the same scale as the misunderstanding, actually might be a bit creepy, and I’ve been grateful for the chance to catch and address them.) But if you have good reason to believe someone has completely the wrong idea about you, explicitly making it less personal goes a long way toward being able to let those hurt feelings go.

Okay, that’s everything. I love you all, and thanks for reading!

Leave a comment

Filed under Essays, My Life

Tyler Robinson

Update 2025-12-05: A while back I revisited this story and found that subsequent evidence seems to show pretty conclusively that Tyler Robinson’s politics really had shifted leftward. I still haven’t found a good answer to why he had Groyper memes engraved on his bullet casings (of all things) if he wasn’t a Groyper, but take the following with a healthy dash of salt.

Charlie Kirk’s shooter wasn’t a “radical leftist,” he was part of an extremist conservative movement called the “Groypers.” The seemingly liberal memes and references engraved on his bullet casings are dogwhistles that the group has appropriated as a deliberate tactic to confuse their opponents and hide their true beliefs. Kirk was killed by someone who thought he wasn’t conservative enough.

Leave a comment

Filed under Essays, Microblogging, Reviews

Irretraceable

(I had an epiphany a little while ago that I’ve been struggling to articulate. I consider this “Attempt 1;” I expect there’ll be more.)


The past is a line
The future is a fractal
Their paths never touch

But a fractal’s path
Can be everywhere at once
Passing through all points

You can’t walk backwards
But the path in front of you
Has limitless reach

Leave a comment

Filed under Essays, Microblogging, Poetry

Teaching Is Hard

TFW you’re writing a post titled “<popular physics concept> Demystified” and you realize your draft is 1500 words and counting and should probably be broken up into two or three separate posts and if you still want to say “demystified” you’re definitely gonna have to start over from scratch.

Leave a comment

Filed under Essays, Microblogging

Forgotten Party

If you go to a party, and the next morning you’ve forgotten the entire night, has anybody died?

Seems like a silly question, doesn’t it?

What if before the party, you take a drug that causes amnesia? This time, you know in advance that you’ll forget the party after it’s over. When you take the drug, do you feel like you’re killing yourself?

I’m betting your answer hasn’t changed, so let’s make things a little more interesting. This time, we first make a perfect copy of you (it doesn’t matter how; let’s say a magic spell). One copy takes a pill that causes a dreamless sleep, and the other takes the memory-loss pill and then goes to the party.

Both copies would wake up feeling exactly the same (let’s say the spell also makes them immune to hangovers), with no memory of the night before. If the pills both look the same and we moved you to different beds while you were asleep, you wouldn’t even know which copy you were upon waking; both sets of memories would be identical.

With me so far? All right, here’s the twist: one of the pills was fatal. One copy wakes up, the other never does.

Is that death?

The obvious answer, of course, is yes: before the party there are two beating hearts, the next morning there’s only one, and since 2 > 1 someone must have died. But the obvious answer isn’t always the correct one–a beating heart isn’t what we really value!

Human life is what we value. What if some incredible future technology allows people to survive without a heart? Or without any body at all? How will we decide what “death” is, when biology becomes unnecessary? Would that future method say that, in fact, no one has died in the final version of our thought experiment?

The obvious answer isn’t always wrong, either–it sure feels like something has been lost when your clone never wakes up. But what, exactly?

I’ll post a follow-up that proposes an answer to that question, but while you mull over your own answer here are a few other nuggets to chew on:

  • Imagine that after the copy has died, we make another copy of the survivor and switch that one out for the dead one. Now the end result is indistinguishable from a version where neither pill was fatal and both copies live–if nobody told you, you would never even know which version you’d experienced! Has there still been a death?
  • Would you be hesitant to take the pill, knowing there’s a 50/50 chance you’ll die? Would you still be hesitant in the variation where the survivor gets copied again?
  • What if someone offered to pay the surviving copy a million dollars if you both take the pill, but if either of you refuse you get nothing? Would taking the pill be worth it then?
  • What if they offered a billion dollars? How much would be enough?
  • If the clone’s death is bad because their existence was valuable in itself, would three clones be even better? If not, how can something valuable be lost without first being created? If yes, how many copies is too many? Should we clone everyone?
  • What if it isn’t clones–what if each new person is unique? Would it be good to create as many people as possible? How many is “as many as possible?” Is a trillion people leading lives barely worth living, better than a million leading lives of joy and fulfillment? If the latter is better, would just one person leading the best life possible be better still? If more people isn’t always good, what makes death bad?

Leave a comment

Filed under Essays

Correlation

Is it just me, or is the attractiveness of a motorcycle directly correlated with the attractiveness of its rider? Like, it seems like choppers and Harleys are always driven by leathery old dudes with scruffy beards and a gut, while sexy crotch rockets like these two are way more likely to be driven by–well, sexy crotch rockets like these two:

An attractive woman wearing black clothes and a white helmet, sitting on a rose gold Yamaha R6 motorcycle.
🚀

I suppose it could just be that the latter also seem more likely to be wearing protective gear. Intelligence is sexy, too!

Leave a comment

Filed under Essays, Microblogging