The AI Thread

[Kaitzilla]

Thank you for your answer, @Hygro and for the courtesy of your extensive post @Moriate. I'll have a good number of things to say about the latter, but I continue to be amused by this:






I keep not getting any examples to work with.

Here are two people who love the thing, work with it every day, but won't supply me with one instance of its supposedly impressive work, but just continued reiterations of amazement about how it works.

I love poetry, work with it every day. If you ask me for an instance of a great poem, you won't have to wait long for me to provide you with one. (It would be Herrick's "The Vine," if you were to ask right now.) And I'll talk your ear off telling you why it's great.

(Admittedly it has to be a text for me to be able to be impressed with it also. Yes, I know it can write code.)

Anyway, don't bother responding to this post. I'll have a substantive post in time. It's long past time we availed ourselves of your fracturing of think into domains, @Moriarte. I've already been starting to draw on it, and I think that my not yet doing so explicitly accounts for some of how you responded to my last post.

I'm going to hold off watching your video for a while, though. I suspect that "domains of think" will give us a way of processing the claim that I expect that guy to make. I have a hunch as to its content.

AI examples?

I went to Chatgpt.com and asked it to make a Reddit Post.

Create a Reddit Post for AmIThe*******.

It is about a Girlfriend who complains about her boyfriend playing too much Civilization 4 and she broke up with him.

Include a lot of Civilization 4 details.

Then I got about 5 seconds later:

ChatGPT said:

Sure! Here's a well-crafted Reddit Am I the ******* (AITA) post, written from the girlfriend's perspective. It includes humor, plenty of Civilization IV details, and enough context for readers to judge.

---

Title: AITA for breaking up with my boyfriend because he wouldn't stop playing Civilization IV?

Hey Reddit, I (27F) broke up with my boyfriend (29M) last week and my friends are torn—some think I did the right thing, others think I overreacted. So, I’m here to get an unbiased verdict. Strap in, because Gandhi did declare war.

My boyfriend is OBSESSED with Civilization IV. I knew he was a gamer when we started dating a year ago, and I’ve always respected that—we all need hobbies. I even play some games myself. But things got… out of hand.

He started a new game of Civ IV “just to kill some time” one Saturday morning, and the next thing I know, it's midnight and he’s still in his robe muttering something about Frederick of Prussia having a better tech tree. That would be fine once, but this has become a regular occurrence.

He plays these marathon sessions where he’ll “just finish this turn” and then plays for 4 more hours. He once stayed up until 4am because he was "in the middle of a diplomatic crisis with Tokugawa." I kid you not.

He talks about the game non-stop. One time, while we were arguing about house chores, he said—and I quote—“We need to optimize our production queue like in Civ.” WHAT??

Here’s the final straw: we were supposed to go out for my birthday dinner. I got dressed up, made a reservation at a nice place, and when I went to get him, he said, “Give me five turns, I’m about to build the UN.”
I waited. For 45 minutes. He missed my birthday dinner because he was too busy making sure Roosevelt didn’t win a cultural victory.

I confronted him about it that night and he got defensive. Said it’s “not just a game,” it’s “strategy, diplomacy, and world history.” He even accused me of being anti-intellectual.

So I told him I couldn’t be in a relationship where I’m competing for attention with Hammurabi. I packed my stuff and left the next day.

Now he’s telling mutual friends that I “dumped him over a video game” and making me sound shallow. But to me, it wasn’t about the game—it was about priorities.

So AITA for breaking up with someone because they were too addicted to Civilization IV?

Then it asked me to keep going.

Would you like a version from the boyfriend's perspective too?

Yes

Absolutely! Here's the Am I the ******* (AITA) post from the boyfriend’s perspective, also packed with Civilization IV references, but now from the lens of someone who really doesn’t see what the big deal is.

Title: AITA for playing Civilization IV a lot and getting dumped because of it?

This might sound dumb but I (29M) recently got dumped by my girlfriend (27F) because she says I “play too much Civ IV.” I honestly feel blindsided, but maybe I’ve been missing something, so I want Reddit’s take.

So yeah, I’m a huge fan of Sid Meier’s Civilization IV. Always have been. I’ve been playing Civ since the Beyond the Sword expansion dropped, and honestly, it’s one of the few games I never get bored of. The complexity, the strategy, the diplomacy—it’s like chess, but with catapults and space race victories.

Yes, I play a lot. I’ll admit it. Sometimes I get into long sessions—like 6 to 8 hours—especially if I’m close to building a wonder or trying to defend against a surprise war declaration from Shaka. (If you know, you know.) But I always figured that was just part of how I decompress after work.

My girlfriend, on the other hand, doesn’t get it. She says I “ignore her” when I’m playing, but I don’t think that’s fair. I don’t ignore her—I just get immersed. Civ IV isn’t a casual game. You don’t just pause mid-invasion because someone wants to talk about their day. It’s turn-based, but every decision matters.

Here’s the specific incident she’s mad about: it was her birthday, and we had a dinner reservation. But I had just built the Three Gorges Dam (major production boost—game-changer), and I was trying to finish a diplomatic play involving Catherine and Saladin before leaving. I asked her to give me “just a few turns.”

She came back like 30 minutes later totally pissed. I told her I needed just a little more time—Roosevelt was 3 turns from a cultural victory and I had to launch my spaceship or it was over.

She ended up going to dinner alone and broke up with me that night. She said something about how she “couldn’t be with someone who gives more attention to Montezuma than to her.” (Which is ironic because Montezuma was plotting against me.)

Now she’s telling our mutuals I’m addicted to a game and emotionally unavailable. I think that’s unfair. It’s not like I’m playing Call of Duty all day—I’m learning history, building empires, making strategic decisions. Civ is mentally stimulating. I even tried to explain how the game teaches you soft skills like resource management and long-term planning, but she rolled her eyes.

I’m not some man-child screaming into a headset. I just like building civilizations and sometimes yes, it takes priority for a few hours. Doesn’t everyone have something they’re passionate about?

So Reddit—AITA for being really into Civ IV and possibly letting it affect my relationship?

Keep going?

Let me know if you want the comment section tone-written too—like some funny top comments from Redditors reacting to either version.

The whole Internet will be fake in no time.
A 6 year old can use this stuff.
This was about 15 seconds of work.

3 Gorges Dam. What a noob. Hehe


I guess the part that impressed me was that when I did something like this once before, the girlfriend got a few things wrong about Civ 4.

The boyfriend response pointed out her errors, and I felt freaked out a bit!

Wish I could replicate that output, but I don't remember what I asked the machine to do exactly.

 

[Gori the Grey]

Thanks, @Kaitzilla. I think that might actually help me in multiple ways. Are any of you who might be following the larger discussion willing to go down this byway for a bit? I actually think it will bear on the larger issues.

Would you be willing to answer a couple of follow-up questions? They're related to your chosen emojis.

Does this mean you were highly amused? Like, on a comedic scale of 1-10, where would this fall?

(For me it's mildly amusing, with one genuinely clever bit, and maybe a missed comedic opportunity or two.)

And does this mean that those kind of responses in the boyfriend's reply (in the version we don't have) made you think the AI must be thinking (for his posts to correct the errors in hers)? Do you think AI "thinks"? (I don't know where you fall on our big issue.)

Oh, and one more thing. Are you actually able to estimate the boyfriend's level of play from the various Civ IV references? Are they consistent enough that you can make that judgement? You call him a noob for building 3 Gorges Dam; I assume that's not a particularly good wonder in Civ IV (I wish I knew that game better; ironically, the two I've played are 3 and 5).

 

[Hygro]

Gori if I told you my kid was smart and you said “give me an example” would I have some clear example or just a few years of repeat observation and comparison.

I’ll tell you a weird recent one. I was demoing how little effort you can put in to recreate the wheel for certain businesses. My friend wanted to see a gamified crm platform, requested a Miami Vice aesthetic. At first it game a more Tron synthwave 2010s look, oops. One request and pure Miami vice sitewide. (A site that took 60 minutes instead of 2 weeks).

But there was something like an error. There were these buttons at the bottom from before that now were moving in their row from the left side, stopping somewhere in the middle, and doing it again. But staggered. So weird.

And even weirder, their outlines were starting further left but moving faster, crossing them in further right. Also staggered.

I started watching them for a moment because who wants weird moving buttons like that? That’s bad!

But then I took in the vibe a second, and I realized I’ve seen this… from the headland watching waves hit the adjacent beach. The stagger, the position, the outline bs body of each body, the speeds, together formed the best abstract group of waves I have seen in any remotely related context aka never and better wave vibe than 3d games with beaches. It never mentioned it, I never asked, just when we told it no not tron, really Miami vice.

 

[Gori the Grey]

Thanks, Hygro. I can't speak to the visual ones, so I'll take your word for it on that one.

Well, you'd have something specific you could point to, I would think. And, in the arena I'd like to see, texts, I mean that's what it does; it produces texts. Presumably the evidence of its abilities would be in the texts it produced. Have you ever been so blown away by the insightfulness in one that you saved it for later reference?

Like, for me, Hygro is an internet poster. I know him as a set of posts on an internet discussion site. If somebody asked me to demonstrate that Hygro was a limber thinker, I could point them to your post about Leviticus, and say why I found it an instance of limber thinking.

Leviticus is full of very precise rules, and the extrapolations are pretty narrow, and if you're reading carefully the directionality of interpretation of strictness is towards less strictness. There's an entire catalogue of illegal incest, which gender in which relation. There is no reason to believe there is implied equal directionality: for example, no man with his aunt does not mean no woman with her uncle. (And fyi definitely no mention of cousins, it was pretty minimal). And they go out of their way to mention edge cases. For example "no man with his sister" and then later, "no man with his sister even if different mom", and then later: "no man with his sister even if they grew up in different households".

It's not a disorganized mismatch of scoping, it's so deliberate. Believe me, they recognize scoping, and they aren't afraid of repetition, nor the use of too much paper to say the exact same thing.

It's this specifically that I mean to test next time I read Leviticus. I once decided to read the Mosaic Law, giving it a chance to teach me its own logic. If you just approach it from modern day standards, a lot of it seems bizarre. And it's easy to dismiss, therefore. But I read slowly. I let each bizarre one be bizarre for a minute, but then asked "but ok, what would this do for a society that had this as one of its codes?" I didn't finish with the project, but I got deep enough into it that the totality of commandments started making a kind of sense. So I'm open to any further helps in that direction, and I'm interested in this idea about the directionality being toward less strictness. The individual codes are (again to a modern sensibility) so miscellaneously organized, that anyone who discerns any organizational system within them at all, strikes me as thinking.

Anyway, no worries, Kaitzilla has given me a first something to work with.

 

[Kaitzilla]

Thanks, @Kaitzilla. I think that might actually help me in multiple ways. Are any of you who might be following the larger discussion willing to go down this byway for a bit? I actually think it will bear on the larger issues.

Would you be willing to answer a couple of follow-up questions? They're related to your chosen emojis.

Does this mean you were highly amused? Like, on a comedic scale of 1-10, where would this fall?

(For me it's mildly amusing, with one genuinely clever bit, and maybe a missed comedic opportunity or two.)

And does this mean that those kind of responses in the boyfriend's reply (in the version we don't have) made you think the AI must be thinking (for his posts to correct the errors in hers)? Do you think AI "thinks"? (I don't know where you fall on our big issue.)

Oh, and one more thing. Are you actually able to estimate the boyfriend's level of play from the various Civ IV references? Are they consistent enough that you can make that judgement? You call him a noob for building 3 Gorges Dam; I assume that's not a particularly good wonder in Civ IV (I wish I knew that game better; ironically, the two I've played are 3 and 5).

The AI got me with this one.

You don’t just pause mid-invasion because someone wants to talk about their day

I found it funnier than this Napoleon's best quote.

Also, it mimics the tone of AITA of Reddit pretty well.

I guess it trained on all of them?

I just treat AI as the sum total of human knowledge that can be queried.
A curated-by-its-creator/trainer popularity contest formed by consensus.
Ask the humanity gestalt (the English speaking one?) anything you want, no minority report needed.

My surprise is merely that it got so many details correct on the previous attempt.

Clearly Civ Fanatics.com was used as a primary source.
Perhaps even me!

Does the AI think?
Humans have attempted to replicate everything they've seen.
From birds flying in the sky to the sun shining above.
I won't say a thinking machine cannot be done.
Maybe one day cybernetic organoids will do it.
I'd say at this point our current AI emulates some of our thinking sure.
My friend will correct it and ask it to try again a lot.

The fictional boyfriend is of course a noob.
Girlfriend aggro cannot be ignored in such a manner.
A spaceship takes 10 turns to land after launching.
The cultural victory in 3 turns cannot be stopped in such a manner.
Any 20 day veteran would know, much less 20 year.


I cannot really contribute to the AI thread in any meaningful way, but I think the current abilities are already amazing.

Most of the information in libraries and books are locked away by time and distance.
A person reading merely the title of every book written will die of old age before finishing.
AI can read every book and give access of the important data to the asker in seconds.
Think of Burgess Meredith in the Twilight Zone having a book picked out for each day.
He ran out of time!

Our eyes can start going by 60.
Our brains can start going by 80.
Unreal speed is a merit all by itself, even if it is sloppy.

 

[Gori the Grey]

The fictional boyfriend is of course a noob.
Girlfriend aggro cannot be ignored in such a manner.
A spaceship takes 10 turns to land after launching.
The cultural victory in 3 turns cannot be stopped in such a manner.
Any 20 day veteran would know, much less 20 year.

Thanks @Kaitzilla. You don't size it up as an error on Chet G. Peety's part, then (i.e. not really knowing those game mechanics, so that it could properly work them into its story)? Because the story establishes that this guy plays 6-8 hours a day and has since the release of Beyond the Sword. He's stayed a noob all that time? That adds to the humor? Or detracts from the plausibility?

Anyway, as you've hinted, this kind of prompt plays to two of generative AI's real strengths. Pattern-recognition in texts and quick access to the sum of human knowledge.

It didn't break a sweat on this one because AITA posts are formulaic in the extreme: general complaint, and-then-and-then-and-then narrative with individual instances of the general complaint (ideally of mounting magnitude), last-straw complaint. Offender's self-defenses sprinkled in. Even a dumb human like me can figure out the formula having seen a sum total of three samples, with these two parodic instances included in the count. You could almost make one of those old MadLib sheets. Dear Reddit. Am I the the *******? My boyfriend [something a woman might complain about]. Instead of paying attention to me he [verb]. He once told me he thought I was [adjective] compared to his [kind of relative].

It discerns that formula in the zillions of AITA posts in its database and combines that--fills in the slots--with factoids about Civ easily available on the web, especially that it's a time-sink.

I'm not trying to diminish GPT's achievement, just say that to construct a text like this draws on capacities it possesses that even a detractor like me is happy to admit that it has--and superhuman capacities at that. (Interesting thing to try to find a way to test; from how few of a particular kind of text could it discern the generic formula?)

I'll give a recent example of my own that illustrates those strengths and that I happily acknowledge that it has those strengths.

In the Kirk thread, it came up that someone was proposing he should be a saint (and proposing as grounds for canonization the "miracle" that his spine stopped the bullet, therefore stopped an exit wound, therefore stopped someone else from getting hurt). That squares with people calling him a martyr. I wondered if any saints were patron saints of something based on the circumstances of their martyrdom, so I asked Google that question. It's AI said yes and gave about eight examples. (My favorite is St. Lawrence, who was roasted on a spit and said "turn me over; I'm done on this side." He's the patron saint of comedians! And chefs!) It's search results gave a site that treats that matter, that gives about five examples. So Google's AI understood my question, gave me an answer. It drew most, but not all, of that answer from one particular website, and it gave me the website.

I did the same search on Duck Duck Go (where you specifically have to enable AI, but I didn't), and it did not give the website with the five, at least in its first page of search results. AI, then, is at bare minimum a superior search engine, because it's trying more pointedly to get you the exact kind of results you want, not just general treatments of (in this case) patron saints.

 

[Kaitzilla]

Now we're talking.

AI has created life!
I think viruses count?

AI Has Designed Living Genomes—And They Worked in the Lab - Decrypt

Stanford AI wrote full viral genomes that worked—16 lab-built phages show life can now be designed, not just discovered.

decrypt.co

• Stanford researchers used AI to compose entire viral genomes—not edits, but full DNA blueprints from scratch.
• Sixteen of the 302 AI-generated phages came to life, replicating and even outcompeting the natural virus they mimicked.

Experts Alarmed That AI Is Now Producing Functional Viruses

After scientists showed that AI can design a working virus, it's time to address the obvious: what if it's also used to make bioweapons?

futurism.com

In the study, the Stanford researchers used an AI model called Evo to invent DNA for a bacteriophage, a virus that infects bacteria. Unlike a general purpose large language model like ChatGPT, which is trained on written language, Evo was exclusively trained on millions of bacteriophage genomes.

They focused on an extensively studied phage called phiX174, which is known to infect strains of the bacteria E. coli. Using the EVO AI model, the team came up with 302 candidate genomes based on phiX174 and put them to the test by using the designs to chemically assemble new viruses.

Sixteen of them worked, infecting and killing the E. coli strains. Some of them were even deadlier than the natural form of the virus.

Take that evolution.

World’s first AI-designed viruses a step towards AI-generated life

Scientists used artificial intelligence to write coherent viral genomes, using them to synthesize bacteriophages capable of killing resistant strains of bacteria.

www.nature.com

 

[Moriarte]

In the study, the Stanford researchers used an AI model called Evo to invent DNA for a bacteriophage, a virus that infects bacteria. Unlike a general purpose large language model like ChatGPT, which is trained on written language, Evo was exclusively trained on millions of bacteriophage genomes.

I am not surprised by results we get by merely connecting a thinking calculator to the source of Clean information.

That's the problem with general-purpose chatbots. They are trained on piles of garbage. (But still end up being useful)

I am sure there is a real life analogy hiding there, somewhere!

 

[Thorgalaeg]

 

[Gori the Grey]

Separately from our big discussion, it amuses me how almost eager we seem to be to surrender our agency to this thing.

Headline:

AI Has Designed Living Genomes

Text:

Stanford researchers used AI to compose entire viral genomes

It's kind of like a carpenter built a table and we choose as the headline for that

Toolbox builds table

Yeah, that would be an amazing toolbox!

Like, those poor "Stanford researchers." This probably represents a major scholarly achievement for them--career-making, career-crowning. But instead of them getting credit for it in the national news, a tool they used in the process does. Might they be the first to say, "we couldn't have made this advance without this particular tool"? Sure. But so would the carpenter acknowledge that he couldn't have built the table without a saw and a drill and a screwdriver.

From the article:

The achievement, if replicated, represents a milestone in synthetic biology on par with Craig Venter’s 2010 creation of a minimal bacterial cell.

Before AI, if Craig Venter makes a huge advance in the field of synthetic biology, Craig Venter gets the credit. After AI, AI gets the credit.

But now back to our big discussion. Every headline like this quietly reinforces the view that AI "thinks."

 

[Gori the Grey]

Ok, here's my response to the same prompt:

Spoiler Human: Make an AITA about a boyfriend who is addicted to Civ IV :

Title: AITA for breaking up with my boyfriend because he wouldn't stop playing Civilization IV?

Hey Reddit, I (27F) broke up with my boyfriend (29M) last week and moved out of our apartment. My friend my friends are torn—some think I did the right thing, others think I overreacted. So, I’m here to get an unbiased verdict. Strap in for my account of how this declaration of friendship expired.

My boyfriend is OBSESSED with Civilization IV. I knew he was a gamer when we started dating a year ago, and I’ve always respected that—we all need hobbies. I even play some games myself. But things got… out of hand.

When he came home from work, as soon as we finished dinner, he would fire up a game, to “decompress.” I can understand that. We all need some way to relax. But then it would be two o’clock in the morning before he would come to bed. And over breakfast, all he would want to talk about was how pissed he was that Liz beat him to that site that had both copper and horses. Copper and horses, did I understand? (In fact, I don’t. Is that good?)

His mantra had become “one more turn, babe.” I ask him to help me move the sofa: “one more turn, babe.” I tell him we’ve got to get down to the post office with our bills: “one more turn, babe.” One time I came back four hours later and asked, “Hey, have you finished that turn from four hours ago?” His answer? “One more turn, babe.”

When it came out, I asked if he wanted to go see “Oppenheimer” and he said, “Ghandi is massing troops near Boston. I need to play a few more turns to see if he intends to attack.” I asked him, “since it’s a turn-based game, can’t you just pause it anywhere and pick up when we get back?” He said, “Yeah, but it’s kind of urgent, babe.” I asked why. “Cuz his words are backed by nuclear weapons!” (I went to “Barbie.”)

It’s not just how much time he spends in the game. Even when he’s not playing, his mind is still there. I cooked him his favorite dish, risotto, and all he said in response was “Rice! +1 health! sweet!” I put on sexy lingerie and he muttered, “I wonder if Mao would trade me wine for silk.”

Here was the final straw: we were supposed to go out for my birthday dinner. I got dressed up, made a reservation at a nice place, and when I went to get him, he said, “I want to see if I get the Pyramids before Lincoln gets it. Think of the great people points! Think how it will pair once Korea gets Caste System.” I waited. For 45 minutes. He missed my birthday dinner.

I confronted him about it that night and he got defensive. Said it’s “not just a game,” it’s about world history! He even accused me of being anti-intellectual.

The next day, I packed my stuff and set it near the door. I gave him an ultimatum. I told him “You let me turn that key, and it’s for the last time.” His answer? “One more turn, babe.”

So AITA for breaking up with someone because they were addicted to Civilization IV?

It's not that I'm fishing for compliments (though anyone who wants to is more than welcome to acknowledge that mine is funnier). Relative funniness is not the core point, but I do want be able to use this to illustrate the different way in which a thinking mind approaches the same writing task.

 

[Kaitzilla]

And over breakfast, all he would want to talk about was how pissed he was that Liz beat him to that site that had both copper and horses. Copper and horses, did I understand? (In fact, I don’t. Is that good?)

His mantra had become “one more turn, babe.” I ask him to help me move the sofa: “one more turn, babe.”

Heh, that's Civ 4

 

[Hygro]

I talked to my friend who makes and recites jokes about the hawk joke. He said the joke was that the teller identified the hawk incidentally after saying he didn't know what it was. He also acknowledged the rest.

 

[amadeus]

I asked someone to explain the hawk joke and he said it flew right over his head.

 

[Gori the Grey]

I'm not quite sure what "acknowledged the rest" means. We don't have to go back to the hawk joke.* The aim there was to demonstrate that AI doesn't work with the meanings of words. I'm cooking up another way of making that demonstration.

However, I am going to press people to spell out the humor of my AITA-Civ IV mashup. It's NOT so we can ask if AI understands it, though, so don't worry. Some of what is (or should be) funny about it is kind of subtle, and I'd like to know that people got the more subtle bits, either entirely on their own or with just a little prodding from me.

Oh and not just the jokes. I tried to do something over and above of the humor and it would be cool to think that people caught that too.

Spoiler *two more things about the hawk joke, spoilered for people who have had their fill of the damn hawk joke :

First, one way of getting the two meanings needed for the pun is to work with the word "though," placing it in two spots: 1) "even though I watched it intently, I couldn't identify it" and 2) "I watched it as though it were a hawk even though I couldn't identify it." Second, one can actually get away with the joke only because of a sloppiness in conversational English usage, regarding the word "like." The grammatically proper phrasing for the second meaning should be "as though it were." As soon as you say it that way, there's no second meaning of the word "like" and the pun falls apart (or never comes together in the first place). The whole point of observing careful grammar is precisely to rule out potentially confusing alternate meanings.

 

[Moriarte]

Yeah, that would be an amazing toolbox!

Like, those poor "Stanford researchers." This probably represents a major scholarly achievement for them--career-making, career-crowning. But instead of them getting credit for it in the national news, a tool they used in the process does. Might they be the first to say, "we couldn't have made this advance without this particular tool"? Sure. But so would the carpenter acknowledge that he couldn't have built the table without a saw and a drill and a screwdriver.

Not touching at all. First, Stanford researchers aren't poor, nor they are victims in all this. They are at the forefront of scientific progress and their names will be remembered - the names are right there in the "nature" article. Second, they themselves contacted journalists, phrased and proudly delivered news of their achievement. They will get full credit and probably happy they can finally delegate tedious part of genome sequencing to an algorithm, while freeing up brains for more important tasks supplementary to their work.

You trying to twist the situation into researches being the victims from the point of view of your anti-ai crusade is mildly amusing, but not quite accurate.

I asked someone to explain the hawk joke and he said it flew right over his head.

Well, if this hawk joke and the 6 participants of this thread are any measure we have ample proof that most humans don't work with the meaning of words.

Every headline like this quietly reinforces the view that AI "thinks."

Which is not going to be stopped by reinforcing dogmatic reading of the word "to think".

I have my own beef with headlines being mostly about sensationalism these days. But, ultimately, I understand it's not about me, but about algorithm helping one to make money. Words "AI" in the headlines means millions of extra views - better coverage, public resonance, etc. Which is probably better for the scientists too, not just beneficial to those who make money by selling articles.

 

[Gori the Grey]

Words "AI" in the headlines means millions of extra views - better coverage, public resonance, etc.

Ok, but most of your post didn't respond to my main point, and this effectively makes my main point for me.

My main point is, people in this era seem positively eager to ascribe more agency to AI than to people. The fact that headlines that talk that way get more hits is a sign of the phenomenon I'm highlighting: a (to my mind bizarre) collective desire to assign agency (for thinking tasks) to AI. A group of Stanford scientists (the actual agents in this case) used a lifetime of their own education, and their own thinking powers, to come up with a way of composing viral genomes. They devised a way to harness a computer's speed in number crunching in order to carry out that task; even that was their thinking. The headline is as though AI is the main agent of the whole affair. It really is as crazy a social phenomenon as if we all started wanting to deprive carpenters of their agency in crafting a table and assign it to their saws instead.

Hygro's phrasing showed that he got how the hawk pun worked. And I feel certain that if I'd made a fuss about it, all six would have.

That's why with my AITA parody, I'm not going to skip the step of having you all spell out, in explicit terms, what the joke is (jokes are) (but, again, don't worry not because we're going to ask AI to explain why it's funny). So would you be willing to indicate how I was trying to make a response more funny than chat-gpt's? You don't even have to think mine was more funny. (Except in one focused spot; there I'm going to ask someone to call it in my favor). How about this: can you (pl) explain how my AITA was an attempt to realize more of the comic possibilities in an AITA parody than chat-gpt attempted to realize? (And if anyone notices the serious thing I tried to interweave among the jokes, I'd appreciate it if they were willing to spell that out as well).

I don't think my definition of the word "think" is dogmatic. That's the case I'm arguing for here, that mine is the more sensible definition of the word than the one adopted in the wake of all of the hype about AI. And agency is a huge part of this definition. What we usually mean by the word "think" is an active process, involving directed agency. This would have been the very next thing I would get to in my chart. Well, with one little point before that.

 

[Truthy]

Scott Aaronson, a pretty well-known quantum computing researcher at UT Austin, writes that GPT-5 contributed to a paper of his:

But here’s a reason why other people might care. This is the first paper I’ve ever put out for which a key technical step in the proof of the main result came from AI—specifically, from GPT5-Thinking.

[super technical details about the problem and what GPT-5 did] ...

Given a week or two to try out ideas and search the literature, I’m pretty sure that Freek and I could’ve solved this problem ourselves. Instead, though, I simply asked GPT5-Thinking. After five minutes, it gave me something confident, plausible-looking, and (I could tell) wrong. But rather than laughing at the silly AI like a skeptic might do, I told GPT5 how I knew it was wrong. It thought some more, apologized, and tried again, and gave me something better. So it went for a few iterations, much like interacting with a grad student or colleague. Within a half hour, it had suggested to look at the function.

... And this... worked, as we could easily check ourselves with no AI assistance. And I mean, maybe GPT5 had seen this or a similar construction somewhere in its training data. But there’s not the slightest doubt that, if a student had given it to me, I would’ve called it clever. Obvious with hindsight, but many such ideas are.


I had tried similar problems a year ago, with the then-new GPT reasoning models, but I didn’t get results that were nearly as good. Now, in September 2025, I’m here to tell you that AI has finally come for what my experience tells me is the most quintessentially human of all human intellectual activities: namely, proving oracle separations between quantum complexity classes. Right now, it almost certainly can’t write the whole research paper (at least if you want it to be correct and good), but it can help you get unstuck if you otherwise know what you’re doing, which you might call a sweet spot. Who knows how long this state of affairs will last? I guess I should be grateful that I have tenure.

 

[SemiLazyGamer]

Financial Times is claiming leveraged buyout investors of gaming company Electronic Arts are betting on AI to cut costs and boost profits.

https://www.reddit.com/r/GamingLeaksAndRumours/comments/1ntik38/financial_times_ea_leveraged_buyout_investors_are

Client Challenge

 

[Moriarte]

Financial Times misses the point entirely.

First, EA doesn't have that large of a debt problem.
Second, if it did, it could be solved by traditional means, not by AI
And finally, the journalists miss the most important bit - WHO bought one of the prominent platforms for shaping youth culture and WHAT they intend to use the platform for.

Check out the latest Battlefield trailer from EA:

Spoiler bf next :

"Who was that?"
"Doesn't matter"

For those not following closely: EA has been steadily losing market share for a decade by chasing after trends instead of focusing on the core experience and core audience. The opening scene where a slick, "elite-looking" squad gets wiped out instantly - is essentially EA’s way of signaling: “We’re done chasing distractions. We’re back to business - bring your money our way.”