Category: Tech

  • Tyrannosaurus’s Tiny Arms: Has the Age-Old Mystery Finally Been Solved?

    A CNN article from May 2026 covering the century-old mystery of why the carnivorous dinosaur Tyrannosaurus had such tiny forelimbs set off a lively discussion on 5ch’s Science News+ board. The article cites research suggesting that “the sturdier a dinosaur’s skull, the more its forelimbs tend to shrink” — the idea being that once an animal specialized in killing prey with its jaws and teeth, its arms simply became less important. In the thread, posters pointed out that the same article had already been posted six months earlier, raised objections about confusing correlation with causation, and floated alternative theories — such as the arms being needed to brace the body in an ambush stance — all mixed in with jokes, but touching on real scientific points along the way.

    2026.11.13 Fri posted at 19:57 JST

    (CNN) — Why are the arms of the carnivorous dinosaur Tyrannosaurus rex so small? A study published this past May may finally have solved that mystery.

    This oddest of features on the mightiest of dinosaurs has long been fodder for jokes, and the reason behind it has been debated for over a century.

    Source: cnn.co.jp / Read the original article here

    2AnonymousNov 15, 2026 18:16
    That's because there's no one inside wearing a costume.
    3AnonymousNov 15, 2026 18:31
    Seriously?
    Triceratops looks like it'd have a pretty thick skull too,
    but it's got perfectly normal arms.
    7AnonymousNov 15, 2026 19:09
    Re: #3
    Well yeah, it's got four legs.
    4AnonymousNov 15, 2026 18:32
    Those arms are just for show! The bigwigs just don't get it!
    5AnonymousNov 15, 2026 18:39
    >"Dinosaurs with especially sturdy skulls are highly likely to have especially tiny forelimbs."

    Even a total amateur could figure that out lol
    If a dinosaur with that massive a skull also had sturdy, bulky forearms,
    it couldn't walk on two legs.
    And it'd run slower too.
    If you've got huge jaws and sharp teeth, you don't really need front legs
    to pin down prey. Even a grade-schooler could follow that logic.
    8AnonymousNov 15, 2026 19:10
    Re: #5
    Go look up Allosaurus.
    6AnonymousNov 15, 2026 18:44
    But it's way less romantic if it's basically just a big bird.
    9AnonymousNov 15, 2026 19:17
    Image:
    Allosaurus: King of the Jurassic
    Tyrannosaurus: King of the Cretaceous
    10AnonymousNov 15, 2026 19:31
    Is there any chance it actually had bat-like wings that just didn't survive in the fossil record?
    11AnonymousNov 15, 2026 19:42
    Maybe it's because it was a snake in a past life,
    and it's a bird in its next life, so that kind of tracks.
    12AnonymousNov 15, 2026 19:47
    This lineage evolved further into birds, right? lol Or at least shares an ancestor with birds lol
    13AnonymousNov 15, 2026 20:25
    It's so it doesn't get in the way during the Enoshima Tyrannosaurus Race.
    14AnonymousNov 15, 2026 21:02
    Maybe it's like,
    it uses its arms as a kid, but once it grows up its head gets too big and the arms become useless?
    15AnonymousNov 15, 2026 22:21
    I'd have loved to see a world where dinosaurs survived and coexisted with mammals.
    16AnonymousNov 15, 2026 23:15
    I mean, they're basically chicken-wing fossils, so of course they're small.
    17AnonymousNov 15, 2026 23:43
    Ostrich: has wings
    Emu: wings shrunk to practically nothing

    If you can handle daily chores with just your mouth, having half-useless arms just gets you hurt in a fight.
    Crocodiles don't seem to use their forelimbs for much besides moving around either (though females need them for digging nests)
    18AnonymousNov 16, 2026 00:06
    Basically it's just a ridiculously huge chicken.
    19AnonymousNov 16, 2026 00:30
    They're running this same story again?

    2026/05/26 (Tue) 22:35:22.37 ID:xXUxS+On
    Why are the arms of the carnivorous dinosaur Tyrannosaurus so tiny? Has the age-old mystery finally been solved?
    (CNN) Why are the arms of the carnivorous dinosaur Tyrannosaurus rex (T-rex) so small?
    https://egg.5ch.io/test/read.cgi/scienceplus/1779802522/-100
    20AnonymousNov 16, 2026 00:31
    Trying to become a master head-butter and ended up losing its arms entirely —
    just like Bodhidharma (the legendary monk said to have lost the use of his limbs from years of meditation).
    21AnonymousNov 16, 2026 01:37
    Maybe it's like a shoebill stork?
    22AnonymousNov 16, 2026 02:04
    Arms are just for show!
    Weak people just don't get it, see!
    28AnonymousNov 16, 2026 11:05
    Because the reconstruction is wrong.
    29AnonymousNov 16, 2026 13:27
    Re: #28
    Go get yourself mauled to death by a T-rex.
    30AnonymousNov 16, 2026 13:52
    Looking at a sparrow's skeleton,
    this little guy could totally turn into a Tyrannosaurus given the right conditions.

    https://img06.shop-pro.jp/PA01137/526/product/181982556_o1.jpg
    31AnonymousNov 16, 2026 14:07
    Dinosaurs basically hunted by ambush.
    When prey got close, they'd crouch and wait like a sprinter in the blocks at the 100-meter dash —
    kind of like the tachi-ai stance at the start of a sumo match.
    During that waiting crouch, the arms were incredibly important —
    those small arms were what barely held the body back from lunging forward too soon, and the moment the prey came close, it would let go and dash.
    32AnonymousNov 16, 2026 14:19
    Looking at sparrows and chicks, yeah, they're basically mini T-rexes.
    33AnonymousNov 16, 2026 14:39
    Because it's a preliminary stage on the way to growing feathers.
    34AnonymousNov 16, 2026 14:50
    Even as small as they are, they've probably got enough strength to beat a human at arm wrestling with room to spare.
    36AnonymousNov 16, 2026 15:48
    I see.
    So then it doesn't need legs either,
    and it doesn't need a big body either,
    so… does it just turn into a snake?

    Background and Key Points of This Discussion

    The article behind this discussion was originally published by CNN in May 2026, and as post #19 pointed out, the very same thread had already run on 5ch once before, six months earlier — making this effectively a repost of old news. The explanation offered for the shrinking arms is a correlation — “theropods with sturdier skulls tend to have smaller forelimbs” — framed as a kind of biomechanical trade-off (once the jaws and teeth became the sole weapon, the arms had less work to do). But as post #24 notes, this is still just a statistical correlation; it doesn’t actually spell out the causal mechanism behind why the pattern holds. Elsewhere in the thread, post #31 raises a different functional theory not covered in the article — that the arms braced the body to keep it from lunging forward too early during an ambush. In reality, the role of T-rex’s forelimbs (whether as a hunting aid, for gripping during mating, or as support when getting up) remains a genuinely open question in paleontology. Plenty of folk theories drawing comparisons to living animals like birds and crocodiles circulate too, but none of them amount to an academically settled conclusion.

    ※This article is excerpted and summarized from the 5ch (Science News+) thread “Why are the arms of the carnivorous dinosaur Tyrannosaurus so tiny? Has the age-old mystery finally been solved?“.

  • AI traffic signal control to expand to 8,000 locations across Tokyo — 5ch: “One hack and it’s all over”

    The Tokyo Metropolitan Police Department (MPD) plans to expand AI-based traffic signal control, which is used to predict congestion, to roughly 8,000 locations across the whole of Tokyo by fiscal 2027, the Nikkei newspaper reported. A trial run since fiscal 2023 confirmed a 10% reduction in the length of traffic jams, with the aim of smoothing out logistics. At intersections on arterial roads, the system predicts congestion 30 minutes ahead; at smaller intersections, it detects conditions in real time and switches the signals accordingly. The news set off debate on 5ch, with safety worries (“one hack and the whole transport network is finished”) clashing with doubts about the choice of technology (“wouldn’t sensors be enough without AI?”).

    To ease congestion in central Tokyo, the Metropolitan Police Department will expand the artificial intelligence (AI) control it already uses at some traffic signals to cover the entire city, possibly as early as fiscal 2027. A trial that began in fiscal 2023 confirmed the approach eases congestion, and the department judged it would help keep logistics running smoothly.

    Under AI signal control, intersections where arterial roads meet will have their switching adjusted based on predictions of congestion 30 minutes out, drawn from traffic volume and past data. At relatively small intersections, congestion is detected in real time and the signals swit…(continued in the paid edition, 764 characters remaining)

    Nikkei, November 16, 2026, 5:00 AM

    Source: nikkei.com / Original article here

    2AnonymousNov 16, 2026 08:41
    Get the transport network hacked and it's game over
    25AnonymousNov 16, 2026 08:48
    Re: #2
    That's just as true with old-school signals too
    54AnonymousNov 16, 2026 08:54
    Re: #2
    Wasn't there something like that in Mega Man Battle Network (Rockman EXE)?
    88AnonymousNov 16, 2026 09:03
    Re: #2
    Just station a guy with a flag in the middle of the intersection
    4AnonymousNov 16, 2026 08:43
    If the cloud goes down, do the traffic lights go down too?
    45AnonymousNov 16, 2026 08:53
    Re: #4
    It's obviously running on local AI
    If it isn't, this country's hopeless
    53AnonymousNov 16, 2026 08:54
    Re: #45
    An AI running standalone might go haywire on its own
    93AnonymousNov 16, 2026 09:04
    Re: #45
    …this country's bureaucrats really are hopeless, you know?

    Just this past Friday, during that worldwide Outlook outage, some central government ministry actually phoned around saying, "Delays in submitting daily reports are not acceptable in principle, so be sure to submit by fax or courier by the deadline."

    They probably had bureaucrats working overtime typing all those daily figures, received by fax and on paper, into Excel.
    7AnonymousNov 16, 2026 08:43
    AI messes up now and then — is this really okay?
    If it turns every light green at once, we're done for
    32AnonymousNov 16, 2026 08:49
    Re: #7
    It's not generative AI, you know
    22AnonymousNov 16, 2026 08:46
    Please put this in at the Keio Line crossings
    Some idiotic algorithm has turned it into a crossing that never opens
    147AnonymousNov 16, 2026 09:29
    Re: #22
    Are there still stations like that?
    29AnonymousNov 16, 2026 08:49
    What's the point of putting AI in this, idiot
    Sensors would do the job
    129AnonymousNov 16, 2026 09:17
    Re: #29
    There used to be a lot of signals that turned green via sensor when a car pulled up, but you don't see those much anymore
    37AnonymousNov 16, 2026 08:51
    If this gets hacked, it's going to be a disaster…
    42AnonymousNov 16, 2026 08:52
    Re: #37
    Yeah, that's a real concern
    57AnonymousNov 16, 2026 08:55
    There's bound to be someone who rigs it so the light's always green on their own route
    198AnonymousNov 16, 2026 10:05
    Re: #57
    The push-button signal near my place started syncing with another signal some distance away. It used to change right away, but now that it's linked, it takes forever — some driver probably complained about it.

    Now, after I press the button, I check the other signal, and if it doesn't look like it's about to change, I just walk to the next one.

    End result: cars get stuck at a signal nobody's even crossing at, which is just inefficient.
    62AnonymousNov 16, 2026 08:56
    This year's Kanji of the Year has got to be 愛 ("ai," meaning love — a pun on "AI")
    69AnonymousNov 16, 2026 08:58
    Re: #62
    "Because of love, people must suffer. Because of love, people must grieve. Because of love…" (a famous Gundam line)
    67AnonymousNov 16, 2026 08:57
    Everyone's going on about hacking, but why not just keep it offline?
    79AnonymousNov 16, 2026 09:00
    Re: #67
    Can you even reference traffic volume and historical data (presumably updated constantly) or detect congestion in real time while offline?
    73AnonymousNov 16, 2026 08:58
    It's so dumb having to sit at a red light at night when there's zero cross traffic
    112AnonymousNov 16, 2026 09:09
    Getting stuck at a red light when nobody's coming is ridiculous
    120AnonymousNov 16, 2026 09:11
    Re: #73 Re: #112
    Especially in summer, way more people are out walking dogs or just walking themselves at night and in the early morning hours, so drive with that in mind
    75AnonymousNov 16, 2026 08:59
    Can't these idiots who just heard about it on the news imagine that whatever concern or question pops into their head the instant they hear it has obviously already been addressed? lol
    78AnonymousNov 16, 2026 09:00
    Re: #75
    It hasn't been addressed — that's exactly why there've been leaks all over the place
    The risks of AI have been talked about since the theory was first established
    81AnonymousNov 16, 2026 09:01
    Re: #75
    Given all the recent unauthorized-access incidents, you can't really say that
    85AnonymousNov 16, 2026 09:02
    Traffic signals can also be switched manually, so even if they get hacked, it'll only be temporary
    91AnonymousNov 16, 2026 09:03
    Re: #85
    That "temporary" part is exactly the scary bit
    It was pure panic during the big blackout
    99AnonymousNov 16, 2026 09:05
    Re: #85
    Just like that false earthquake alert recently, being able to deliberately disrupt a wide area even briefly is bad news
    Not everyone slows down at intersections, so depending on the timing, this could cause a serious accident
    98AnonymousNov 16, 2026 09:05
    Re: #91
    Of course it's going to stop during a major disaster
    101AnonymousNov 16, 2026 09:06
    Re: #98
    Which is exactly why even a temporary disruption can do real damage, right?
    108AnonymousNov 16, 2026 09:08
    Re: #99
    Such a negative outlook
    Technological progress requires sacrifice
    If something goes wrong once, you just fix the bug afterward
    116AnonymousNov 16, 2026 09:10
    Re: #108
    Ah, you're the type who's always popping up in AI threads
    Sorry for engaging
    139AnonymousNov 16, 2026 09:22
    Rather than AI, wouldn't ordinary programmed control be fine
    Aren't they just running on timers right now anyway, not that I really know
    141AnonymousNov 16, 2026 09:24
    Re: #139
    I think AI is only used for the traffic-prediction part
    143AnonymousNov 16, 2026 09:25
    Re: #139
    If you actually measure it, you'll see it's slightly different every time

    Background and key points of this topic

    The Tokyo Metropolitan Police Department has been running a trial of AI-based traffic signal control at some intersections since fiscal 2023. At intersections where arterial roads meet, the system predicts congestion 30 minutes ahead using traffic volume and historical data; at smaller intersections, it detects congestion in real time and adjusts signal switching accordingly. Having confirmed this cuts the length of traffic jams by 10%, the MPD plans to expand it to roughly 8,000 locations across the whole of Tokyo by fiscal 2027. The thread split mainly into two points of debate: one was a safety worry — “if it runs through the cloud, couldn’t hackers hijack the signals?” — and the other was doubt about the choice of technology — “is AI even necessary for congestion prediction, wouldn’t sensors be enough?” A common point of confusion was that the “AI” here isn’t generative AI for text or images but a machine-learning model for traffic prediction, and that the signals themselves are designed to allow manual switching even during power outages — several rebuttals in the thread drew on that distinction.

    *This article is excerpted and summarized from the 5ch (News Flash+) thread “AI traffic signal control to expand to 8,000 locations citywide — 10% shorter congestion smooths logistics.”

  • Petition for Former NVIDIA CEO Jensen Huang’s Reinstatement Tops 140,000 Signatures, Musk Among Signers

    NVIDIA co-founder Jensen Huang stepped down as CEO effective November 12, 2026, following an affair scandal involving a woman suspected of being a spy. Despite the fact that whether she was actually a spy has never been confirmed, critics say Huang was forced out regardless — sparking a petition drive calling for his reinstatement. The petition, which has drawn support from current employees, has surpassed 100,000 signatures and stood at 140,000 at the time of writing. Reports note that high-profile figures including Elon Musk have also signed on. On 5ch, discussion threads split between questioning the circumstances of the resignation, its possible ties to China-export and tech-leak suspicions, and how much a petition could realistically do to overturn the company’s decision.

    Petition Calling for Jensen Huang’s Return as NVIDIA CEO Tops 100,000 Signatures, Elon Musk and Others Sign – GIGAZINE

    November 15, 2026, 5:53 PM

    NVIDIA co-founder Jensen Huang resigned as CEO effective November 12, 2026 local time, after an affair scandal involving a woman suspected of being a spy became public. However, whether the woman was actually a spy has never been confirmed, and a petition is now underway calling on the board of directors to reinstate Huang as CEO, arguing that he was effectively ousted without due process. Current employees have joined the petition, which has surpassed 100,000 signatures in total and reached 140,000 at the time of writing.

    Source: gigazine.net / Original article here

    2AnonymousNov 15, 2026 18:04
    Is this a honey trap situation?
    9AnonymousNov 15, 2026 18:07
    Re: #2
    Just a leather-jacket guy's affair, that's all. ("leather jacket guy" = Jensen Huang's nickname online)
    14AnonymousNov 15, 2026 18:09
    They should just make Leather Jacket Guy explain whether he knew about the diversion to China, and give a reasonable explanation for why he sold such massive volumes of product to Singapore.
    15AnonymousNov 15, 2026 18:11
    Haul Jensen in front of a US Congressional hearing, and if he's cleared, do whatever you want after that.

    Wall Street money is already keeping its distance from Nvidia anyway.
    What's spilled can't be put back in the bottle.
    16AnonymousNov 15, 2026 18:12
    "Come back!" Are you crying? lol
    17AnonymousNov 15, 2026 18:12
    (´・ω・`) I'm disappointed in Musk.
    Don't tell me Musk is also subletting servers on the side?
    18AnonymousNov 15, 2026 18:12
    They were keeping the money flowing on nothing but verbal promises — once the cycle stops, the whole bicycle tips over.
    19AnonymousNov 15, 2026 18:12
    If he really was diverting chips, the fines are going to be massive.
    21AnonymousNov 15, 2026 18:13
    If Leather Jacket Guy starts a new company, all the talented people will follow him there.
    35AnonymousNov 15, 2026 18:20
    Re: #21
    There's probably a non-compete clause barring him from doing that for a few years.
    38AnonymousNov 15, 2026 18:23
    It said that once the founder is gone, the company loses its cohesion — well, yeah, of course it would.
    42AnonymousNov 15, 2026 18:25
    Re: #38
    When someone who's been steering the company for over 30 years leaves with zero handover, of course that's going to happen — anywhere would fall apart.
    55AnonymousNov 15, 2026 18:33
    So? Where'd the Chinese woman go?
    60AnonymousNov 15, 2026 18:35
    Re: #55
    Beijing.
    She's not coming back.
    Obviously.
    If she set foot in the US in this situation, they'd squeeze a confession out of her by any means necessary.
    There's no way the FBI backs down and says 'oops, our mistake' at this point.
    They'll fabricate whatever they need.
    Still zero actual evidence to this day.
    62AnonymousNov 15, 2026 18:36
    Re: #60
    The CIA is scrambling to deny she was a spy, though.
    72AnonymousNov 15, 2026 18:48
    Musk is sharp — there's a solid chance he ends up landing Leather Jacket Guy for himself.
    75AnonymousNov 15, 2026 18:50
    Re: #72
    There's no way those two get along — they're both one-man-show types.
    74AnonymousNov 15, 2026 18:50
    All those connections the Ajinomoto president worked so hard to build, trashed in an instant.
    77AnonymousNov 15, 2026 18:52
    Re: #74
    He still physically controls the actual product, so if anything, they'll be the ones coming to bow and beg instead.
    95AnonymousNov 15, 2026 19:03
    A petition for the return of Huang, who was soft on China, huh. Smells like China's fingerprints are blowing this wind.
    209AnonymousNov 15, 2026 22:24
    Re: #95
    It's just a stock-price play.
    165AnonymousNov 15, 2026 20:23
    What a mess, lol
    170AnonymousNov 15, 2026 20:31
    Re: #165
    Isn't this the same as TikTok — the US just taking it over?
    171AnonymousNov 15, 2026 20:32
    Re: #170
    Yeah, it's really pathetic.
    193AnonymousNov 15, 2026 21:38
    Re: #170
    Wasn't it a US company to begin with though?
    195AnonymousNov 15, 2026 21:58
    Wasn't it called some mysterious semiconductor company just like 10 years ago? Getting ousted this easily over an affair makes it sound like it was basically a crude small-business old guy up until then.
    Maybe sudden money just drove him crazy, lol
    198AnonymousNov 15, 2026 22:01
    Re: #195
    It took off big with graphics cards and became well-known back in the Windows 98/XP era, so it's had a presence for over 20 years.

    Background and Key Points of This Discussion

    The starting point for this debate is that Huang’s resignation moved forward even though the core question — whether his affair partner was genuinely a spy — was never confirmed. NVIDIA is a company where co-founder Huang had been at the center of management for more than 30 years, and the thread raised the view that no proper succession plan was in place for a successor. It’s easy to overlook that the petition drive has drawn support from current employees, meaning it’s not simply outside noise but carries an element of internal backing as well. On the other hand, some see China-export and tech-leak suspicions as the real background to the resignation, and opinions were divided over what Elon Musk’s motives might be in signing on. The truth of the spy allegation itself remains unconfirmed as of this writing, so any definitive take on it is still just speculation.

    *This article is compiled as an excerpt and summary from the 5ch (News Flash+) thread “Petition Calling for Jensen Huang’s Return as NVIDIA CEO Tops 100,000 Signatures, Elon Musk and Others Sign.”

  • EU’s New AI ‘EUROPA’ Aims for 1 Trillion Parameters, But Overseas Forums Are Skeptical

    EUROPA, an AI model developed by French AI company Domyn with EU backing, is reportedly aiming for nearly 1 trillion parameters, sparking discussion on the overseas forum Lemmy’s c/europe community. While attention focused on the EU touting this as an “open” AI model, the thread’s first reaction was suspicion about where the training data actually came from — commenters suggested it likely relies on scraped or pirated data, just like other major models. The weights (the model’s actual data) haven’t been released either, and Domyn hasn’t given a firm release date; a June report said Domyn was aiming for release “within the year.” Discussion also turned to just how hard it would be to run a trillion-parameter-class model on personal hardware.

    EUROPA, the EU’s “open AI” model, is aiming for nearly 1 trillion parameters

    Source: maccus.fr / Read the original article here

    3Anonymous2026-11-15 13:07
    Curious what they're actually using for training data. Isn't this just going to be pirated/scraped data like everyone else, in the end?
    13Anonymous2026-11-15 14:30
    Re: #3 Yeah, the answer's probably yes. Honestly, I'm already sick of this topic.
    14Anonymous2026-11-15 15:25
    Re: #3 Exactly — 1 trillion parameters is basically "almost everything humanity has ever written" territory.
    4Anonymous2026-11-15 11:15
    Nice! I wonder how much a home server would even cost to run this at a decent speed (tokens/sec)?
    6Anonymous2026-11-15 11:40
    Re: #4 Judging by the size of other cutting-edge models on Hugging Face, running this at home looks impossible. Right now the ceiling for squeezing something into 64GB of RAM is around 70–80B parameters at Q4 quantization. Push the compression any harder than that and performance really tanks.
    7Anonymous2026-11-15 14:19
    Re: #4 If you want a more realistically-sized European open-weight model, check out Ministral. It's French-made — basically a scaled-down version of full-size Mistral. The 27B Ministral 3 model runs just fine on a 12GB card.
    10Anonymous2026-11-15 11:35
    Re: #4 Domyn hasn't released EUROPA's weight files yet (the files containing the actual parameters that would let anyone run it on their own hardware). Apparently the plan is to release those alongside details of the training methodology once EUROPA ships, though the article gives no specific date. Back in June, Declic Media reported Domyn was aiming for a release "within the year." Guess we just wait and see.
    5Anonymous2026-11-15 14:21
    Is this going to be open source, or just open weights?
    8Anonymous2026-11-15 14:16
    Re: #11 Systems with unified memory (like AMD's or Apple's) can scale up to 512GB of RAM.* *Assuming you're willing to pay about as much as a mid-size new car, that is.
    11Anonymous2026-11-15 13:32
    Re: #6 So right now the ceiling for consumer GPUs is 64GB?
    9Anonymous2026-11-15 13:46
    Re: #6 Is that 64GB of VRAM or regular RAM? If it's VRAM, that's going to run you about as much as a cheap car.
    12Anonymous2026-11-15 14:01
    Re: #9 VRAM, obviously.

    Background and Key Points of This Discussion

    To start, the thread never really settled on what “open” even means here. “Open weight” — releasing only the weights (the parameters themselves) — is different from “open source,” which includes the training code and data too, and many so-called “open” AI models, including Meta’s Llama series, are in practice only open-weight. EUROPA is no different: developer Domyn still hasn’t released the weights and hasn’t even committed to a release date. On top of that, a trillion-parameter-class model simply isn’t something you can run on a personal PC — and even if you could, it would take specialized, expensive hardware. The rule of thumb raised in the thread — “70–80B parameters on 64GB of RAM” — already assumes compression via quantization, and that’s still nowhere close to 1 trillion parameters. As for where the training data comes from, lawsuits over scraping web data without copyright holders’ permission have been piling up across multiple countries, and the underlying point being made is this: even if the EU builds its own “open” model, if it’s using the same methods as everyone else, it’s hard to see how it’s really any different from the major models that came before it.

    *This article is translated and excerpted by our editorial team from the overseas forum thread on Lemmy c/europe, “EUROPA, the EU’s open AI model, is aiming for nearly 1 trillion parameters.” Our site takes responsibility for the accuracy of this translation.

  • GTA6 Physical Edition Exploit: Redemption Code Visible From Outside the Box

    The physical edition of GTA6 doesn’t include a disc — instead, it ships with a redemption code. But a report on X revealed that shining a powerful flashlight at an angle through the side of the box lets you read that code without ever opening it. The thread split over this: some argued “the code can’t be used until it’s scanned at checkout,” while others countered “then shining a light on it in-store isn’t a problem either.” Suspicion also spread that so-called “unopened” used copies might secretly already have redeemed codes. On top of that, reports originating from Reddit about defective copies with the code field left blank (unprinted) are also making the rounds.

    1AnonymousNov 14, 2026 21:34
    tps://x.com/GTAVI_Countdown/status/2121273862751146002
    A fan discovered that shining a powerful flashlight at an angle through the side of the box lets you read the entire redemption code for the physical edition of GTA 6 without opening it
    3AnonymousNov 14, 2026 21:36
    Grand Theft… Auto, am I right?
    Easy money right there
    4AnonymousNov 14, 2026 21:40
    The game begins the moment you try to get your hands on it, huh
    6AnonymousNov 14, 2026 21:48
    So you could unseal it, lift the code, and resell it as "used"?
    67AnonymousNov 14, 2026 23:05
    Re: #6
    Exactly
    That's why "unopened" copies sold on Mercari (Japan's big resale app) are probably already redeemed
    7AnonymousNov 14, 2026 21:50
    Feels like a very GTA kind of story.
    Maybe Rockstar built this in on purpose.
    8AnonymousNov 14, 2026 21:51
    So you could just check it in-store and skip buying it altogether
    9AnonymousNov 14, 2026 21:54
    This is seriously messed up
    Shining a light on it in the store isn't even shoplifting lol
    64AnonymousNov 14, 2026 23:02
    Can't wait to watch Re: #9 get "arrested" for this lol
    10AnonymousNov 14, 2026 21:55
    So you just shine a light, snap a pic with your phone, and pull off easy, legal shoplifting? lol
    12AnonymousNov 14, 2026 21:57
    Guess that means "unopened" can't be trusted at all anymore lol
    13AnonymousNov 14, 2026 21:58
    If someone actually pulled this off, could the code be traced back to them?
    14AnonymousNov 14, 2026 22:00
    What crime is it even to use a code off an unopened box in the first place?
    15AnonymousNov 14, 2026 22:00
    If a "used, unopened" copy turns out to have an already-used code, can you return it to the store as defective?
    16AnonymousNov 14, 2026 22:01
    Re: #15
    GTA6 probably won't even qualify for used buyback
    17AnonymousNov 14, 2026 22:04
    Re: #15
    Doubt it
    There's no way to prove the customer didn't use it themselves
    18AnonymousNov 14, 2026 22:05
    Nah, normally the actual stock would just be kept in the back of the store, right?
    21AnonymousNov 14, 2026 22:07
    No receipt, but my stomach's got the code memorized
    22AnonymousNov 14, 2026 22:08
    Re: #19
    Bet the culprit just turns around and sells that code online
    26AnonymousNov 14, 2026 22:10
    Isn't this just the old "Big One Gum" trick — peeking through a hole to see the prize without opening it? (a classic Japanese candy scam)
    50AnonymousNov 14, 2026 22:38
    Sorry, nice try
    The code can't be used until it's scanned at the register~
    54AnonymousNov 14, 2026 22:50
    Re: #50
    Oh really
    So it's totally fine if people are popping camera flashes at the shelf then!
    59AnonymousNov 14, 2026 22:55
    Re: #50
    Wish that were true
    But it's not, which is exactly why this is such a problem
    70AnonymousNov 14, 2026 23:13
    Looks like defective copies with the code not even printed are turning up
    Reddit's full of posts about it
    80AnonymousNov 14, 2026 23:26
    Re: #70
    Adding insult to injury — well, kind of self-inflicted, but seriously, what's going on with their quality control?
    86AnonymousNov 14, 2026 23:32
    Re: #70
    If it's just the code that's missing, that's not so bad
    But if it's literally a blank sheet of paper in there with nothing printed at all
    then there's no way to even prove it's defective
    97AnonymousNov 14, 2026 23:42
    Re: #70
    A few cases are circulating on X too
    Can't believe that actually happens
    71AnonymousNov 14, 2026 23:15
    Wait, why doesn't this come with a poster?
    Actually it does come with a huge poster, so wouldn't that make this trick impossible?
    75AnonymousNov 14, 2026 23:21
    Re: #71
    Being able to peek probably rakes in more impressions, right?
    74AnonymousNov 14, 2026 23:17
    Hold on, why's everyone talking like the box you can pick up off the shelf is the "real product"??

    There's no way that's not a dummy/display box
    Are there even stores these days that put the actual product out on the shelf?
    78AnonymousNov 14, 2026 23:24
    Re: #74
    Overseas stores just put the real product out though?
    90AnonymousNov 14, 2026 23:37
    Does using this code once lock everyone else out of downloading with it?
    93AnonymousNov 14, 2026 23:38
    Re: #90
    Why would you even question that it's single-use?

    Background and Key Points of This Topic

    Physical editions of video games increasingly skip the disc entirely, packaging only a download redemption code while the game itself is delivered as a digital purchase. The core issue here is a physical, structural one — since the code is printed on the inside of the box, shining a strong light through the side lets it be read from outside — and it isn’t a defect unique to GTA6. The thread split between two theories: one holding that even if the code is read and used, it shouldn’t actually activate unless the purchase is processed (linked) at the register, and the other countering that if that’s true, shining a light on the box in-store to check it causes no real harm and isn’t illegal. The actual specifications of the retailer’s sales system couldn’t be confirmed within the thread. There were also separate concerns about the possibility that “unopened” used copies already have redeemed codes, as well as reports of defective copies where the code field itself was left blank due to a printing error — worth noting this is a distinct quality-control issue, separate from the flashlight trick itself.

    *This article is compiled as an excerpt and summary of the 5ch (Hardware/Industry [Geha]) thread “[Good News] Trick Discovered for Checking Physical Edition GTA6’s Redemption Code Without Opening the Box!“.

  • Norinchukin Bank invests in AI semiconductor ETF — 5ch calls it a ‘picture-perfect shoeshine-boy signal’ (a classic market-top omen)

    A Nikkei report has revealed that Norinchukin Bank, which manages around ¥45 trillion in assets, has invested in semiconductor-related ETFs (exchange-traded funds). The move is unusual for the bank, which has traditionally stuck to broad, diversified investment across the whole market — though overseas pension funds are reportedly also buying the dip on AI-related stocks. On 5ch, opinions clashed: some mocked the timing as a “top signal” given AI stocks’ recent sluggishness, while others argued that with the hype already priced out, now is exactly the time to buy. Many posts also dredged up the bank’s past massive losses from subprime-mortgage-backed securities and U.S. Treasury bond investments, voicing unease about this latest move.

    Norinchukin Bank, which manages roughly ¥45 trillion, has gone ahead with an investment in a semiconductor stock ETF (exchange-traded fund). It’s an unusual move for a bank that has stuck firmly to broad investment across the whole market. Overseas pension funds are also reportedly buying the dip on AI-related (artificial intelligence) stocks.

    AI stocks have been trading sluggishly, but long-term capital that still believes in the growth story is starting to move.

    (Continued at the link below)

    Source: nikkei.com / Read the original article here

    2AnonymousNov 13, 2026 22:14
    There it is — the AI bubble's top signal.
    74AnonymousNov 13, 2026 22:47
    Re: #2
    Rice shortage again, huh…
    154AnonymousNov 13, 2026 23:13
    Re: #2
    A picture-perfect shoeshine-boy move
    AI's done for a while now
    Another massive loss, and they'll paper it over with rice again (i.e., rice price hikes)
    15AnonymousNov 13, 2026 22:19
    Nice limit order
    Say what you will, AI's supply-demand picture is crystal clear
    Since the bad news is already priced in
    buying here is the right call
    109AnonymousNov 13, 2026 22:58
    Re: #15
    That's what you'd think, huh?
    21AnonymousNov 13, 2026 22:21
    Lol, right as everyone's asking whether the AI and semiconductor bubble already popped
    195AnonymousNov 13, 2026 23:39
    Re: #21
    If it already popped, that's a buying opportunity lol
    It's still got plenty of room to climb lol
    34AnonymousNov 13, 2026 22:30
    Norinchukin's glorious track record

    During the Lehman Shock, booked a massive final loss of about ¥570 billion from subprime-mortgage-backed securities and the like

    ↓

    Scared off complex derivatives, so switched to 'simple' ultra-long-term Western government bonds
    Then racked up unrealized losses from rapid rate hikes and ended up booking a massive ¥1.8 trillion loss anyway

    ↓

    If complex financial products AND government bonds are both no good, it's gotta be stocks!
    So now, of all times, they're scooping up AI semiconductor ETFs
    48AnonymousNov 13, 2026 22:35
    Re: #34
    Lmao
    Is this the 'Gifu Rampage' routine again lol (apparent reference to a comedic wrestling gimmick known for repeat failures)
    43AnonymousNov 13, 2026 22:33
    Is it even true that they sold off government bonds to cover the losses in the first place?
    52AnonymousNov 13, 2026 22:37
    Re: #43
    They did sell off their US Treasuries
    56AnonymousNov 13, 2026 22:39
    Re: #43
    If they sold Japanese government bonds now, it'd just mean even bigger losses
    44AnonymousNov 13, 2026 22:33
    So after a massive loss on bonds, now they're riding the AI bubble?
    Total self-destructive gambler of an incompetent organization
    46AnonymousNov 13, 2026 22:34
    Re: #44
    Rice farming itself is basically a gamble to begin with
    67AnonymousNov 13, 2026 22:43
    For some reason, Japan's 'farmers' bank,' Norinchukin, holds the largest CLO portfolio in the world
    They're getting played for suckers by the whole world
    Farmers need to wake up to the danger, fast
    Well, it's probably too late even if they do
    71AnonymousNov 13, 2026 22:46
    Re: #67
    Did Norinchukin kill your parents or something…?
    78AnonymousNov 13, 2026 22:49
    Re: #67
    That's exactly the kind of thing that blows up badly in a downturn.
    69AnonymousNov 13, 2026 22:44
    They say 'AI-related,' but it doesn't actually say semiconductor-related
    It could just as easily be SoftBank, Hitachi, Ajinomoto, Daikin, or other physical-AI names
    86AnonymousNov 13, 2026 22:51
    Re: #69
    Apparently it tracks the Nikkei Semiconductor Stock Index
    79AnonymousNov 13, 2026 22:50
    Re: #71
    They're holding ¥13 trillion worth of garbage CLOs
    Anyone who works in finance should know how scary that is
    103AnonymousNov 13, 2026 22:56
    Re: #79
    It's common knowledge among finance people
    that AAA-rated CLO tranches have never actually blown up…
    95AnonymousNov 13, 2026 22:53
    Norinchukin went on an MBS-buying spree back in the day too, didn't it
    102AnonymousNov 13, 2026 22:56
    Re: #95
    This time it's CLOs
    That's just insane — pretty much everything except Hikari Tsushin corporate bonds from Fukutsu, Fukuoka
    101AnonymousNov 13, 2026 22:56
    Before the Lehman Shock
    buys subprime mortgage bonds

    During the COVID shock
    buys ultra-long-term US Treasuries at rock-bottom rates

    During the AI bubble
    buys AI stocks —
    that's Norinchukin
    107AnonymousNov 13, 2026 22:57
    Re: #101
    A genuine genius at the flying-leap catch (i.e., always diving in right at the top)
    140AnonymousNov 13, 2026 23:08
    If it fails, they can just double the price of rice — simple as that
    144AnonymousNov 13, 2026 23:09
    Re: #140
    The ones who really get screwed are the farmers
    143AnonymousNov 13, 2026 23:08
    ¥45 trillion is wild
    Controlling the rice market is one hell of an advantage
    147AnonymousNov 13, 2026 23:10
    Re: #143
    They're probably burning through the national pension savings grandpas and grandmas scraped together like it's nothing

    Background and Key Points of This Discussion

    Norinchukin Bank is a financial institution established under the Norinchukin Bank Act to invest funds pooled from agricultural cooperatives (JA) nationwide, and it has traditionally followed a policy of broad diversification across domestic and overseas stocks and bonds. Against that backdrop, the reported investment in a semiconductor ETF is unusual. The bank has a history of booking massive losses in the past — about ¥570 billion from subprime-mortgage-backed securities, and another huge loss from ultra-long-term U.S. Treasury holdings — and many posters in the thread cited that history to voice concern that the bank might suffer another big loss. Others, however, praised the move as ‘buying the dip at a point where the hype has already faded’ from AI stocks, reflecting a split between those focused on short-term price action and those betting on long-term growth. Note that posts linking the bank’s past losses to rising rice prices are speculation on the part of individual posters and are not mentioned in the original source article.

    ※This article is excerpted and summarized from the 5ch (News Flash+) thread「Norinchukin makes its move, buying AI stocks at the bottom — betting on long-term stable growth, “growth over near-term moves”」.

  • Magnetic Tape Back in the Spotlight in the Generative AI Era — 5ch: “Slower Than Cup Noodles”

    Fujifilm announced it will expand its production capacity for data-storage magnetic tape fivefold, responding to a surge in data volumes driven by the spread of generative AI. Magnetic tape is getting a second look because it can be stored disconnected from any network, lowering the risk of data leaks from unauthorized access. Over on 5ch, the thread filled up with back-and-forth over the math behind 100TB-per-cartridge capacities and transfer speeds, nostalgic reminiscing about past storage formats like cassette tapes, MiniDiscs, DVDs, and Blu-ray, and it even spiraled into worries about SSD lifespans and a durability debate over whether paper or stone outlasts anything digital.

    Published 11/13 (Fri) 11:42

    TV Asahi (ANN)

    Fujifilm has announced it will increase its production capacity for data-storage magnetic tape fivefold. Amid growing calls for countermeasures against information leaks from unauthorized access, the old-school technology of storing data on tape disconnected from any network is getting fresh attention.

    Source: news.yahoo.co.jp / Original article here

    4AnonymousNov 13, 2026 20:41
    Recording data on cassette tapes — that's some 'Pasokon Sunday' (a classic retro Japanese TV show about home computers) era stuff right there
    24AnonymousNov 13, 2026 20:50
    Re: #4
    One cartridge holds 100TB, you know
    5AnonymousNov 13, 2026 20:41
    It'd be nice if there were a drive that could read and write digital data onto videotape. Cartridges have that general-purpose appeal going for them
    39AnonymousNov 13, 2026 20:56
    Re: #5
    Sansui used to make a PCM processor meant for use with VCRs
    9AnonymousNov 13, 2026 20:44
    Does magnetic tape actually outlast optical discs and flash memory?
    20AnonymousNov 13, 2026 20:49
    Re: #9
    If it's longevity you're after, punched tape is the real champion
    17AnonymousNov 13, 2026 20:48
    Even magnetic tape is hitting terabyte scale now
    Makes you wonder if an overnight backup would even finish in time
    31AnonymousNov 13, 2026 20:53
    Re: #17
    Even defragging a hard disk on Windows 85 sometimes took all night, so there's no way magnetic tape finishes in time
    18AnonymousNov 13, 2026 20:48
    https://asset.fujifilm.com/www/jp/files/2021-08/6c229b66fcdc6b310f1a3554b163118b/hero_ltotape_01_sp.jpg
    23AnonymousNov 13, 2026 20:50
    Even a full format on a multi-terabyte hard disk can take a couple of days, so who knows how many days it'd take to fill one of these all the way up
    134AnonymousNov 13, 2026 21:52
    Re: #23
    The fastest spec out right now tops out at
    400MB/sec
    so for 100TB that works out to about 250 seconds (4 min 10 sec).

    Faster than instant udon (5 min), but slower than Cup Noodles (3 min), apparently.
    26AnonymousNov 13, 2026 20:51
    In the end, paper really is the strongest medium. It can preserve records 1,000 years into the future and needs no playback device at all
    32AnonymousNov 13, 2026 20:54
    Re: #26
    For sheer durability, writing onto stone or ceramic has to be the strongest option
    History actually backs this up too — things like the Rosetta Stone prove information can survive since before the common era
    The only downside is that once it's written, rewriting it is basically impossible
    30AnonymousNov 13, 2026 20:53
    It's been in constant use this whole time — there's nothing to 'rediscover' about something that never actually went away
    38AnonymousNov 13, 2026 20:55
    Re: #30
    Is magnetic tape still in active use where you work?
    What do you use it for?
    35AnonymousNov 13, 2026 20:54
    Love this trend of analog stuff getting re-appreciated
    76AnonymousNov 13, 2026 21:15
    Re: #35
    Hanko (personal seal) shops are about to make a huge comeback
    82AnonymousNov 13, 2026 21:18
    Re: #76
    Contracts have gone digital with DX (digital transformation) too, so if the cloud ever goes down entirely, that's going to be a disaster
    40AnonymousNov 13, 2026 20:56
    Is it really okay to put an employee's full name out there like that?
    46AnonymousNov 13, 2026 20:59
    Re: #40
    Why would you assume that's a problem?
    42AnonymousNov 13, 2026 20:58
    Open reel
    ↓
    Cassette tape
    ↓
    DAT, LaserDisc
    ↓
    CD
    ↓
    MD (MiniDisc)
    ↓
    DVD
    ↓
    Blu-ray
    155AnonymousNov 13, 2026 22:22
    Re: #42

    Spare a thought for HD DVD every once in a while too
    197AnonymousNov 13, 2026 23:31
    Re: #42
    Zip disks….
    48AnonymousNov 13, 2026 21:02
    >>+42
    You left out the L-cassette
    Start over
    183AnonymousNov 13, 2026 23:04
    Re: #48
    Way too much missing — Micro Ace cassette, VHD, and more
    54AnonymousNov 13, 2026 21:04
    You could air-gap a current SSD from the network just as easily
    Are they just forcing this into a 'ramp up tape production' story?
    56AnonymousNov 13, 2026 21:07
    Re: #54
    Tape lasts 200 years, but SSDs degrade over time, don't they?
    58AnonymousNov 13, 2026 21:07
    Re: #54
    True
    Networks exist for real-time sharing
    If all you need is to preserve records with today's latest systems, just keep them isolated from the network
    78AnonymousNov 13, 2026 21:17
    Re: #54
    Heard SSDs can just die suddenly without warning, which is scary
    Haven't experienced it myself though
    73AnonymousNov 13, 2026 21:14
    Sorry for staying quiet about it until now (apparently a line from an old cassette-tape commercial jingle)
    79AnonymousNov 13, 2026 21:17
    Re: #73
    Wasn't that for AXIA (Fujifilm's old cassette tape brand)?
    88AnonymousNov 13, 2026 21:19
    Re: #73
    Re: #76
    Saito Yuki is still going strong, you know
    169AnonymousNov 13, 2026 22:41
    Re: #73
    Those were pretty excuse-laden lyrics about a woman two-timing someone
    Lyrics by Natsuo Giniro, music by Kenjiro Sakitani
    107AnonymousNov 13, 2026 21:28
    They should put all those VHS tapes people already own to good use and make a VHS deck that can record data
    151AnonymousNov 13, 2026 22:17
    Re: #107
    Betamax too, please
    157AnonymousNov 13, 2026 22:25
    Re: #107
    The Beta deck I used to own could do PCM recording if you attached the optional PCM module
    Which means it could already write digital data
    139AnonymousNov 13, 2026 21:54
    This is the one

    18TB for 24,380 yen

    https://lto.co.jp/pic-labo/archival_life01.jpg
    168AnonymousNov 13, 2026 22:40
    Thought this was about data tape, then everyone veered off into audio tape talk, and now we're right back to data again lol
    By the way, do they even sell data cassette decks for regular consumers? Or do you just run that screechy modem-style noise through a regular audio deck?
    171AnonymousNov 13, 2026 22:43
    Re: #168
    Nobody's actually reading or writing data on cassette tape these days anyway
    173AnonymousNov 13, 2026 22:44
    Re: #168
    Kansas City standard
    'Sapporo City standard' (a joking riff on that old data-over-cassette format)

    Background and key points of this topic

    Magnetic tape has all but disappeared as a consumer storage medium, but it’s alive and well in the enterprise world under the LTO standard, still widely used for backups today. The reason it’s being ‘rediscovered’ now comes down to the rising value of ‘air-gapped’ storage — keeping data physically disconnected from the network — as a defense against ransomware. The thread lit up with nostalgia over old storage formats, but there were also level-headed comments pointing out that there’s nothing new about something that’s ‘still in use anyway,’ with some even pushing back on the article’s framing of it as a ‘rediscovery.’ While magnetic tape is considered to outlast SSDs for long-term storage, its slow read speed and poor searchability remain real drawbacks.

    ※This article is excerpted and summarized from the 5ch (News Express+) thread “Magnetic Tape Back in the Spotlight — Generative AI Boom Drives Massive Data Growth, Demand Rises as Leak Risk Stays Low.”

  • AI Is Crashing Kids’ Grades, Says Tokyo University Professor — 5ch: ‘Isn’t This Just the Old Manga-Ban Argument All Over Again?’

    Kunie Sakai, a professor at the University of Tokyo’s Graduate School and a language neuroscientist, has reportedly argued in the December issue of Bungei Shunju that AI use in education “should be banned outright.” His reasoning: if children give up thinking in language by leaning on AI, it risks stunting the development of brain functions — including the grammar center — and he claims abnormalities are already showing up even in current Tokyo University students. On 5ch’s science news board, reactions split between posters sympathetic to the risk of AI dependence eroding critical thinking, and skeptics who doubted the causal link had been proven, with many noting the resemblance to past moral panics over manga and video games.

    Leading language neuroscientist and Tokyo University professor Sakai sounds the alarm: “Ban young people’s AI use entirely” — “Kids’ academic ability is plummeting, and abnormalities are even showing up in current Tokyo University students…”

    Something is shaking education to its very foundations.

    Kunie Sakai, a language neuroscientist and professor at the University of Tokyo’s Graduate School of Arts and Sciences, has issued an urgent proposal in the December issue of Bungei Shunju (on sale November 10): AI use in educational settings “should be banned outright.”

    Source: bunshun.jp / Original article here

    2AnonymousNov 10, 2026 22:22
    "Not thinking at all" is bad news.
    Even grad students these days just dump AI's hallucinated garbage straight into their papers,
    so I get why people are worried about younger kids picking up even worse habits.
    3AnonymousNov 10, 2026 22:27
    Behold his honorable face (forum slang for "here's a photo of the guy")
    https://bunshun.ismcdn.jp/mwimgs/0/2/1500wm/img_023d0d6be58a6a4e69dcf16cea5b6afa150792.jpg
    4AnonymousNov 10, 2026 22:58
    If you use AI, the world's cutting edge becomes your starting line.
    Anyone who can't even manage that
    might as well get a lobotomy.
    5AnonymousNov 10, 2026 23:25
    You can use AI to get smarter, but you can also use it to cut corners and get dumber.

    Left to their own devices, most people end up as the latter.
    7AnonymousNov 10, 2026 23:32
    As someone who's dabbled in cultural anthropology, I'm kind of curious whether kids raised entirely on AI will turn into "people from before writing was invented" lol.
    The "voice of god" from the bicameral mind hypothesis, lol.
    As long as you live by whatever AI-sama tells you, there's no problem at alllll~
    10AnonymousNov 10, 2026 23:50
    Re: #7
    A regression to the state before Socrates' critique? (He famously warned that writing itself would weaken human memory — same argument, new technology.)
    8AnonymousNov 10, 2026 23:38
    Joking aside, I really think the best fix is building the habit of fact-checking what AI tells you — or making kids do it.
    It comes up a lot in cooking talk: AI states something that's chemically wrong,
    and when you ask for a source, it turns out to be some copy-paste site dressed up to look legit.
    9AnonymousNov 10, 2026 23:43
    Well, idiots who mistake AI's abilities for their own
    exist everywhere,
    no matter where you look.
    11AnonymousNov 11, 2026 00:47
    Kids not bringing notebooks or pens because they just photograph the blackboard with their phone — people have been saying that for ages already.
    12AnonymousNov 11, 2026 00:49
    Cameras mean kids never develop the ability to sketch things in the field.
    It was never a needed skill to begin with, so no problem there.
    Situations where you can't use a camera are so nonexistent it's almost absurd to even picture one.
    13AnonymousNov 11, 2026 00:53
    So it ends up being a convenient idiot-manufacturing machine for rulers who want to govern a dumbed-down populace?
    14AnonymousNov 11, 2026 01:58
    This is exactly the kind of setup you'd see in Showa-era sci-fi, isn't it?
    Someone has a personal computer that advises them on everything, and when a system glitch cuts off the advice, they're left helpless, unable to do anything on their own.
    15AnonymousNov 11, 2026 03:09
    Well, I can't really talk big since I barely took notes even when I did bring them.
    Apparently Socrates also said that taking notes makes you stupid,
    but still, yeah, I do think this AI thing isn't great.
    16AnonymousNov 11, 2026 03:55
    So it's like one of those retired-bureaucrat guys ("amakudari" officials parachuted into cushy post-retirement jobs) whose actual skills atrophied from farming everything out to subcontractors?
    17AnonymousNov 11, 2026 04:03
    Work keeps throwing up new problems and the business keeps changing, so you can't stay lazy forever —
    but compulsory education's the same every single year, so AI's probably already got it down perfectly, huh.
    18AnonymousNov 11, 2026 05:18
    Humanities-side brain scientists only ever say half-baked stuff anyway.
    19AnonymousNov 11, 2026 05:38
    Kraftwerk — "The Robots"

    I'm your servant
    I'm your worker
    We're charging our battery
    And now we're full of energy
    We are the robots
    We're functioning automatik
    And we are dancing mechanik
    We are the robots
    I'm your servant
    I'm your worker
    We've just been programmed
    to do
    whatever you want us to
    20AnonymousNov 11, 2026 06:35
    We finally got to a point where we could cut corners, and this just means more fact-checking work instead.
    Keep that up and AI ends up being something you can only use for harmless small talk.
    21AnonymousNov 11, 2026 06:38
    Kids who read Dazai Osamu and that sort of thing end up writing copycat prose and turning into pretentious literary wannabes,
    so reading literary fiction should be banned for young people too.
    Literary pretension — that's basically like catching the measles.
    37AnonymousNov 11, 2026 18:36
    Re: #21
    Wouldn't it be fine if you could actually pull off imitating Dazai well?
    That stuff's easy once you've got reading comprehension — without it, it just comes across as needlessly convoluted.
    22AnonymousNov 11, 2026 08:00
    That's rough.
    The correlation between declining reading comprehension and AI is a logical leap, and the rest is just assumption with zero evidence.
    Maybe it's professor quality that's declining here, if this is coming from a Tokyo University professor?
    23AnonymousNov 11, 2026 08:21
    The Reiwa-era "Drive to Eliminate Harmful Books" (a nod to the 1950s anti-manga panic campaign)

    Let's get Reiwa White Postboxes (the real bins once used to collect "harmful" manga for disposal — now repurposed for "harmful apps") set up in front of train stations.
    24AnonymousNov 11, 2026 08:22
    Forget AI — kids probably shouldn't be using PCs or smartphones at all during their growth years.
    For people with disabilities they're a necessary tool, but if you've got full use of your body, handwriting and paper books are what actually stick with you.
    25AnonymousNov 11, 2026 08:56
    Manga and anime, then games after that —
    "it'll make kids dumb, it'll push them into crime, so keep them away" — that's the kind of half-baked stuff "experts" have been peddling on daytime TV forever.
    This guy's probably just more of the same breed.
    27AnonymousNov 11, 2026 09:52
    I mean, the skills involved are just different.
    Using AI, mimicking what AI says — sure, you build that particular skill.

    But organizing information with notebook and pencil trains you up through: organizing information yourself → explaining it to others based on that → the ability to revise it.

    Use AI well and you can even push past human cognitive limits in that area — but lean on it too hard and… there's a real risk it just never develops.
    28AnonymousNov 11, 2026 10:42
    Maybe we'll start seeing "thinking gyms," like sports gyms.
    30AnonymousNov 11, 2026 13:18
    I don't think today's so-called "digital natives," who get treated like they're something amazing, are actually that impressive.
    They're just smartphone addicts. Whatever they can do is well within the range you'd expect.
    But AI natives are going to be different.
    35AnonymousNov 11, 2026 18:16
    What's wild is that it never occurs to him to just teach kids how to use AI properly.
    At least understand deduction and induction first.
    Before banning AI use, go ban dictionary use while you're at it.
    36AnonymousNov 11, 2026 18:26
    Sounds like a sensible teacher. AI is a tool you pick up after you've already built basic thinking and judgment skills — not before.
    38AnonymousNov 11, 2026 18:39
    "The answer's always written in the textbook" —
    so we're just going to get more idiots like those economists, huh.
    39AnonymousNov 11, 2026 18:42
    It'll be genuinely interesting to see what happens to someone who relies on AI from birth all the way to adulthood.
    40AnonymousNov 11, 2026 18:53
    ">'Giving up on thinking in language by relying on AI will impede the development of innate brain functions, starting with the grammar center. It may be each adult's own responsibility if they use AI after having trained their brain in a pre-AI era — but encouraging children and students still in their developmental years, from elementary school all the way through university, to make use of it is tantamount to a reckless human experiment.'

    That's probably the real point he's making.
    Europe's been in the news over restricting mobile devices in elementary, middle, and high schools —
    I wonder if AI's going to start generating that same kind of news eventually too.

    Background and key points of this debate

    Kunie Sakai is a language neuroscientist known for his research on the brain’s grammar center, and this proposal was published not in an academic journal but as a magazine contribution to Bungei Shunju. Where the thread split was over one specific point: whether a causal link between AI dependence and actual academic decline has ever actually been established. Supporters pointed to real cases of current students using AI output without checking it, while skeptics argued the structure is identical to past moral panics claiming manga or video games “corrupt” children. One thing readers can easily miss is that the proposal targets use in educational settings from elementary school through university — it doesn’t reject AI use by working adults themselves, so it’s worth keeping in mind that this is an age-bracketed argument, as the article itself frames it. The comparison to Europe’s ongoing restrictions on mobile devices in schools also hints that AI regulation could end up following a similar trajectory.

    ※This article excerpts and summarizes the 5ch (Science News+) thread “[Proposal] Leading Language Neuroscientist Sounds the Alarm: “Ban Young People’s AI Use Entirely”.”

  • NVIDIA President Rumored for Divorce, Wife Could Get Trillions of Yen — 5ch: “Now That’s a Dream”

    Speculation about a divorce between NVIDIA CEO Jensen Huang and his wife Lori has emerged after Huang admitted to an affair. Should the couple divorce, it could trigger an asset split potentially worth tens of billions of dollars; on November 9, 2026, a sharp drop in NVIDIA’s stock price even briefly knocked Huang out of the world’s top 10 richest people. On the thread, posters expressed surprise that U.S. marital property division can extend to assets acquired before marriage, contrasted it with Japan’s system, and some brought up Bill and Melinda Gates’ divorce as a precedent.

    Divorce Speculation Surrounds NVIDIA CEO Who Admitted to Affair — Wife Could Receive “Over ¥11.8 Trillion”

    Divorce speculation has surfaced involving NVIDIA CEO Jensen Huang and his wife Lori Huang, after Huang admitted to having an affair. Should the couple divorce, Lori could receive assets worth tens of billions of dollars (trillions of yen). The exact amount she would actually receive is unknown.

    Huang ranks among the world’s top 10 wealthiest people. On November 9, 2026, Eastern Time, NVIDIA shares plunged, briefly knocking him out of the top 10 on the wealth rankings. As of 4 p.m. that day, his net worth was estimated by Forbes at approximately $155.2 billion (about ¥24.5216 trillion).

    Source: news.yahoo.co.jp / Original article here

    19AnonymousNov 11, 2026 20:42
    The wife: "Now that's livin' the dream"
    91AnonymousNov 11, 2026 22:09
    Re: #19
    Wife: "Hey boss~, that's way too cheap for me~ ♡"
    164AnonymousNov 12, 2026 00:59
    Re: #19
    Only women get to sit down next to everything you spent your whole life desperately building — in one second flat
    37AnonymousNov 11, 2026 20:53
    And then this wife defects to China after the divorce
    119AnonymousNov 11, 2026 23:09
    Re: #37
    Nobody actually gains anything from that
    202AnonymousNov 12, 2026 07:31
    Re: #37
    The world really is rational!
    38AnonymousNov 11, 2026 20:56
    Capitalism really is messed up, huh
    121AnonymousNov 11, 2026 23:10
    Re: #38
    You only think it's messed up because you're poor lol
    49AnonymousNov 11, 2026 21:12
    It's over lmaooooo

    "Jensen Huang's affair partner smuggled $7.5 billion worth of NVIDIA AI chips to China"
    Posted 11/11 (Wed) 9:43 AM
    https://news.yahoo.co.jp/articles/a3177fe5ca3316fe707a94c1977e48fbfeb7f32e
    53AnonymousNov 11, 2026 21:17
    Re: #49
    Whoa, this is way bigger than just a divorce now lol. Could this end in an arrest?
    57AnonymousNov 11, 2026 21:20
    Re: #49
    Isn't this basically a life-sentence case at this point?
    61AnonymousNov 11, 2026 21:30
    Bill Gates' ex-wife made out pretty well too.
    Let's ask an AI.

    Estimates put the total asset split at roughly $7.3 billion to $76 billion (around ¥8 trillion at the time).
    100AnonymousNov 11, 2026 22:31
    Re: #61
    >roughly $7.3 billion to $76 billion

    That's… quite a wide range
    62AnonymousNov 11, 2026 21:31
    Must be nice being a woman, having this kind of easy dream available to you
    71AnonymousNov 11, 2026 21:39
    Re: #62
    "Easy"? Do you have any idea how many years ago this wife married Mr. Leather Jacket? (a nickname for Jensen Huang, after his trademark leather jacket)
    72AnonymousNov 11, 2026 21:39
    If the wife and the spy turn out to have been in on it together, that'd be quite a drama
    88AnonymousNov 11, 2026 21:57
    Re: #72
    Oh, nice plot. If the guy billed as "the spy's husband" turns out to be a spy himself too, and he's been double-cheating with [Huang's] wife on top of it — that's a movie deal right there lol
    80AnonymousNov 11, 2026 21:46
    But she's already up there in age, so what good is inheriting that much money anyway
    94AnonymousNov 11, 2026 22:18
    Re: #80
    It'd be great if she set up some kind of foundation and put it toward AI startups or scholarships for engineering students
    Realistically, what's an old woman going to do with trillions of yen anyway
    95AnonymousNov 11, 2026 22:22
    Guys who've spent their whole lives buried in science and research are exactly the type to fall for even the most obvious Chinese honey trap
    Because they've never actually played the field before
    97AnonymousNov 11, 2026 22:25
    Re: #83
    Huang got married in '85 and NVIDIA was founded in '93, so even under Japan's rules this would get split 50/50
    147AnonymousNov 11, 2026 23:43
    Re: #97
    Japan's 50/50 split only covers assets built up after the marriage.
    An inheritance from your parents stays yours.
    In America it's everything — premarital assets, and of course even scholarships and debts get split too.
    104AnonymousNov 11, 2026 22:36
    Why does a wife who barely did anything get to cash in this big
    Seriously, alimony and inheritance laws are messed up
    106AnonymousNov 11, 2026 22:39
    Re: #104
    In his case, she stuck with him through all the hard times, so the adultery apology he gave his employees reflects that
    105AnonymousNov 11, 2026 22:38
    Still, the US must've had intel on this — they sure let it run loose for a long time
    108AnonymousNov 11, 2026 22:44
    Re: #105
    It's basically like Trump was getting in the way of it
    141AnonymousNov 11, 2026 23:34
    I get that the husband and wife share the surname Huang, but why does the affair partner have the name Huang too?
    180AnonymousNov 12, 2026 02:39
    Re: #141
    He's Huang Renxun (Jensen Huang), the affair partner is Huang Le

    From Baidu Baike
    https://baike.baidu.com/ja/item/%E9%BB%84%E5%A7%93/1132633
    "Demographics: As of 2006, the population with the surname Huang was about 27 million, roughly 2.2% of the national population. By ranking, the Huang surname rose from 8th to 7th most common."
    146AnonymousNov 11, 2026 23:41
    Actually, more than the affair itself, the part about secretly exporting chips to China seems like the real problem here
    152AnonymousNov 11, 2026 23:46
    Re: #146
    What are you talking about.

    That's literally what's been the issue from the start.

    Background and Key Points of This Story

    The asset figures in the reporting are based on Forbes’ estimate of personal net worth, and the actual division of assets would depend heavily on whether a prenuptial agreement exists and on court rulings. On the thread, the idea that “in America, even premarital assets get split 50/50” was widely assumed, but in reality this varies by state law and by contract, and the premise differs from Japan’s principle of dividing only assets built up during the marriage. Another point of discussion was that the woman reported as the affair partner is also said to be connected to the separately reported allegations of NVIDIA AI chips being smuggled to China — on the thread, there was more concern about the fallout from export-control violations and espionage suspicions than about the divorce itself. Some posts cited the Bill and Melinda Gates divorce as a precedent, but the estimated settlement range varies widely depending on the source, and the article itself doesn’t note that no figure has actually been confirmed.

    *This article is compiled as an excerpt and summary from the 5ch (News Express+) thread “U.S. NVIDIA CEO Jensen Huang Faces Divorce Speculation — Wife He Met in College Could Receive “Over ¥11.8 Trillion”“.

  • Is ChatGPT Taking Our Jobs? 5ch: ‘For $100 a Month, You Get a University of Tokyo Grad as Your Partner’

    OpenAI, together with freelance platform Upwork, released survey results on November 10 comparing ChatGPT usage against demand for freelance work. Analyzing 63 job categories, they found that the more a field relied on ChatGPT, the lower the share of job postings on Upwork tended to be. On 5ch, opinions split: some said they already feel illustration and writing work drying up, others countered that AI is just an efficiency tool and doesn’t actually take jobs away, and still others worried about whether this would spread to general corporate jobs and video production too.

    Is “ChatGPT Is Stealing Freelance Jobs” True? Upwork Surveys 63 Fields

    On November 10, OpenAI released survey results conducted jointly with freelance platform Upwork comparing ChatGPT usage against demand for freelance work. Analyzing 63 job fields, they found that categories with relatively heavy ChatGPT use tended to see a decline in the share of job listings on Upwork.

    See the source for details. 2026/11/11

    Source: news.yahoo.co.jp / Original article here

    3AnonymousNov 11, 2026 20:50
    Back in the day, unknown illustrators could pick up quick little gigs for pocket money, but even that's been wiped out by AI now.
    29AnonymousNov 11, 2026 21:16
    Re: #3
    Yeah, it churns out way better illustrations in just a few minutes.
    198AnonymousNov 12, 2026 00:25
    Re: #3
    The sharp unknown illustrators are already delivering AI output plus touch-ups.
    And honestly, that's fine as it is.
    22AnonymousNov 11, 2026 21:06
    General office workers will be taken over by AI within a few years.
    Whichever office drones become redundant will be first on the layoff list.
    24AnonymousNov 11, 2026 21:09
    Re: #22
    For now, ChatGPT ("Chappy," as it's nicknamed here) says companies are absorbing people internally by shifting them to sales and the like, so there hasn't been mass firing yet.
    194AnonymousNov 12, 2026 00:12
    Re: #22
    There's no shortage of work checking and fact-checking whatever AI spits out.
    23AnonymousNov 11, 2026 21:09
    Well, it really is impressively good at stitching together scraped info into text.
    183AnonymousNov 11, 2026 23:35
    Re: #23
    The wording may be slick, but it doesn't actually understand the content — read closely and it often falls apart.
    36AnonymousNov 11, 2026 21:25
    Re: #24
    With domestic demand already shaky and price hikes unavoidable,
    getting shuffled off to sales sounds like hell.
    175AnonymousNov 11, 2026 23:21
    Re: #36
    There's plenty of work out there — bus and truck drivers, rebar workers, scaffolders, you name it.
    38AnonymousNov 11, 2026 21:28
    Isn't it less about "taking" jobs and more about boosting efficiency?
    That just frees you up to do other work, right?
    53AnonymousNov 11, 2026 21:42
    Re: #38
    That's fine if "other work" is actually available to that person — but if it isn't, they just lose their job outright.
    And since everyone who gets pushed out floods into the jobs AI can't take,
    competition spikes there too, on top of pay getting driven down.
    42AnonymousNov 11, 2026 21:31
    Hey you, celebrating how AI's made things easier —
    you're living on borrowed time until someone realizes humans aren't needed at all.
    51AnonymousNov 11, 2026 21:40
    Re: #42
    Whether you celebrate it or not, once "humans aren't needed" becomes reality, everyone's livelihood ends equally, so what's the difference?
    Are you planning on becoming a god or something?
    55AnonymousNov 11, 2026 21:43
    Then freelancers should just use ChatGPT to make money themselves.
    How long are you idiots going to keep doing unnecessary manual work?
    62AnonymousNov 11, 2026 21:45
    Re: #55
    Requests on sites like Lancers (a Japanese freelance marketplace) are going to crater.
    Except for the ones with political connections, freelancers are basically done for.
    56AnonymousNov 11, 2026 21:43
    Writers are in real trouble.

    AI is seriously going to take their jobs.

    Eventually AI will be winning the Naoki Prize and the Akutagawa Prize (Japan's top literary awards).
    70AnonymousNov 11, 2026 21:49
    Re: #56
    Didn't a past winner already use it?
    65AnonymousNov 11, 2026 21:46
    To mere bystanders like you it probably looks hopeless, but speaking as someone actually doing the work hands-on, AI wins about 90% of the time — though in maybe 10% of cases humans still come out ahead.
    69AnonymousNov 11, 2026 21:49
    Re: #65
    The moment you're competing against an AI that runs 24/7/365 without ever sleeping, you've already lost.
    67AnonymousNov 11, 2026 21:48
    I wish it would hurry up and take video work like movies and TV dramas away from actors too.
    72AnonymousNov 11, 2026 21:50
    Re: #67
    Nah, footage that's obviously all-AI just kills the vibe.
    73AnonymousNov 11, 2026 21:51
    Re: #69
    That's a bystander's opinion. Even against something that runs around the clock, I've crushed AI before with a Boolean algebra hack.
    80AnonymousNov 11, 2026 21:55
    Re: #73
    But no matter how hard they try, humans still end up passing out from exhaustion, don't they?
    131AnonymousNov 11, 2026 22:49
    Re: #1
    Workman (a Japanese workwear retailer) said they no longer need a design agency for marketing and ads — not even freelancers.
    They said something that used to cost 3 million yen to make now costs zero.
    Given how many ads they run, of course they'll use it if it works.
    133AnonymousNov 11, 2026 22:50
    Re: #131
    They also said it's huge that location shoots are no longer needed.
    In other words, it's not just labor costs —
    they're cutting total expenses across the board.
    137AnonymousNov 11, 2026 22:52
    Re: #131
    Thing is, that design agency was outsourcing work to freelancers instead of doing it in-house anyway.
    132AnonymousNov 11, 2026 22:49
    Just think of it as getting a University of Tokyo grad as your partner for $100 a month.
    135AnonymousNov 11, 2026 22:51
    Re: #132
    That's not quite right.
    Why University of Tokyo specifically?
    That framing itself is the mistake.
    Everyone and everything has its own strengths.
    141AnonymousNov 11, 2026 22:55
    Re: #132
    An all-around capable person is rare.
    That's why people band together to form companies.
    But an AI that pools all of human knowledge can do everything.
    142AnonymousNov 11, 2026 22:55
    Re: #132
    University of Tokyo? That's trash.
    148AnonymousNov 11, 2026 22:58
    Local government posters have turned into nothing but terrible AI art.
    The resolution's rough and pixelated too.
    149AnonymousNov 11, 2026 22:59
    Re: #148
    Things keep changing fast.
    If all you look at is dumb work, you end up dumb yourself.
    152AnonymousNov 11, 2026 23:00
    Re: #148
    I cracked up seeing a JR train-car ad with a disclaimer that said "This was made by AI."
    157AnonymousNov 11, 2026 23:03
    Re: #148
    Just for you:

    A new image-generation AI called "MiniMax-H3-Image" has arrived, capable of generating 16-megapixel images —
    it runs locally and supports image editing and multi-image reference too.
    gigazine.net/news/20261111-minimax-h3-image/
    150AnonymousNov 11, 2026 22:59
    Re: #141
    It's decent at image generation, programming, and PC-related stuff,
    but in fields without clear-cut answers, it answers questions with total confidence while often being wrong.
    155AnonymousNov 11, 2026 23:02
    Re: #150
    Calculations especially still have a long way to go,
    and for law it mostly just regurgitates interpretations and precedents favoring the drafter's intent — relying on that alone often isn't good enough.
    156AnonymousNov 11, 2026 23:03
    Re: #150
    "It answers" and "that's wrong" — isn't that just a one-way exchange?
    You're using it wrong.
    It's exactly like Japan's test-based education — ask a question, grade the answer —
    and that approach is completely wrong.

    Background and Key Points of This Debate

    It’s worth noting that this survey is a correlation analysis OpenAI itself ran jointly with Upwork. The finding that “fields with heavier ChatGPT use see fewer job listings” is a correlation — it doesn’t prove AI directly took anyone’s job. Within the thread, individual creative freelance work like illustration and writing tended to get lumped together with salaried corporate roles like general office jobs, but the two differ in how easily they can actually be replaced. There’s also a view that, rather than AI completely replacing human labor, new jobs — like checking and fact-checking AI-generated output — are emerging at the same time, which frames this less as work disappearing and more as the nature of work shifting. On the other hand, companies really are cutting back on outsourcing to freelancers in favor of in-house production and AI use, and in some fields this is already showing up as lower per-job rates and fiercer competition.

    *This article is excerpted and summarized from the 5ch (News Speed+) thread “Is “ChatGPT Is Stealing Freelance Jobs” True?“