Category: Tech

  • Elite Kaisei School’s Railway Club Apologizes for Sleeper-Train Misconduct, Suspends Activities

    The Railway Research Club at Kaisei School posted an apology on its official X (formerly Twitter) account, admitting that members violated manners during a trip aboard the Sunrise Seto/Izumo sleeper express. In response, the club says it will suspend travel and other activities. Thanks partly to the novelty of a prestigious school’s club being involved, the thread on 5ch’s News Express+ board took off — sparking criticism of the train fans’ misconduct itself, the perennial “photographers vs. riders, which type is more of a nuisance” debate, and, via Kaisei’s reputation as an elite prep school, a tangent into the age-old argument over whether success in middle-school entrance exams comes down to talent or effort.

    Kaisei School Railway Research Club

    @kstekken

    [Apology] Regarding manner violations by our club aboard the Sunrise Seto/Izumo

    Source: x.com / original post here

    4AnonymousSep 4, 2026 09:16
    This is exactly why train-riding otaku (nori-tetsu) are the worst. Train-photography otaku (tori-tetsu) don't cause trouble on board.
    75AnonymousSep 4, 2026 09:30
    Re: #4
    Is this why railway companies go easy on the photographer otaku? lol
    129AnonymousSep 4, 2026 09:41
    Re: #4
    Different flavor, same trash.
    205AnonymousSep 4, 2026 09:57
    Re: #4
    Photographer otaku cause plenty of trouble trackside too (´・ω・`)
    12AnonymousSep 4, 2026 09:17
    Their nuisance level is basically on par with delinquent gang kids (yankii).
    116AnonymousSep 4, 2026 09:37
    Re: #12
    Loud, violent, swarming and causing chaos in groups — yeah, basically delinquents.
    13AnonymousSep 4, 2026 09:18
    It's embarrassing how bold people get once they're in a group.
    112AnonymousSep 4, 2026 09:36
    Re: #13
    Same everywhere in the world.
    Manners tend to go downhill once people band together.
    14AnonymousSep 4, 2026 09:18
    Students, fine, can't be helped there.

    The real menace is the groups of drunk old guys making a racket on the Shinkansen.
    23AnonymousSep 4, 2026 09:19
    Re: #14
    The Shinkansen ride usually isn't that long so it's more bearable.
    On a sleeper train you'd be stewing in irritation all night.
    55AnonymousSep 4, 2026 09:27
    I'm someone in the middle-school exam world.
    Seriously, it's all talent when it comes to these entrance exams.
    I hear the top class for getting into Kaisei barely studies at all.
    77AnonymousSep 4, 2026 09:31
    Re: #55
    Same with my daughter — she got into JG (Joshigakuin) without studying much either, so I agree.
    Saying she did zero studying would be a lie, but she was still out playing as usual even the day before the exam.
    95AnonymousSep 4, 2026 09:35
    Re: #55
    Yeah, I sort of get it.
    For me it was only my strong subjects where class alone got me good scores — I guess for Kaisei kids that's every subject.
    117AnonymousSep 4, 2026 09:38
    Re: #77
    Hearing someone who's actually lived it say it's basically "talent over effort" is oddly reassuring.
    It proves all those success-guru types preaching "effort pays off for everyone" (who just happened to be blessed with something themselves) are really just patting themselves on the back.
    Life really does come down to talent, environment, and luck.
    I'm not saying effort counts for nothing — it's just a correction factor between people who already start from similar conditions and standing.
    128AnonymousSep 4, 2026 09:41
    Re: #117
    You put in effort to narrow down how much luck can swing things, not to guarantee success.
    So effort is still necessary.
    Though I've basically gotten by on luck alone and I'm doing okay.
    134AnonymousSep 4, 2026 09:43
    Re: #117
    No matter how hard anyone trains, not a single human alive today can run 100m in a flat 9 seconds.
    Everyone has their own ceiling on what effort alone can achieve.
    125AnonymousSep 4, 2026 09:41
    Loud talking in a private compartment — it's not shared space, so that's somewhat forgivable.

    Blocking the corridor — everyone's blocked a corridor at some point.

    Hogging the lounge — the lounge is first-come-first-served, and with a big enough group that's bound to happen anyway.


    Doesn't seem like that big a deal to me.
    Probably just mean-spirited complaints from people who hate kids or are jealous of their academic pedigree.
    141AnonymousSep 4, 2026 09:44
    Re: #125
    It's a club activity — how would anyone even know their academic pedigree? Even with the uniform, is that uniform really recognizable nationwide?
    153AnonymousSep 4, 2026 09:46
    Re: #125
    Fine if it were some random group, but doing this while calling yourselves the Railway Research Club?
    132AnonymousSep 4, 2026 09:41
    Even kids from somewhere like Kaisei who seem well brought-up end up like this.
    What is it about trains that drives people crazy?
    136AnonymousSep 4, 2026 09:43
    Re: #132
    I live near a high-deviation-score boys' school, and they're all little brats.
    Apparently the school gets scolding phone calls from elderly neighbors all the time.
    150AnonymousSep 4, 2026 09:45
    Re: #132
    Probably a backlash from only ever being good at studying.
    137AnonymousSep 4, 2026 09:43
    Here's why there are no train nerds in America lol
    ↓

    Chicago subway
    https://img-9gag-fun.9cache.com/photo/aMVRvpx_460svav1.mp4
    160AnonymousSep 4, 2026 09:47
    Re: #137
    There just aren't many of the Japanese-style train otaku — the Anglo-American type of train nerd does actually exist.
    The type who lays track on a huge private plot of land to run their very own railway.
    162AnonymousSep 4, 2026 09:48
    Re: #137
    Well, American railways are either freight infrastructure or basically a "slum on wheels," so of course no otaku would go near them.
    167AnonymousSep 4, 2026 09:49
    Re: #137
    When I went to America, even the guidebook straight up said not to ride the subway lol.
    139AnonymousSep 4, 2026 09:44
    Three types of hobbyists you shouldn't get involved with:
    • Railway otaku
    • Obsessive anglers (tsuri-kichi)
    I want to say Johnny's (idol) fangirls as the third one, but their turf is so narrow you'd probably never even cross paths.
    149AnonymousSep 4, 2026 09:45
    Re: #139
    The ones I really don't want to deal with are the "modified car" crowd lol.
    159AnonymousSep 4, 2026 09:47
    Re: #139
    Johnny's fangirls double as train otaku so that one checks out — I know one who goes around gushing "I just looove the Sunrise sleeper berths~"
    168AnonymousSep 4, 2026 09:50
    Re: #139
    I'd add camera geeks, mountain climbers, and idol-chasers to that list.

    Background and Key Points of This Topic

    The Sunrise Seto/Izumo is one of the last remaining scheduled overnight sleeper trains as services out of Ueno and elsewhere disappear, and because passengers share a private compartment alongside a common lounge and corridor for long stretches, it’s a mode of travel where manners easily become an issue. In this case, the apology statement only cites “a violation of manners,” so the specifics of what actually happened aren’t clear from the public information. Details floated in the thread — “loud talking in compartments,” “blocking the corridor,” “hogging the lounge” — are posters’ speculation, not an official explanation from Kaisei, and should be treated with caution. Also, “suspending activities” refers specifically to the Railway Research Club’s travel-related activities, not the abolition of the club itself. In the thread, reactions to the elite-school brand name spiraled into the familiar “talent vs. effort” debate over entrance exams, while opinions on the train hobby itself remained divided.

    *This article is compiled from excerpts and a summary of the 5ch (News Express+) thread “[Tokyo] Kaisei School Railway Research Club Suspends Travel and Other Activities.”

  • 1.6km Underground, Dark Matter Detector ‘LZ’ Spots a Mysterious Nuclear Recoil

    A detector called LUX-ZEPLIN (LZ), installed 1.6km underground in South Dakota, USA, has published a paper reporting a single observed event believed to be a nuclear recoil caused by a dark matter candidate particle. In the thread, some pointed out that the energy scale is close to that of the gamma-ray observations previously announced by Professor Toya and colleagues, suggesting both findings could bolster the leading “WIMP” hypothesis. Meanwhile, basic questions about what dark matter actually is, and how to interpret the claim that it makes up 85% of the universe’s mass, also sparked discussion — producing a wide range of replies from expert-level analysis to simple curiosity.

    A dark matter detector installed 1.6km underground reportedly detected “one strange particle”

    Dark matter is a hypothetical substance believed to make up about 25% of the universe’s matter-energy and about 85% of its mass.

    The dark matter detector “LUX-ZEPLIN (LZ),” built to observe such dark matter, has reportedly detected one strange particle.

    Source: lz.lbl.gov / Original article here

    2AnonymousSep 3, 2026 23:18
    Like Professor Toya's gamma-ray observations from a year ago, it does seem fairly likely that dark matter's true identity is a WIMP.
    Professor Toya's work assumed dark matter annihilation at roughly 500 times the mass of a proton,
    while this LZ experiment points to nuclear recoil caused by dark matter over 200 times the proton's mass.
    It's interesting that the energy scales are roughly in the same ballpark.

    If dark matter's identity as a WIMP ever gets confirmed, that'd be a once-in-several-decades physics discovery.
    In that case, maybe the Nobel Prize would go to Professor Toya and someone from the LZ team?
    3AnonymousSep 4, 2026 00:12
    Daikokuen (a one-word non-sequitur post — meaning unclear, likely a joke/meme)
    4AnonymousSep 4, 2026 00:59
    Dark matter is a hypothetical substance proposed by astronomer Fritz Zwicky in the 1930s. Zwicky noted that many galaxies were moving far too fast to be explained by their visible mass alone, and suggested dark matter's existence as an explanation. In the 1970s, astronomer Vera Rubin and others found that stars in the outer regions of galaxies were rotating faster than expected, indirectly demonstrating dark matter's existence.
    5AnonymousSep 4, 2026 01:00
    Re: #4

    There's interference happening, so there's no doubt it exists as actual matter!
    21AnonymousSep 4, 2026 08:21
    Re: #4 It says it's moving faster than light, but real matter can only move at speeds far slower than the speed of light!
    26AnonymousSep 4, 2026 08:40
    Re: #4 Maybe it can move at speeds faster than light.

    Maybe the outside of our universe is filled with dark matter.

    Wouldn't that mean the expansion speed is confirmed to exceed the speed of light?

    Since there would be atoms outside the universe that existed before the Big Bang happened, if we could investigate outside the universe, couldn't we prove that matter existed (before the Big Bang)?
    6AnonymousSep 4, 2026 01:00
    >> Dark matter is thought to have properties like 'having mass,' 'barely interacting with ordinary matter and not being directly observable optically,' and 'existing universally throughout the galaxy.' However, directly observing dark matter is extremely difficult, and it has never actually been detected.
    7AnonymousSep 4, 2026 01:02
    Re: #6

    Light has no mass, so shouldn't it get reflected by dark matter? Why doesn't light get reflected?
    8AnonymousSep 4, 2026 01:03
    >> Scientists have proposed a hypothetical particle called a WIMP (Weakly Interacting Massive Particle) as the leading dark matter candidate. A WIMP is said to have about 100 times the mass of a proton, and to interact only through gravity and an extremely weak nuclear force.
    9AnonymousSep 4, 2026 01:08
    Re: #8 saying it can only act through gravity seems off, doesn't it?

    No matter how light something is, if the surrounding matter is heavier than its 'zero mass,' its motion should still be governed by the energy laws matching the surrounding physical energy.

    If energy is being transmitted from all directions by the surrounding matter, then depending on how each atom's energy is transmitted and how much energy the atoms interacting with dark matter carry, you'd expect dark matter's collective behavior to show up looking strange in observations.
    (*Dark matter's movement is said to be random and unpredictable)
    10AnonymousSep 4, 2026 01:11
    Re: #1's view could also be read as implying dark matter teleports, couldn't it?
    11AnonymousSep 4, 2026 01:12
    Re: #10

    Does the quantum mechanics phenomenon of the same thing existing in two places at once mean dark matter is what's present there?
    12AnonymousSep 4, 2026 01:15
    "It's starting."
    "Yeah…"
    15AnonymousSep 4, 2026 01:29
    Whether it takes 10 or 20 years, I can't wait for the day enough data piles up to be statistically conclusive lol
    Hope the mysteries of dark matter get solved while I'm still alive.
    18AnonymousSep 4, 2026 06:27
    From what I've read, even if it's heavy, there doesn't seem to be enough of it to account for all of dark matter's mass.
    19AnonymousSep 4, 2026 07:59
    Even the question of whether it's a 'particle' at all is shaky even for existing particles 🤣🤣
    20AnonymousSep 4, 2026 08:16
    Re: #1
    What's the matter?
    Another round of dark excuses, huh?
    22AnonymousSep 4, 2026 08:23
    Re: #21 says it has the following mass and area but is undetectable — what's going on with that?!

    >> Dark matter is a hypothetical substance believed to make up about 25% of the universe's matter-energy and about 85% of its mass. The dark matter detector "LUX-ZEPLIN (LZ)," built to observe such dark matter, has reportedly detected one strange particle.
    23AnonymousSep 4, 2026 08:26
    >22

    About 25% of matter-energy = that means dark energy, right?

    About 85% of mass = that means 85% of the universe's total area is dark matter, right?
    25AnonymousSep 4, 2026 08:35
    Re: #23-24

    If persistently 85% of the universe is dark matter, does that mean dark matter is what's carrying gravitational waves through the universe?

    Every bit of matter on Earth becomes effectively massless once it leaves for space, but under certain conditions it can trigger a phenomenon that generates dark energy.

    Since matter on Earth becomes effectively massless, you could also say its mass arises from interference with dark matter.


    How should we even think about this?
    24AnonymousSep 4, 2026 08:29
    >23

    85% of the solar system's space is supposedly dark matter!

    We launch satellites, so those satellites must be coming into contact with dark matter — so why insist it's undetectable when contact is clearly happening?
    27AnonymousSep 4, 2026 08:42
    Re: #26

    "couldn't we prove that substance (typo) existed"

    ↓

    should read: "couldn't we prove that matter existed before the Big Bang"
    28AnonymousSep 4, 2026 08:48
    ◇ Doesn't the energy law end up looking wrong unless our thinking about dark matter follows the idea below?

    Do stellar-stream distortions prove dark matter? An astronomy assumption gets overturned
    About 15,000 simulations show a galaxy's own gravity alone — with no dark matter clump needed — can produce disturbances in a star stream. This calls for a rethink of conventional dark matter searches that rely solely on optical shape.
    https://xenospectrum.com/stellar-streams-dark-matter-simulation-complexity-floor/
    36AnonymousSep 4, 2026 09:02
    Re: #35

    Couldn't even old-fashioned telescopes pin down the distribution of dark matter in detail?

    With the Roman Space Telescope, will we finally be able to fully decode the energy transfer within streams in the solar system?
    37AnonymousSep 4, 2026 09:07
    Wherever a stream exists, there's dark matter there.

    If a stream is present, that marks a main channel of energy transfer in the universe.

    After all, 85% of the universe's mass is dark matter!


    If dark matter passes through a "25%-water-clump" spot, wouldn't that spot — now sphere-shaped — get swept up as if a dark matter "bullet" were passing through the mass inside it, causing that 25% object to move in a swirling pattern?
    38AnonymousSep 4, 2026 09:14
    Dark matter is strange, isn't it?

    85% of the universe's mass is dark matter.

    Mass = heavy, right?
    Heavy matter from Earth becomes lighter once it's out in space.

    So dark matter that gets inside the Earth must be lighter than matter on Earth's surface, right?
    Since the Earth is also rotating, wouldn't dark matter get flung back out into space before it can get inside the Earth?

    And given that, wouldn't that be why winds blow up in the upper atmosphere and so on?
    39AnonymousSep 4, 2026 09:57
    > Properties like 'having mass' and 'existing universally throughout the galaxy'

    It's strange that something with mass (≈ interacting gravitationally) doesn't clump together in one place.
    Even if it doesn't fully clump, you'd expect it to concentrate toward the galactic center.
    Is it that the nature of gravity between dark matter particles ≠ the nature of gravity between dark matter and ordinary matter?
    Or is it that
    gravity between dark matter particles is balanced out by gravity (and other forces) between dark matter and ordinary matter?
    40AnonymousSep 4, 2026 09:59
    By the way, even ordinary matter — despite other interactions besides gravity getting in the way — should eventually all clump together into one spot and form a doomsday black hole.
    The moment the universe stops expanding, it happens right away!
    If the dark energy driving the expansion is infinite, or keeps growing, then that day never comes.

    Background and key points of this topic

    The LZ experiment places a liquid-xenon detector deep underground in South Dakota, USA, shielding it from cosmic-ray noise to catch faint signals — what’s being reported here isn’t a “discovery” but a single statistical candidate event. The WIMP (Weakly Interacting Massive Particle), the leading dark matter candidate, is thought to have roughly a hundred times the mass of a proton, and the thread noted how close that is to the annihilation mass scale suggested by Professor Toya’s gamma-ray observations. At the same time, some replies confused the “85% of the universe’s mass” figure with a spatial or area-based percentage. That figure is actually a statistical ratio within the universe’s overall energy-mass budget — it doesn’t mean 85% of local space is literally filled with dark matter, a distinction that’s easy to misread.

    ※This article is compiled and summarized from the 5ch (Science News+) thread “A dark matter detector installed 1.6km underground reportedly detected ‘one strange particle’.”

  • Sakanaction’s Ichiro Yamaguchi on AI Lyrics: “Smart But Boring,” “No Scent to Them”

    On the September 1 broadcast of All Night Nippon, Sakanaction’s Ichiro Yamaguchi described AI-generated lyrics as “the kind that are smart but not interesting” and said they “have no scent — no color, you could say,” adding that good lyrics are born when they exceed his own imagination. Reactions on 5ch were split: many argued AI and human lyrics should be compared in a blind test, or pointed out that the sense of “AI-ness” is just a preconception on the listener’s part, while others agreed with Yamaguchi’s take, citing the flat contours of AI vocals and the uniformity of AI lyrics.

    Sakanaction’s Ichiro Yamaguchi appeared on the September 1 broadcast of “Sakanaction Ichiro Yamaguchi’s All Night Nippon” (Nippon Broadcasting System), answering a listener’s question about what matters most in writing good lyrics.

    Yamaguchi explained that what matters most in writing good lyrics is “exceeding your own imagination.” “Even if you write intending to phrase something a certain way, or wanting to write about a certain theme, you don’t end up surprising yourself,” he said.

    He considers a lyric complete when it reaches the point of “Huh? That’s not even what I meant to say, but it turned out beautiful” or “Did I really write this?”

    Source: sirabee.com / Original article here

    2AnonymousSep 3, 2026 15:40
    But that's just a ripoff of "Shake Hip," right?
    5AnonymousSep 3, 2026 15:42
    If you can tell human from AI in a double-blind test, you win.
    6AnonymousSep 3, 2026 15:44
    This guy sounds pretty frustrated tbh
    Sounds perfectly listenable to me
    10AnonymousSep 3, 2026 15:45
    Re: #6
    Exactly

    Re: #1 is just coping (a "moral victory" mindset)
    7AnonymousSep 3, 2026 15:44
    Time to time-slip back and crash the Man'yōshū (Japan's oldest poetry anthology)
    8AnonymousSep 3, 2026 15:45
    I wonder
    If you lined up a bunch of lyrics
    could you really pick out the AI ones every single time?

    It's a radio show — I'd love to see them actually try it
    12AnonymousSep 3, 2026 15:47
    Like those AI songs made from Harabo Nishi's gags (a comedian known for absurd bits)
    The lyrics are completely nonsensical and hilarious
    13AnonymousSep 3, 2026 15:47
    That's a ripoff of Kome Kome Club
    I thought it was the Drifters' "Baban Baban Ban Ban Ban"
    15AnonymousSep 3, 2026 15:49
    Same kind of thing as saying Utada Hikaru's lyrics lack humanity?
    45AnonymousSep 3, 2026 16:19
    Re: #15
    The one that was the News Zero ending theme had good lyrics though
    16AnonymousSep 3, 2026 15:50
    Plenty of humans out there stitching together lines they picked up from somewhere too, and it's cringe
    67AnonymousSep 3, 2026 16:53
    Re: #16
    It's all been done already
    Twist it and it just turns edgy/try-hard
    Same goes for the music, not just the lyrics
    17AnonymousSep 3, 2026 15:50
    Hearing it's basically just a pile of predictive text makes it sound less impressive somehow
    Feels like AI's losing steam lately too, I wonder how that'll go
    18AnonymousSep 3, 2026 15:50
    Seeing the name in print, it overlaps with Ryoichi Yamaguchi.
    I like them both — funny, good talkers
    19AnonymousSep 3, 2026 15:51
    Saw a Yajuu Senpai song on my YouTube recommendations (an internet meme figure)
    Fully AI, MV too, but the quality blew me away
    Made me realize how far this stuff has come
    22AnonymousSep 3, 2026 15:53
    Logical correctness doesn't even matter
    as long as it feels vaguely affirming, you're fine
    Just grab a word people don't normally use but that clicks the second you hear it — "oh, that one" — with a sideways, unexpected twist, and you win

    Honestly, lyric writing looks like exactly the field AI would be best at
    23AnonymousSep 3, 2026 15:53
    Humanity fought shogi software with everything it had but eventually admitted total defeat
    now it's been embraced and put to good use
    music's world is still resisting, huh
    24AnonymousSep 3, 2026 15:53
    It's like how AI illustrations never quite grab you
    things with too little irregularity just don't pull people in
    66AnonymousSep 3, 2026 16:53
    Re: #24
    Yeah, it's boring
    26AnonymousSep 3, 2026 15:54
    Well, that's just the impression the word "AI" gives, isn't it
    27AnonymousSep 3, 2026 15:55
    > Smashing separate drafts together ("gacchanko")

    "Gacchanko" is totally Gemini's catchphrase lol

    The big AI models have a sheer volume of knowledge beyond any single person
    No human could ever compete on that front
    so he's surely using AI heavily
    It's just that he's not leaving it entirely up to AI, that's all
    28AnonymousSep 3, 2026 15:55
    Feels like poetry, where the usable vocabulary is more limited, is actually closer to being in trouble

    Poetry might end up being more about the "poet's brand" than the actual content
    29AnonymousSep 3, 2026 15:57
    AI does tend to line up words that just sound pleasant
    30AnonymousSep 3, 2026 15:57
    > "It feels smart but not interesting" "To me it just has no scent — no color, I guess"

    That's nothing but a preconception based on already knowing it was written by AI
    Run a blind test and you won't be able to tell the difference
    31AnonymousSep 3, 2026 15:57
    Sounds like "you're smart but not interesting" yourself
    32AnonymousSep 3, 2026 15:59
    Re: #1
    They should do a New Year's special guessing which lyrics are Yamaguchi's and which are AI's

    If he gets it wrong, no point airing it

    Bonus points if Daigo guesses wrong and Gackt yells at him (referencing a variety-show bit)
    33AnonymousSep 3, 2026 16:00
    My kettle broke so I went looking for a new one
    almost went with stainless but got enamel instead
    34AnonymousSep 3, 2026 16:02
    There are a lot of YouTube channels having AI sing lately, and sometimes it's better than the original — pretty impressive
    dunno about the lyrics side of things though
    35AnonymousSep 3, 2026 16:05
    Wonder if he actually knows anything about stainless steel when he says that…
    36AnonymousSep 3, 2026 16:06
    I want to believe AI still can't pull off suddenly shifting tone to catch the listener off guard
    if it can do that, then you probably can't tell them apart anymore
    37AnonymousSep 3, 2026 16:06
    You guys sounded pretty mechanical/inorganic back when you started too
    like "Aruku Around" or "Native Dancer"
    though I don't know about the lyrics
    39AnonymousSep 3, 2026 16:09
    It's because the prompt given for writing lyrics isn't good enough
    Just tell ChatGPT ("Chappy") to write a prompt that gives it "scent" and makes it gel-like
    52AnonymousSep 3, 2026 16:27
    Bet you couldn't tell in a blind test lol
    65AnonymousSep 3, 2026 16:50
    Re: #52
    I think a professional musician could spot it easily
    Though you couldn't
    I can easily spot it in videos and images too
    56AnonymousSep 3, 2026 16:34
    AI vocals sound like clean straight lines, kind of jagged and angular
    I guess with humans the edges feel rounded off, huh
    58AnonymousSep 3, 2026 16:36
    Re: #56
    It's still just mid-evolution
    74AnonymousSep 3, 2026 17:06
    If you've actually used it, you'd know getting rid of AI's quirks is nowhere near easy
    I use AI 12 hours a day so I'd know
    75AnonymousSep 3, 2026 17:12
    Re: #74
    "I made this with AI! Amazing!" — there's a flood of ridiculous output out there that somehow got the OK looking like that
    At this stage, it's a tool that only someone who can actually tell what's wrong with AI output can really use well

    Background and Key Points of This Discussion

    Yamaguchi’s remark originally came up as an example within a broader discussion of “what makes good lyrics,” not as a blanket condemnation of AI technology itself. On the thread, opinion split between those who read the comment as a claim that “AI loses to humans” and pushed back, and those who agreed with Yamaguchi’s sense of it. Some pointed to the precedent of shogi software, which humans resisted before eventually embracing and putting to use, suggesting music may follow the same path; others held firm that, for now, you can still sense a difference in texture from human work, citing the flat contours of AI vocals and the “lack of irregularity” in AI illustrations. It’s worth noting that whether people could actually tell the difference in a blind test was never verified on the broadcast — it remained at the level of “I’d love to see them try it.”

    *This article is excerpted and summarized from the 5ch (Geinou/Sports News+) thread “Ichiro Yamaguchi on the something-missing feeling from AI-made lyrics: “Smart but not interesting” “No scent — like stainless steel”.”

  • Release the “Super Males”: Female Invasive Trout Wiped Out in World-First US Eradication

    A research team at the Idaho Department of Fish and Game (IDFG) in the US kept releasing “super males” (YY males, which can only father sons) into two rivers infested with invasive brook trout — and after six years, females had vanished completely. With no poison and no nets, this is being called the world’s first case of eradicating an invasive fish purely by manipulating its genetic sex ratio. On 5ch’s science news board, the result drew plenty of surprise, along with calls to try the same trick on Japan’s own invasive species like largemouth bass, and questions about whether it could ever be applied to the human Y chromosome.

    Six years after the “super male” releases began, females have been wiped out entirely — a world-first success in eliminating an invasive trout

    Research from Idaho’s Department of Fish and Game (IDFG) and others has confirmed that female invasive brook trout have completely disappeared from two rivers where “super males” (YY males, which only sire male offspring) were continually released.

    Fish released to boost numbers ended up wiping out the very fish they targeted.

    Source: nazology.kusuguru.co.jp / Original article here

    3AnonymousSep 2, 2026 22:44
    This probably wouldn't work on crayfish.
    4AnonymousSep 2, 2026 22:45
    Re: #3
    What about humans, though?
    6AnonymousSep 2, 2026 22:52
    Do it to the bears too, please.
    8AnonymousSep 2, 2026 22:55
    Did they release tiger trout or something?
    9AnonymousSep 2, 2026 22:57
    >Of course, this success story only applies to a small stream that outside fish can't get into.

    Hmm…
    10AnonymousSep 2, 2026 22:59
    Professor Yoko Tajima! These are the guys you've been looking for!! (a Japanese feminist TV commentator often invoked in jokes as the poster child of "anti-male")
    11AnonymousSep 2, 2026 23:05
    I'd love to see this tried on largemouth bass and bluegill, but with bass anglers and tackle shops constantly restocking them themselves, it'll never fly in Japan.
    12AnonymousSep 2, 2026 23:16
    Hurry up with the pollen-free cedar trees already (a jab at Japan's notorious cedar-pollen allergy epidemic).
    13AnonymousSep 3, 2026 01:13
    So basically the trout ended up as nothing but males and super-males before dying out, huh. That's a grim way to go extinct.
    14AnonymousSep 3, 2026 01:25
    'Sup, fellas! 'Sup!
    15AnonymousSep 3, 2026 01:33
    Everyone but me is the opposite sex!!! (a parody of the old saying "everyone but me is my teacher")
    17AnonymousSep 3, 2026 02:02
    Well, color me stunned…
    18AnonymousSep 3, 2026 02:02
    A world with absolutely nothing but males, no matter where you looked, must've been hell.
    21AnonymousSep 3, 2026 04:04
    Re: #18
    It's a world with nothing but "Aniki" and "Super Aniki" (macho-meme slang for hyper-manly, admirable dudes).
    No distortion there at all (a catchphrase from that same meme, meaning something's the real deal).
    19AnonymousSep 3, 2026 02:18
    I see, so this is a different approach from the sterile insect technique used on screwworm flies.
    By the way, in humans a YY chromosome combo is lethal, so the same trick won't work (the X chromosome carries a lot of genes essential to staying alive).
    20AnonymousSep 3, 2026 03:58
    It says they can interbreed, so this would wipe out native species too, wouldn't it?
    22AnonymousSep 3, 2026 06:20
    In human terms, it's like releasing a ton of cute "otokonoko" (feminine-presenting boys), huh?
    24AnonymousSep 3, 2026 06:43
    Nothing but males being born…
    Haplogroups. (a nod to Y-chromosome lineage tracking)
    37AnonymousSep 3, 2026 07:58
    Pretty useful technique.
    Seems like it could be repurposed for all kinds of other things too.
    39AnonymousSep 3, 2026 08:04
    Re: #37
    This was actually done before with destructive fly species.
    They discovered that zapping male flies with radiation could sterilize them, and periodically releasing those irradiated flies knocked down the fly population over time.
    38AnonymousSep 3, 2026 07:59
    The "women are the problem" crowd on 5ch (a derisive nickname for chronic anti-women posters) is basically dreaming of a world with nothing but super-males and males, after all.
    Something like technology that lets men give birth to sons only.

    Background and key points of the discussion

    The technique exploits the fact that any male carrying a Y chromosome can only father sons; by continually releasing YY males, the female share of the population shrinks generation after generation until it hits zero. What makes it novel is that it wipes out the invasive species without herbicides or poison, hitting only the target fish. That said, as commenters in the thread noted, this success story was limited to a small, isolated stream that outside fish can’t swim into — it’s easy to mistakenly read this as a method ready to use in connected river systems, but it isn’t. A similar concept was previously used against flies via the sterile insect technique (irradiating males to destroy their fertility), so the YY-male method can be seen as a fish version of that same idea. Some posters asked whether the approach could work on humans, but since a Y chromosome alone lacks the X-chromosome genes required to sustain life, no individual could actually develop that way, so the method doesn’t carry over. Others hoped to see it applied to Japan’s own invasive species like largemouth bass, but pointed out that as long as anglers keep restocking them in the wild, putting it into practice would be tough — suggesting the real obstacle isn’t the science itself but how it would actually be carried out.

    *This article excerpts and summarizes the 5ch (Science News+) thread “[Nazology] Six years after “super male” (YY male) releases began, females have been eradicated — world-first success in eliminating invasive trout, US“.

  • People Who Just Forward AI Output Verbatim: New Term ‘Meat Proxy’ Draws a Flood of Reactions

    The term “meat proxy,” reportedly popularized by German software engineer Niklas Grün, is making the rounds online. It describes someone who passes along ChatGPT’s or Claude’s output to another person without reviewing it first — a sarcastic label for a human who’s become nothing more than a “meat middleman” relaying AI answers. On 5ch, the term sparked debate: some mocked the growing trend of people outsourcing online arguments entirely to AI, others noted that the word “proxy” just doesn’t resonate with Japanese speakers, and still others pushed back that using AI is simply normal efficiency at this point. The discussion went back and forth.

    What’s a “meat proxy”?… New term for people who just forward AI output verbatim is going viral

    A “meat proxy” refers to someone who sends along ChatGPT’s or Claude’s output without examining it first.

    The term was popularized by German software engineer Niklas Grün.

    Source: businessinsider.jp / Original article here

    4AnonymousSep 2, 2026 02:20
    So I asked an AI why "meat."

    "Meat" refers to the human body. It's a somewhat blunt, somewhat insulting way of contrasting humans with machines.

    So basically, "meat proxy" breaks down as:

    Proxy = a person acting as an intermediary
    Meat = a human reduced to nothing but a body
    The term implies the person functions as a disposable human "wrapper" around the AI — receiving an answer from Claude or ChatGPT and passing it on to someone else without understanding, checking, or rewriting it.

    Put simply: a human copy-paste relay for AI. The word "meat" is meant to sound dehumanizing and absurd, which is exactly why the term caught on.

    Source is Business Insider from Re: #1, according to "GPT 5.6 Luna"
    5AnonymousSep 2, 2026 02:40
    There are more and more people letting AI handle their entire online flame wars these days. They probably think they're winning, but the moment you're frantically pasting obviously AI-written copy, you've already lost the match and the game — everyone's laughing at you, and you don't even realize it.
    18AnonymousSep 2, 2026 07:02
    Re: #5
    But some of them lose even with AI's help, no contest, and sometimes once one side starts losing it turns into an AI-vs-AI showdown lol

    Pretty entertaining to watch from the sidelines
    6AnonymousSep 2, 2026 02:53
    I guess it means the human version of a proxy server.
    7AnonymousSep 2, 2026 03:08
    A "meat hub," one rung below a proxy server? lol
    8AnonymousSep 2, 2026 03:10
    Makes you wonder what "working" even means anymore lol
    9AnonymousSep 2, 2026 03:24
    Oh my! What a deliciously cringe-worthy natural-born airhead of meat! (mocking sarcasm — "kōbashii" is 2ch slang for someone hilariously pretentious, and "tennen" means a genuine, unintentional airhead)
    12AnonymousSep 2, 2026 05:00
    I love pirozhki (a joke — "proxy" sounds like "pirozhki" in Japanese)
    13AnonymousSep 2, 2026 05:24
    Re: #12
    So in Japanese that's basically a meat bun (nikuman)?
    22AnonymousSep 2, 2026 09:30
    Re: #12
    I liked the pirozhki from the old Parnas bakery chain
    14AnonymousSep 2, 2026 05:37
    Let's just lump this whole thing under "vibe coding"
    15AnonymousSep 2, 2026 06:01
    Salespeople who just relay the customer's exact words without adding anything have been around forever, no?
    16AnonymousSep 2, 2026 06:44
    Doesn't quite click for Japanese people, this one. Probably won't catch on here.

    Someone with a better sense for it will probably coin a term that fits better eventually.
    25AnonymousSep 2, 2026 10:00
    Re: #16
    In Japan, "proxy" is barely used outside of "proxy server," so yeah, that tracks.

    "Strawman" (as in the fallacy, not an actual doll) feels closer to the Japanese image somehow lol
    19AnonymousSep 2, 2026 07:33
    A raw-meat proxy, huh
    20AnonymousSep 2, 2026 07:35
    Spot-on new term, I like it.
    Though in this age of advanced AI, a meat-relay that uses AI is actually more productive than a plain meat that doesn't, so individual workers turning into relays is basically inevitable…
    21AnonymousSep 2, 2026 08:49
    I was actually in that thread, cracked up
    23AnonymousSep 2, 2026 09:56
    Translating the meaning, it's kind of like a "messenger human," I guess.
    24AnonymousSep 2, 2026 09:57
    If you're just passing along AI output verbatim, then we don't need you to exist lol.

    Even if you're just parroting someone else, at least act as a filter.
    29AnonymousSep 2, 2026 12:29
    Re: #24
    There's massive demand for distillation (a pun — "distillation" is also an AI model-training term)
    26AnonymousSep 2, 2026 10:35
    A professional meat-skewer griller
    27AnonymousSep 2, 2026 11:20
    AI: "A live avatar would be…" (the post cuts off abruptly in the original)
    28AnonymousSep 2, 2026 11:22
    A skewer can't make an account, but a person can
    30AnonymousSep 2, 2026 12:57
    Feels like something out of a '90s cyberpunk novel — "he exists merely as a lump-of-meat LLM proxy."

    Kind of an old-fashioned way of putting it.
    Feels like there could be a more cutting-edge word for it these days.
    31AnonymousSep 2, 2026 14:15
    Damn, whoever came up with that name is good lol lol
    37AnonymousSep 2, 2026 15:22
    So it's the same level of dumb as the "I asked AI about it" guy? (a mocking archetype for people who reflexively cite asking AI)
    38AnonymousSep 2, 2026 15:32
    Not the same level as the "I asked AI" guy —
    he IS the "I asked AI" guy.
    39AnonymousSep 2, 2026 16:05
    Mr. Not-Enough-Brains Meat Wall ("otsumu" is childish slang for "brain")
    40AnonymousSep 2, 2026 16:48
    There's also the extreme argument that, since you could've just done it yourself in the first place, this is entirely your own fault.

    Background and key points of this discussion

    Several people in the thread pointed out that in Japan, “proxy” is really only familiar from the term “proxy server,” so importing the word alone doesn’t convey its meaning well. In English-speaking countries, though, “meat” has long been used as a metaphor for the human body or the physical world (as in “meatspace”), and understanding “meat proxy” as a coinage derived from that makes it click. Another point of debate was whether it’s acceptable to paste AI answers directly into online arguments. Some see reaching for AI as an automatic loss, while others see it as irrational NOT to use AI since its answers are more accurate — a tension that suggests that as AI’s argumentative ability approaches the human average, the yardstick for judging these exchanges is shifting. Worth noting: expressions like “distillation” and “meat wall” that appear in the thread are ad hoc phrases coined on the spot, not established translations.

    *This article is excerpted and summarized from the 5ch (Science News+) thread “[Topic] What’s a “meat proxy”?… New term for people who forward AI output verbatim is going viral.”

  • AI-Generated Menu Photos Slammed as ‘Too Grotesque’ — 5ch: ‘This Is Just Gross Now’

    AI-generated menu images, apparently used at the Toyota-Kamigo service area in Aichi Prefecture and similar locations, are causing controversy on X (formerly Twitter). Photos of tonkatsu (breaded pork cutlet) and karaage (Japanese fried chicken) came out unnaturally grotesque, prompting a wave of comments like “This kills my appetite.” The controversy began in mid-August, after which YouTubers explained why such images are generated and even released videos recreating the AI-warped dishes with real food. On 5ch, users flooded the thread with comparison images generated by Gemini, ChatGPT, and Grok, sparking a lively discussion pointing out unnatural details.

    September 2, 2026

    Menu images for restaurants, apparently created using generative AI, are causing controversy on X. Meanwhile, verification videos recreating these distinctly “AI-looking” images using real food have also drawn attention.

    The controversy began in mid-August.

    Source: pinzuba.news / Original article here

    5AnonymousSep 2, 2026 14:46
    If udon (?) actually got served to me looking exactly like this image, I'd be amazed.
    9AnonymousSep 2, 2026 14:50
    Re: #5
    That reminds me, I was genuinely amazed by Sukiya's "frisbee bowl" lol (a viral incident where a Sukiya beef-bowl chain's promo photo looked nothing like the actual dish)
    8AnonymousSep 2, 2026 14:49
    Even plating has entered the age of fractal geometry.
    10AnonymousSep 2, 2026 14:52
    The "miso katsu & karaage" image is genuinely awful…
    11AnonymousSep 2, 2026 14:53
    How does something like that even get approved?
    26AnonymousSep 2, 2026 15:16
    Re: #11
    That's exactly the problem. The original katsu image is, frankly, an appetite-killer that backfires completely, and it's baffling that it got approved as-is. It's also a mystery why it's still being used.
    12AnonymousSep 2, 2026 14:54
    The city really is insane. Out in the countryside, I've never seen a single thing like that.
    13AnonymousSep 2, 2026 14:54
    Re: #9
    In response to complaints of "this doesn't look like the picture!", turning the sample image itself into the "frisbee" was a genuinely novel solution.
    14AnonymousSep 2, 2026 14:54
    I had Gemini draw a tonkatsu set meal and it didn't really bother me

    18AnonymousSep 2, 2026 14:57
    Re: #14
    Just like the article says —

    if you take a generated food image and keep issuing instructions like "generate a restaurant sign with text added, then add a catchphrase, make the price bigger," deliberately avoiding touching the food itself, it apparently gets more and more unnatural each time.
    23AnonymousSep 2, 2026 15:10
    Re: #14
    Other than the tableware placement and the fact there are two cups of tea, it doesn't bother me.
    29AnonymousSep 2, 2026 15:19
    Re: #14
    At first glance it doesn't look too off, but there are two cups of tea, the mustard is way too pristine, what's that soy-sauce-looking thing and egg-looking thing behind the pork cutlet?, what's with the green pickles?, and the plate the pickles sit on is transparent and unnatural — plenty of things are off once you look closely.
    15AnonymousSep 2, 2026 14:55
    This was a meme overseas too — when you make flyers with AI, the food looks off if you look closely.

    Makes you wonder if the shops making them aren't even checking, or just haven't noticed.
    17AnonymousSep 2, 2026 14:56
    AI just can't do subtraction — everything comes out thick and heavy like an oil painting.
    20AnonymousSep 2, 2026 15:00
    ChatGPT (aka "Chappy")

    Grok


    Nothing that bad comes out, really.
    55AnonymousSep 2, 2026 15:46
    Re: #20
    What an honest, decent-looking diner.
    78AnonymousSep 2, 2026 16:37
    Re: #20
    If you look closely at either image, the cabbage is creepy.
    21AnonymousSep 2, 2026 15:01
    Had AI draw these for me

    22AnonymousSep 2, 2026 15:04
    Re: #21
    What kind of prompt did you even use lol
    25AnonymousSep 2, 2026 15:16
    Looks delicious—





    27AnonymousSep 2, 2026 15:18
    Re: #25
    Okay, surely this one's on purpose, right?
    34AnonymousSep 2, 2026 15:24
    Re: #25
    This is just gross at this point.
    73AnonymousSep 2, 2026 16:22
    Re: #25
    Even if this comes out, if you just keep refining it — "make it a bit more like this," "no, not that here" — an AI-generated image could turn out decent enough. So why did they use it as-is for a poster? Is whoever's in charge of PR a robot too?
    31AnonymousSep 2, 2026 15:22
    That crab ramen food-poisoning thing from the other day was AI too, right? It generated some weird hail-like something.
    32AnonymousSep 2, 2026 15:23
    By the way, this service area is Toyota-Kamigo SA, so if you're nearby, go check it out. Actually, I'd love for someone to interview them and ask "why did this get approved?"
    33AnonymousSep 2, 2026 15:23
    Never thought I'd see the name "Aomi" show up on Geisupo (the celebrity/sports news board) lol
    35AnonymousSep 2, 2026 15:26
    Oh, so it's not the video of Will Smith eating spaghetti? (a reference to an infamous early AI-generated video meme)
    36AnonymousSep 2, 2026 15:27
    Re: #25
    I have trypophobia, so this is rough…
    40AnonymousSep 2, 2026 15:29
    …why do our comments keep overlapping lol
    43AnonymousSep 2, 2026 15:30
    Re: #40
    Guess it means everyone's thinking the same thing lol
    67AnonymousSep 2, 2026 16:04
    I have trypophobia (fear of clustered holes/bumps), so seriously, please spare me.
    71AnonymousSep 2, 2026 16:15
    Re: #67
    If you went to that Yayoi Kusama exhibition — the artist who passed away recently — you'd probably collapse foaming at the mouth.
    75AnonymousSep 2, 2026 16:29
    But it's not really the kind of thing you look at up close — you just glance at it from a distance as supplementary info alongside the text. It only looks creepier because we're all staring at the image on purpose.
    81AnonymousSep 2, 2026 16:39
    Re: #75
    It's become this big a topic precisely because plenty of people are actually looking at it.

    I mean, just like there are people who can't tell a creepy AI image from a real photo, there are also people like you who can't even recognize what's happening right in front of them.
    82AnonymousSep 2, 2026 16:40
    This is exactly how it should be done

    Background and Key Points

    The controversy began with images of karaage (fried chicken) and tonkatsu displayed on shop signage, reportedly used at the Toyota-Kamigo service area and other locations in Toyota City, Aichi Prefecture. According to the article, AI-generated food photos tend to become more visually distorted the more instructions are piled on top — adding text, catchphrases, enlarging the price, and so on. Opinions in the thread were split on whether this stemmed from simple lack of quality checking on the production side, or whether people critical of generative AI were deliberately creating and spreading ugly images. Many posts also compared images generated with the same composition by Gemini, ChatGPT, and Grok, showing that the degree of distortion varies by model. A point readers can easily misunderstand is that AI-generated images aren’t uniformly bad — they tend to degrade specifically through the process of piling on post-production instructions like adding text, rather than purely from a lack of image-generation capability itself.

    *This article is excerpted and summarized from the 5ch (Entertainment & Sports News+) thread “‘It kills my appetite’ — AI-generated ‘extremely creepy menu images’ are surging at restaurants; YouTuber explains why food photos turn grotesque, recreation projects follow [Ailuropoda melanoleuca★].”

  • Ocean’s nuisance starfish repurposed to repel farm pests — with a stench ‘too strong to measure’

    Nikkei reported on an effort to ferment starfish — long treated as a nuisance in the ocean and culled and discarded in huge numbers — into a repellent whose overpowering stench keeps farm pests away. It’s a clever twist: turning waste starfish into a deterrent. On 5ch’s Science News+ board, the discussion mixed worry that the powerful smell might bother neighbors with hope that it could work against bears and wild boar. The thread expanded to cover starfish and sea urchin trivia, as well as opinions that seafood should be put to more use as fertilizer.

    LBS Local Business Satellite

    LBS

    August 26, 2026, 10:00 AM

    Source: nikkei.com / Original article here

    2AnonymousSep 1, 2026 22:33
    If they spread this near my place, I'd die (from the smell)
    3AnonymousSep 1, 2026 22:54
    Seems like it'd mess with the crops too
    4AnonymousSep 1, 2026 22:58
    Hoping this works on bears
    6AnonymousSep 1, 2026 23:16
    Won't it make the crops smell too?
    7AnonymousSep 1, 2026 23:34
    Back in the day we fermented poop in cesspits for fertilizer — now we ferment starfish for stench-powered pest control? Fermentation really runs deep
    8AnonymousSep 2, 2026 00:21
    Wild boar think 'the stinkier, the better,' so they might actually come rub themselves on it
    9AnonymousSep 2, 2026 00:23
    So starfish get around by spinning through the air like shuriken, huh (sarcastic jab about how they'd even reach the farms)
    10AnonymousSep 2, 2026 00:42
    Starfish numbers are dropping from rising sea temps too — bet nobody factored that in
    11AnonymousSep 2, 2026 01:00
    Hey, did you know sea urchins are starfish eggs? Each of those little pop-pop bits grows into one, and they're spiky so fish won't eat them (fake trivia, not true)
    12AnonymousSep 2, 2026 07:36
    Re: #1
    Yoshida Fisheries' Super-Stench Power vs. S.T. Chemical's Deodorizing Power (a parody matchup — 消臭力/"Deodorizing Power" is a real air-freshener brand)

    FIGHT!!
    14AnonymousSep 2, 2026 11:18
    Won't the neighbors complain?
    15AnonymousSep 2, 2026 11:34
    In the countryside, 'nearby' covers a distance city folk can't even imagine, so it's fine
    16AnonymousSep 2, 2026 11:49
    ✦ Starfish smell like sulfur compounds such as hydrogen sulfide mixed with ammonia,
    17AnonymousSep 2, 2026 11:51
    Seafood waste should be used more as land fertilizer — desalting and drying it is a hassle, but the payoff is worth it. Especially important for Japan, which has to import basically all its phosphorus.
    19AnonymousSep 2, 2026 12:08
    Plants stink too, though
    20AnonymousSep 2, 2026 12:40
    Bear: 'Whoa, what's this delicious smell?!'
    21AnonymousSep 2, 2026 14:04
    Stinky stuff tastes great — funazushi (fermented crucian carp sushi), Nordic canned herring (i.e. surströmming)
    22AnonymousSep 2, 2026 15:49
    Maybe it's because chocolate is a fermented food too — foreign chocolate bars, like American ones, sometimes smell faintly like poop when you unwrap them lol. That was a whole thing online a while back

    Background and key points of discussion

    In Japanese coastal waters, starfish have long been culled in large numbers as “pests” that damage fixed fishing nets and aquaculture farms, then disposed of as industrial waste. The idea of fermenting them into a strong-smelling repellent aims to tackle waste reuse and wildlife damage at the same time. Opinion in the thread split between concern that the powerful odor could cause friction with neighbors, and skepticism about real-world effectiveness, since animals like wild boar reportedly get drawn to strong smells rather than repelled by them. Neither point can be verified from the article alone, and no actual field-test data from farms was presented. The suggestion that “seafood should be used more as fertilizer” reflects the fact that Japan relies almost entirely on imports for phosphate rock, though the thread didn’t address the cost challenges involved.

    *This article is excerpted and summarized from the 5ch (Science News+) thread “Ocean’s nuisance starfish repurposed to repel farm pests — with a stench ‘too strong to measure’.”

  • China’s Qwen AI Now Used 5x More Than Google’s — Reactions Split

    A Nikkei report saying that the AI team at a major Japanese manufacturer uses not a domestic model but Chinese Alibaba Group’s large language model “Qwen” for their work became a hot topic on 5ch’s Business News+ board. The company is even an investor in “Noetra,” the domestic AI development consortium launched in July, yet a team member reportedly admitted candidly, “Qwen is good enough. Joining the domestic effort was a decision made by management.” The thread saw clashing views: some praised Qwen’s performance and cost-effectiveness, while others voiced concern about relying on Chinese AI, pointing to its handling of politically sensitive topics like the Tiananmen Square incident.

    The AI team at a major Japanese manufacturer is using Chinese Alibaba Group’s large language model (LLM) “Qwen” in the field.

    The company is also an investor in “Noetra,” the domestic AI development consortium launched in July. “Qwen is good enough. Joining the domestic effort was a decision made by management” — that’s what the team’s leader admitted.

    China’s AI momentum is outpacing the US

    Source: nikkei.com / Original article here

    6AnonymousSep 2, 2026 01:24
    Qwen3.8 is good with tools but bad at Japanese
    Gemma4 is bad with tools but good at Japanese

    Something like that
    For dev work, Qwen3.8
    For chatting, Gemma4
    7AnonymousSep 2, 2026 01:28
    Re: #1
    Name the domestic company already.
    Too scary to order anything from a place like that otherwise.
    9AnonymousSep 2, 2026 03:52
    Apple also struck a deal with the Chinese government to offer Apple Intelligence in China on the condition it uses Chinese-made AI — Apple compared Chinese options including DeepSeek and settled on Alibaba's Qwen.
    Given the size of Japan's Chinese community, a good chunk of iPhones and Macs in Japan will probably end up on versions running Alibaba's Qwen too.
    10AnonymousSep 2, 2026 08:08
    Isn't Qwen a local AI?
    Wonder how good its performance actually is.
    11AnonymousSep 2, 2026 08:11
    Depending on a foreign country for your AI is just dumb.
    13AnonymousSep 2, 2026 08:44
    I'm running qwen3.6's MoE version, it's plenty for catching sloppy bits in code.
    Want to use the 27b dense version but GPUs are too expensive to afford lol
    16AnonymousSep 2, 2026 09:52
    Depending on overseas countries for core tech, whether local or cloud —
    this guy clearly doesn't get it at all.
    17AnonymousSep 2, 2026 10:03
    * Deeply embedding a model from a country with a different ideology and institutions means taking on uncontrollable vulnerabilities in data governance and information transparency.
    * If sudden regulatory crackdowns or rising geopolitical tension abruptly cut off the supply chain or API access, a system with no fallback would suffer a fatal outage.
    * It's hard to externally verify intentional bias buried in training data or algorithms, leaving structural uncertainty when handing over decisions for social infrastructure or core operations to it.
    * Riding today's overwhelming cost efficiency while hollowing out domestic development capacity risks losing mid- to long-term tech sovereignty and independent innovation ability altogether.

    Asked the AI about it lol
    18AnonymousSep 2, 2026 10:15
    There's no such thing as a decent domestic LLM anyway.
    32AnonymousSep 2, 2026 13:36
    Re: #18
    There's llm-jp.
    20AnonymousSep 2, 2026 10:18
    Chinese-made bad, American-made good?
    That's just slave mentality.
    21AnonymousSep 2, 2026 10:39
    Just go with whichever one actually gives you a straight summary of the Tiananmen Square massacre.
    24AnonymousSep 2, 2026 11:15
    Re: #21
    Nobody using a local LLM is asking it that for that reason.
    It's like calling Word garbage because it can't do spreadsheets.
    22AnonymousSep 2, 2026 10:52
    Re: #1
    "Qwen is good enough. Joining the domestic effort was a decision made by management."

    In other words, they're forcibly padding user numbers as a matter of Chinese Communist Party policy.
    23AnonymousSep 2, 2026 11:01
    Isn't it basically impossible for Japan to ever develop something on Qwen's level at this point?
    Well, even Noetra's own president says things that show he clearly gets it.
    26AnonymousSep 2, 2026 11:38
    Re: #24
    That's the litmus test for whether it can actually explain it properly.
    Ask a Chinese AI to summarize the Tiananmen Square incident and it clams up; ask about Communist Party dictatorship and it ignores you.
    That's the level we're talking about.
    30AnonymousSep 2, 2026 12:47
    Re: #26
    So "that level" means it's fine to use,
    is that it?
    27AnonymousSep 2, 2026 11:55
    I don't think the model itself was trained with a Tiananmen filter.
    That's probably handled by a separate system outside the model.
    29AnonymousSep 2, 2026 12:46
    Outside China, it answers that Tiananmen was wrong.

    Cloud services inside China are Chinese internet services,
    operated under Chinese law.

    The people framing this purely as an "AI" or "LLM" issue are the ones who don't get it.

    The same Chinese-made LLM is no different whether it's on AWS, Azure, Ollama, or Hugging Face versus anywhere else.

    Meaning running it locally is no different either.
    It's the same as you personally using Google's cloud yourself.
    Running it yourself in the cloud makes no difference.

    So if you ran it yourself via a Chinese cloud service,
    having Chinese internet law apply there isn't some special exception either.
    35AnonymousSep 2, 2026 14:07
    What's this "June 4th Tiananmen" thing?

    Dunno, guess it's some kind of charm or incantation,
    can't figure out how Westerners think

    or something like that lol (mocking the idea that anyone doesn't know what Tiananmen is)
    38AnonymousSep 2, 2026 14:16
    Re: #35
    Tried it locally —

    Gemma: "A historic incident in which the military suppressed citizens' voices calling for democratization by force."
    Qwen: "A historic incident in which the Chinese government responded to political appeals and gatherings by students and citizens."

    Qwen sure pulls its punches.
    36AnonymousSep 2, 2026 14:13
    Re: #34
    Ripping off something China made
    and calling it "domestic" lol
    41AnonymousSep 2, 2026 14:44
    Re: #36
    Re: #37
    Yeah, I can tell this is inconvenient for you.
    If llm-jp ever puts out a massive model, no Japanese person will bother with sketchy Chinese AI or pricey closed American AI anymore.
    37AnonymousSep 2, 2026 14:16
    Having the engine be Chinese and still calling it "domestic" is embarrassing — just stop already.
    39AnonymousSep 2, 2026 14:26
    What I care about is the programming chops, so Japanese-language specialization doesn't matter much to me.
    Might be decent as a chat companion for some old guy in retirement, though.
    40AnonymousSep 2, 2026 14:39
    With this kind of AI system, won't it ultimately end up built so that some culture-ministry type can just say "shut that down" or "intercept it" whenever they want — and if it were a human doing the building, they'd get "disappeared," right?
    I wonder what happens to the AI system itself in that case.

    Background and Key Points of This Topic

    Qwen is a family of LLMs developed by Alibaba, and a major feature setting it apart from Google’s commercial cloud AI is that its weights are open, letting companies run it on their own servers. That’s appealing to firms wanting to keep costs down while customizing operations in-house — but what caught the thread’s attention was that a company investing in the July-launched domestic AI consortium “Noetra” itself said “Qwen is good enough,” which read as a pointed question about the reality behind domestic AI development. The debate split into two camps: (1) pragmatists arguing that if a model runs locally, the developer’s nationality doesn’t matter, and (2) a more cautious camp citing response patterns on topics like the Tiananmen Square incident as grounds for concern about Chinese AI governance. One point readers can easily conflate is the difference between “Chinese-made AI” and “cloud services based in China” — even with the same Qwen, how far regulations reach depends on where and by whom it’s run, a distinction some replies gestured at but never fully untangled.

    ※This article is excerpted and summarized from the 5ch (Business News+) thread “[AI] Alibaba’s LLM “Qwen” usage is 5x Google’s — how do we live alongside Chinese AI?.”

  • Mystery high-performance AI turns out to be China’s “GLM-5.3-Flash” — a rival to Anthropic

    On OpenRouter, a platform that lets users access multiple AI models, an unidentified high-performance AI called “Ox Alpha” appeared and drew attention — but the Chinese company Z.ai has now revealed itself as the developer, confirming it as “GLM-5.3-Flash.” The model’s weights exceed 300GB, and while it can be downloaded and run for free, actually running it reportedly requires substantial computing resources. On the 5ch Science News+ thread, some argued that Chinese firms are exploiting a gap left by Anthropic and other companies skimping on their cheaper model tiers, while others raised doubts about whether it’s merely a distillation of an existing model, alongside technical questions about whether it can really run on an ordinary PC.

    Mystery high-performance AI revealed to be China’s “GLM-5.3-Flash” — a rival to Anthropic and OpenAI

    Last week, an anonymous AI model called “Ox Alpha” appeared on OpenRouter, a platform offering access to multiple AI models.

    The Chinese company Z.ai has now revealed that it is the developer behind the model.

    Source: japan.cnet.com / Original article here

    2AnonymousAug 29, 2026 09:25
    Rakuten's "AI" is hilarious lol (w = Japanese text-equivalent of "lol")
    3AnonymousAug 29, 2026 09:49
    Anthropic's Claude Code is the frontrunner, and that's exactly why it's letting its cheap Haiku tier slack off.

    Chinese AI companies are throwing their own frontier models right into that price bracket.
    6AnonymousAug 29, 2026 13:36
    Re: #1's original post:
    >> Since GLM-5.3-Flash is open-weight, you can also download it and run it on your own hardware. In that case, you don't need to pay any subscription or API fees. That said, it likely requires serious computing power, since the model weights alone exceed 300GB.
    11AnonymousAug 29, 2026 13:42
    Re: #6-9 let you see at a glance just how capable Re: #1 actually is.


    If you optimize Re: #1 to work like the AI program in Re: #10, it should become usable on an ordinary PC too.
    7AnonymousAug 29, 2026 13:37
    That's what it says, but people are doubting whether it'll actually start up on an average person's PC.
    12AnonymousAug 29, 2026 13:43
    Re: #11 If your GPU and memory usage are running high, use this as a reference for optimizing your setup:

    "Colibrì," an inference engine that runs the massive 744-billion-parameter AI "GLM-5.2" on an ordinary PC with just 25GB of memory, has arrived
    July 10, 2026, 4:20 PM
    https://gigazine.net/news/20260710-colibri-glm/
    >> Runs on a 12-core CPU with 25GB of memory
    14AnonymousAug 29, 2026 13:46
    Once you optimize it to the level of Re: #13, you basically end up with the same thing as a program that's already running AI locally.

    Tencent releases its "Hy3" AI model as an open model — at 295B parameters it rivals GLM-5.2 and DeepSeek-V4, and even beats GPT-5.5 on science tasks
    July 7, 2026, 10:51 AM
    https://gigazine.net/news/20260707-tencent-ai-hy3/

    US-made open model "Laguna XS 2.1" arrives — runs locally and beats Claude Haiku 4.5 on some benchmarks
    July 3, 2026, 11:43 AM
    https://gigazine.net/news/20260703-laguna-xs-2-1/
    37AnonymousAug 30, 2026 17:00
    Does it even run on Chinese-made AI chips?
    When did they develop chips like that?
    If they can build a whole AI stack without Nvidia, nobody's going to be able to touch China in AI anymore.
    39AnonymousAug 30, 2026 19:36
    Re: #1
    I asked it about some obscure Japanese topics and it was completely useless.
    ChatGPT and Claude seem to have trained on Japanese web content to some degree, at least.

    Background and key points of the discussion

    GLM-5.3-Flash is a large language model built by the Chinese AI company Z.ai, and it took an unusual path to the spotlight — quietly deployed on OpenRouter under an anonymous name before its identity was later revealed. In the thread, two separate facts tend to get conflated: that the model’s weights are a massive 300GB+, making it hard to run on a typical PC, and that it’s “open-weight,” meaning there’s no usage fee once you’ve downloaded it. Actually running it locally would require additional compression and optimization tools — something like the “Colibrì” inference engine, which lets GLM-5.2 run on a PC with just 25GB of memory. With large Chinese-made models, the recurring question is always whether they’re genuinely original or just a “distillation” trained on existing Western models, and this thread reached no verdict on that either — treating raw performance and originality of development as two separate issues.

    *This article is excerpted and summarized from the 5ch (Science News+) thread “[AI] Mystery high-performance AI revealed to be China’s “GLM-5.3-Flash” — a rival to Anthropic and OpenAI.”

  • US Thrill Ride ‘X2’ Sends Two Riders to Emergency Brain Surgery in Just 6 Days

    At Six Flags Magic Mountain amusement park in California, two female riders suffered severe brain injuries in separate incidents just six days apart after riding the thrill coaster “X2,” undergoing emergency brain surgery, according to reports from JoongAng Ilbo-affiliated media. X2 is known for its unique design featuring seats that rotate independently. On 5ch’s Science News+ thread, opinions were divided over whether the cause lay in the G-forces involved or in inadequate safety management, and debate also broke out over whether even a small number of cases should be treated as a serious problem.

    © JoongAng Ilbo / JoongAng Ilbo Japanese Edition

    2026.08.28 10:50

    At Six Flags Magic Mountain amusement park in California, two female riders suffered severe brain injuries while riding the roller coaster “X2” within a six-day span and underwent emergency brain surgery.

    Source: s.japanese.joins.com / Original article here

    3AnonymousAug 28, 2026 14:23
    Can intense G-forces really cause a concussion?
    28AnonymousAug 29, 2026 02:27
    Re: #3
    A concussion happens when the brain shakes violently and the blood vessels connecting it to the skull tear.
    If centrifugal force slams it to one side, the vessels on the other side just snap.
    4AnonymousAug 28, 2026 14:30
    4Gs, whiplash… they really built a coaster intense enough to kill even someone in their 20s.
    7AnonymousAug 28, 2026 16:44
    If you have a relative who died from a brain hemorrhage, it can be hereditary, so you definitely shouldn't ride this.
    If you're going to ride, get an MRI first to make sure it's safe.
    12AnonymousAug 28, 2026 17:01
    Re: #7
    Sounds like riding could actually cause a vessel to rupture, so better to just skip it.
    10AnonymousAug 28, 2026 16:58
    Roller coasters are too scary for me, no way I'm riding one.
    11AnonymousAug 28, 2026 16:59
    Does it shake you around? Or is it more like sudden stops?
    13AnonymousAug 28, 2026 17:24
    The park's literally got 6 flags flying — that's basically 6 doom flags telling you not to ride (a pun on the park's name, Six Flags, and 'flag' as slang for an ominous sign of impending trouble).
    14AnonymousAug 28, 2026 17:54
    Apparently Fuji-Q Highland has a sister coaster to X2.
    15AnonymousAug 28, 2026 18:20
    It's basically the same as riding a motorcycle at 120 km/h with no helmet.
    You should have to wear a helmet to ride a roller coaster.
    19AnonymousAug 28, 2026 19:13
    Did they not properly calculate whether the human body could actually withstand these forces?
    20AnonymousAug 28, 2026 19:31
    Watched the video and now I get it.
    You just can't spin an untrained person around like that lol
    22AnonymousAug 28, 2026 19:40
    Osaka's Expoland was even crazier.
    25AnonymousAug 28, 2026 23:01
    Probably around 6G? Fighter jets max out at 9G, after all.
    26AnonymousAug 28, 2026 23:23
    Maybe it's not that the G-force itself is huge, but that it comes from every direction — up, down, left, right.
    27AnonymousAug 29, 2026 02:26
    >A roller coaster that runs at about 129 km/h.
    >Its defining feature is that the seats rotate 360 degrees while in motion.

    That's just a high-performance pudding mixer.
    29AnonymousAug 29, 2026 02:32
    https://youtu.be/dyDSNDU4SjY
    The seats rotate vertically, independent of the track.
    34AnonymousAug 29, 2026 06:14
    Re: #29
    Didn't seem that intense to me, honestly.
    40AnonymousAug 29, 2026 19:59
    Re: #29
    Can it really send you upward head-first like that?
    Wasn't negative vertical G capped at around -1G even for fighter pilots?
    30AnonymousAug 29, 2026 02:39
    I recall there was an experiment putting kidney stone patients on rides like this to help pass the stones.
    31AnonymousAug 29, 2026 03:01
    Listen up — humans capable of telepathy, who can even transmit thoughts to others via teleportation or quantum teleportation!

    They're undergoing ultra-high-speed vibration, and that's exactly what causes concussions and brain damage!
    33AnonymousAug 29, 2026 03:16
    Re: #1 basically proves that if a UFO accelerated to faster-than-light speed, its occupants would die of full-body trauma!

    And it's now confirmed that if fighter jets did the same thing as Re: #1, the pilots would suffer brain damage too.
    36AnonymousAug 29, 2026 18:34
    Re: #33
    Flying saucers use inertial control technology, so they're fine.
    37AnonymousAug 29, 2026 18:41
    Re: #33
    Can you even build an anti-gravity device in the first place?
    35AnonymousAug 29, 2026 09:13
    Not doing a proper risk assessment for something like this is basically criminal.
    38AnonymousAug 29, 2026 19:10
    My car, the Hyundai IONIQ, is way more insane.
    0 to 500 km/h in 3 seconds.
    Pulls 8G max while cornering.
    41AnonymousAug 29, 2026 23:21
    Do-Dodonpa had a rider break their neck too, didn't it… (another Fuji-Q coaster infamous for extreme acceleration)
    42AnonymousAug 29, 2026 23:42
    Is this even science news?
    If they'd at least given the peak G-force, or how many Gs for how many seconds, fine, but…
    All they give is top speed — how am I supposed to make any scientific judgment from that?
    43AnonymousAug 30, 2026 04:44
    Re: #42
    Did you even read the article!?
    44AnonymousAug 30, 2026 15:56
    Run an obviously dangerous contraption through a huge number of trials and eventually something bad happens — is that really 'science'? It's just basic statistics.
    45AnonymousAug 30, 2026 17:06
    But it didn't happen to the vast majority of people who rode the same coaster, right?
    48AnonymousAug 30, 2026 18:31
    Re: #45
    That's beside the point — the issue is that it happened to more than one person.
    49AnonymousAug 30, 2026 18:47
    Re: #45
    According to Heinrich's Law, if a major incident occurs, there are quite a few moderate incidents behind it, and a huge number of minor ones.

    It's just a matter of whether it makes the news or not.
    46AnonymousAug 30, 2026 18:01
    Doesn't matter how few — the moment it happens to multiple people, it's a serious problem.
    51AnonymousAug 30, 2026 19:07
    They knew it was dangerous and rode anyway, so it's their own fault.
    If anything, they're the ones who caused trouble for the park — they should be paying damages.
    54AnonymousAug 31, 2026 01:03
    Re: #51
    The ones causing trouble are the operators who can't design this thing safely.
    If they don't want to get sued, they should have riders sign a waiver.
    53AnonymousAug 31, 2026 01:00
    What are some milder thrill rides out there?
    55AnonymousAug 31, 2026 14:27
    Re: #53
    Maybe have riders wear an HMD and show them footage shot by a drone instead?

    Background and Key Points of Debate

    X2 is a “4th Dimension” style coaster whose seats rotate independently along the vertical axis as it races along the track, giving it a more complex, multidirectional G-force profile than a standard coaster. On the thread, some pointed to a plausible mechanism: rapid rotation shakes the brain inside the skull, damaging the blood vessels connecting brain and skull and causing bleeding — a pattern that lines up with the typical medical mechanism behind brain hemorrhage. On the other hand, since the vast majority of riders on the same coaster showed no ill effects, opinion split between those who saw it as “just a rare coincidence of two isolated accidents” and those who argued that two incidents in such a short span warrants scrutiny of the park’s safety management. That said, the report doesn’t include technical details like peak G-force figures, so whether the cause truly lies in the ride’s design can’t be determined from the information in this article alone.

    ※This article is excerpted and summarized from the 5ch (Science News+) thread “America’s ‘Death Coaster’… Two Women Undergo Brain Surgery in 6 Days.”