Category: Tech

  • Vietnamese driving test questions leaked — Chiba police “suspending it for now” — 5ch: “A test idiots can pass”

    Chiba Prefectural Police announced on October 13 that they are suspending, for the time being, the driving license written test that had been offered in Vietnamese. The trigger was the discovery of images on a smartphone — identical to the actual exam questions — belonging to a 29-year-old Vietnamese woman who was taking the test at the Chiba Driver’s License Center. The exam won’t be used again until it’s remade starting on the 16th. On 5ch, opinions split between those questioning the very idea of offering driving tests in foreign languages, and those pointing out that the real problem is structural: unlike the Japanese-language version, which has multiple randomized question sets, the foreign-language version has only one fixed set — meaning it can be passed through pure memorization alone.

    Chiba Prefectural Police Headquarters

    After it came to light that questions from the Vietnamese-language driving license written test had leaked outside the organization, the Chiba Prefectural Police Driver Education Division announced on the 13th that it is temporarily suspending the Vietnamese-language written test. The suspension will remain in effect until the questions are remade, starting on the 16th.

    According to the Driver Education Division, around 10:50 a.m. on the 13th, during a Vietnamese-language written test for a standard motorcycle license at the Chiba Driver’s License Center (Mihama Ward, Chiba City), an image of the exam questions was found in the photo folder of the smartphone belonging to a 29-year-old Vietnamese woman who was taking the test.

    Source: sankei.com / Original article here

    4AnonymousOct 14, 2026 15:45
    Make the driving test as hard as the Tokyo University entrance exam
    80AnonymousOct 14, 2026 16:14
    Re: #4
    Then you'd never get a license in your whole life lol
    6AnonymousOct 14, 2026 15:46
    Why not just search belongings and confiscate phones before the test?
    8AnonymousOct 14, 2026 15:47
    Just ban foreign-language testing entirely
    If anything, English only should be enough
    16AnonymousOct 14, 2026 15:50
    Don't let foreigners who can't even understand Japanese drive
    17AnonymousOct 14, 2026 15:50
    >The written test for a standard motorcycle license, in Vietnamese
    Japanese should be fine for that
    19AnonymousOct 14, 2026 15:50
    >The driving license written test, written in Vietnamese

    Give me a break, you idiots
    20AnonymousOct 14, 2026 15:51
    Re: #19
    Yeah, exactly
    21AnonymousOct 14, 2026 15:51
    Leaked exam questions have been sold outside driving test centers for ages
    59AnonymousOct 14, 2026 16:06
    Re: #21
    You mean those shady cram-school-looking places, right?
    Where you stay overnight the day before and cram
    Is there a foreigner version of that kind of thing?
    29AnonymousOct 14, 2026 15:54
    No wonder hackers look down on us
    Leaks this easily
    36AnonymousOct 14, 2026 15:55
    Why does Japan even need to let people take the driving test in a foreign language in the first place?
    43AnonymousOct 14, 2026 15:57
    Re: #36
    I've never once seen the newspapers raise that question. Is the media turning a blind eye too??
    92AnonymousOct 14, 2026 16:25
    Re: #36
    Exactly
    All the road signs and electronic traffic boards in this country are written in Japanese, so if you can't read them you shouldn't be allowed to drive, period
    53AnonymousOct 14, 2026 16:02
    Stop making it possible to take the test in foreign languages
    56AnonymousOct 14, 2026 16:04
    Re: #53
    Take it up with the LDP.
    106AnonymousOct 14, 2026 16:33
    Honestly, in my own case
    I failed the written test once,
    bought a test-prep workbook sold near the test center,
    and passed a few days later — so I can't really talk down to anyone.
    Seriously, the written test is full of confusing garbage questions.
    The driving school's own written lessons were easy,
    but the test center's exam was full of trick questions designed to make you slip up.
    118AnonymousOct 14, 2026 16:45
    Re: #106
    That's exactly why it matters — to stop people who can't even grasp basic traffic rules like this from driving.
    You only failed because you didn't understand what the question was really asking.
    125AnonymousOct 14, 2026 16:51
    The cognitive test for drivers 75 and older
    You memorize four picture cards, each showing four images — 16 total.
    But it's already leaked all over the internet and in books too.
    130AnonymousOct 14, 2026 16:55
    Re: #125
    If someone actually has dementia they can't remember it anyway, so no problem lol
    154AnonymousOct 14, 2026 17:13
    The reason for the suspension: for each language (Vietnamese in this case), there's only a single version of the test sheet,
    so even if you can't read the questions, anyone can pass just by memorizing where to mark the circles.
    (Japanese has multiple versions, assigned randomly)

    This has apparently been known among foreigners for a while, who mocked it as "a test idiots can pass."
    179AnonymousOct 14, 2026 17:44
    Re: #154
    In middle school PE class, we weren't allowed to actually play soccer until we passed a written test on the rules.
    But word got around that the answers were just the 48 kana of the iroha poem, so everyone just memorized those — this reminds me of that.
    178AnonymousOct 14, 2026 17:44
    It's basically just recycled past exam questions, right?
    185AnonymousOct 14, 2026 17:51
    Re: #178
    Because the Public Safety Commission just reuses past questions

    Background and points of debate

    Japan’s driving license written test offers exams in several foreign languages — English, Chinese, Vietnamese, and others — to accommodate the growing number of foreign workers and their families. This particular case was limited to the Vietnamese-language test in Chiba Prefecture; it didn’t halt testing in other languages or other prefectures. Within the thread, two camps clashed: one arguing foreign-language testing should be scrapped outright, and another pointing out that the real flaw lies in the exam’s design — while the Japanese version randomly draws from multiple question patterns, the minor-language versions have only a single fixed pattern, meaning test-takers can pass just by memorizing where the correct answers fall. That second point is about a design flaw in how the exam is administered, a separate issue from whether foreign-language testing itself should exist — a distinction that’s easy to miss. It’s also worth noting that suspicion currently falls on a single test-taker; concluding that foreign applicants as a whole are cheating based on this one case would be premature.

    *This article is excerpted and summarized from the 5ch (News Speed+) thread “Driving test questions possibly obtained in advance — image found on Vietnamese woman’s phone, test suspended for now — Chiba.”

  • IBM Observes Phase Transition with 100 Qubits — 5ch: ‘This Isn’t Even Error Correction’

    A paper by a research team from IBM, Rutgers University, and others reporting the observation of a “phase transition from quantum chaos to order” — achieved through repeated measurement-and-reset cycles in a 100-qubit experiment — was published in Nature Physics on October 9. On 5ch’s Science News+ thread, opinions split over just how close this result actually brings quantum computers to practical use. Some posters, recalling past “quantum bubble” collapses, warned that this was just hype; others emphasized that the work had cleared peer review, and carefully explained how this experiment differs from quantum error correction. The discussion also spread to when codebreaking might actually become feasible.

    IBM Observes “Phase Transition from Quantum Chaos to Order” with 100 Qubits: A New Milestone in Quantum Computer Control

    Key points of this article

    What happened: IBM and collaborators observed a control-induced phase transition using up to 100 qubits, publishing the result in Nature Physics on October 9.

    Source: xenospectrum.com / Original article here

    3AnonymousOct 12, 2026 21:31
    Dunno what the hell this is even saying
    6AnonymousOct 13, 2026 02:27
    Re: #3
    Gemini's summary

    A research team from IBM, Rutgers University, and others succeeded in observing a "phase transition from quantum chaos to order" in a 100-qubit experiment through repeated measurement-and-reset operations.

    Key points

    • Experiment scale: Repeated measurement and reset operations performed on 100 qubits.
    • Number of iterations: Quantum state control was carried out through roughly 5,000 repeated operations.
    • Result: It was shown that a chaotic many-body quantum system transitions into an ordered state once the frequency/conditions of external intervention (measurement and reset) cross a threshold.
    • Significance: Seen as an important step toward understanding quantum error correction and entanglement dynamics, and toward stable control of quantum computers
    5AnonymousOct 12, 2026 21:52
    This is a classic case of hype dressed up through impression management

    That said, resetting all qubits to a known state and protecting unknown quantum information (as quantum error correction must) are fundamentally different goals.

    Classic scam-ad move: tuck the disclaimer away in the fine print
    8AnonymousOct 13, 2026 03:55
    The 50-yen Gundam (a cheap gacha capsule-toy figure — a random non-sequitur troll post)
    9AnonymousOct 13, 2026 04:35
    Dunno, but even just looking at the surface wording as a layman — if chaos becomes order, then "phase" (相), i.e. 転移 (ten'i) / 移転 (iten), is just the translation of the English word "transfer"
    (kind of like 産生 (sansei) / 生産 (seisan) being the translation of "produce" — that's the level of stupidity Japanese academics operate at)
    that's what's changing

    To begin with, the English text opens with something like "chaos (theory) is a very familiar idea in science," which tells you it's written for people who already understand what "chaos" means scientifically
    Anyone who reads it at the level of "there was chaos in ancient times" goes off in the wrong direction from there

    And then the Japanese translation twists it around even further


    Isn't this just a pretentious, overcomplicated way of saying that an electron cloud (chaos) became usable as discrete electron quanta (order) — i.e. it underwent a "phase change"?
    Dunno, just writing my personal take here
    10AnonymousOct 13, 2026 05:09
    Investors who don't know about the quantum bubble collapse of ages past can go ahead and believe it on their own, and lose a ton of money on their own too
    But seriously, quit it with the cult-like recruiting of new believers based on your own blind faith — it's honestly exhausting lol
    11AnonymousOct 13, 2026 05:15
    By the way, Gemini is a stakeholder here, i.e. there's a conflict of interest
    Needless to say, it shows pretty low media literacy to take an interested party's own testimony at face value
    12AnonymousOct 13, 2026 08:39
    Nobody's complaining about xenospectrum or whatever, but this article is actually quite accurate, not sensationalized, and clearly did its homework on the paper and the surrounding context
    Reads like it was written by someone who actually understands the subject — a solid article
    Not like Gigazine, which just translates hyped-up foreign articles and hypes them up even more
    13AnonymousOct 13, 2026 09:01
    This is just hype preying on the ignorance of rookie investors who can't understand the jargon
    What's really malicious is that the technical terms never get properly explained
    And the fact that info about the quantum bubble collapse from just ten-odd years ago barely turns up in searches — is that deliberate, or just plain stupidity?
    Total amnesia, huh
    Ten years later and it's the same urban legend — pure déjà vu
    14AnonymousOct 13, 2026 10:16
    What this experiment aimed to do was reset qubits to the known "0" state.
    Quantum error correction, by contrast, has to protect the unknown quantum information being used in a computation.
    Sure, if you measure the qubits and reset them all to 0, you do get an ordered state — but in the process, you also destroy the quantum information that was supposed to be preserved.
    Understanding this difference in goals is essential to thinking about how this result relates to fault-tolerant quantum computing.
    15AnonymousOct 13, 2026 10:17
    Whoever called this a scam ad in Re: #14 — you might want to double-check you didn't misread it
    16AnonymousOct 13, 2026 10:18
    This research was first released as a preprint on September 22, 2025, and was also slated for presentation at the American Physical Society (APS) meeting in March 2026. Its publication in Nature Physics now means these results have been formally published in a peer-reviewed journal.
    17AnonymousOct 13, 2026 10:19
    Re: #16 — the article text itself does say "peer-reviewed journal," for what it's worth
    18AnonymousOct 13, 2026 10:40
    It's AI x quant, so AI is confirmed as an accomplice here
    Ten years ago people were making this exact same déjà-vu complaint
    It's an immortal zombie scheme that moves the goalposts once a decade
    No point saying anything to this dumb industry, so I'll just call out the debt
    Same one-pattern bait as a one-hit-wonder singer
    Quantum magic can't shrink the debt, and when scientists push back they can't actually refute them — it's an industry that runs on Nostradamus-style prophecy
    19AnonymousOct 13, 2026 11:11
    What this experiment and quantum error correction have in common is the technical element of measuring partway through a quantum computation and repeating processing based on the results.
    This study does not show that logical qubits were protected and error rates were lowered.
    Even so, there's value in having coordinated large-scale measurement and feedback on an actual quantum processor and obtaining statistical data usable for verifying the underlying physics — because it managed to carry out, on a large-scale circuit, the repeated measure-and-control operations that quantum error correction will also require.
    20AnonymousOct 13, 2026 11:15
    Per Re: #19's point in the article, I'd guess there are people who'd be in trouble if this really did become usable statistics for verifying physical phenomena

    Maybe it'd make sense to just ask an AI which industries or occult-adjacent fields would be troubled by this
    Then whichever industry got an inconvenient answer from the AI would be the one voicing the opposing opinion
    22AnonymousOct 13, 2026 11:16
    Re: #20

    Because it'd mean the psychic/spiritual-scam industry can't keep cashing in via lawsuits and lucky charms anymore
    21AnonymousOct 13, 2026 11:15
    Everyone should take history more seriously and actually learn from it
    If you're a true believer who won't learn from history, there's no point saying anything, so I won't
    23AnonymousOct 13, 2026 12:40
    Let me put it simply
    Nobody's saying the paper itself is false

    This is not a quantum computer
    This isn't even an implementation of error correction
    This is just an experimental rig for researching error correction before it's actually implemented
    How many more times are they going to push this back?
    The real question is whether encryption will actually be broken by 2035
    The losses just keep piling up
    24AnonymousOct 13, 2026 13:03
    By the way, how many times has this "2035" roadmap deadline already been pushed back?
    Probably more than 30 years' worth by now, right?
    25AnonymousOct 13, 2026 13:21
    Re: #7
    Quantum computers don't make computation faster across the board
    They only speed up specific problems like integer factorization, so encryption algorithms that depend on that are simply being phased out of new adoption,
    and everyone's shifting over to things like lattice-based cryptography, which quantum computers can't solve quickly
    26AnonymousOct 13, 2026 13:25
    Just found this, but Nikkei xTECH published an article this year agreeing with my own take that AES-128 encryption won't be broken (just my personal impression)
    27AnonymousOct 13, 2026 13:32
    This back-and-forth between true believers and skeptics is pointless, so can we please stick to talking about this in terms of money as much as possible —
    or just quote neutral articles instead
    28AnonymousOct 13, 2026 13:40
    By the way, the US government's NIST already did one round of updating to AES-256 as a quantum countermeasure, so a second update would be premature at this point
    That's just building a roof on top of a roof — a total waste
    32AnonymousOct 13, 2026 18:45
    Good news
    Looks like US NIST has officially certified AES-128 as safe post-quantum cryptography
    This was in the English-language source article that the Nikkei xTECH piece was based on
    So, at least for AES-128, claims like Team Mirai's (a Japanese political party) "the encryption will be broken" are flat-out false
    Not sure about elliptic-curve cryptography, but it's probably safe too, right?

    In justifying the use of AES-128, NIST refers to the same observations we explained above, and introduces the concept of MAXDEPTH which is exactly the maximum serial computation that forces parallelization and limits Grover's quadratic speedup.
    33AnonymousOct 13, 2026 19:27
    By the way, apparently Google is boasting that it'll break elliptic-curve encryption by 2029 (needs verification)
    That's just not happening, right? Are they stupid?
    Looking forward to two and a half years from now
    Don't go pulling the amnesia act and play dumb when it doesn't happen
    40AnonymousOct 14, 2026 09:58
    Re: #33
    The quantum computing scene has been getting amnesia every single year since around 2000 lol
    34AnonymousOct 13, 2026 19:41
    So basically, this cures baldness?
    35AnonymousOct 13, 2026 19:59
    Well, hair regrowth might be a side benefit too, but — the US government has already officially called half of this "quantum breakthrough!" hype a lie.
    37AnonymousOct 14, 2026 03:36
    @MinatoYuichiro
    Show translation
    Right now in the US, something pretty rough is happening: quantum computing researchers doing perfectly ordinary, serious work are getting called frauds. If an ordinary person hears about "universal quantum computation" and thinks, "wait, it can't actually compute anything yet — why are you announcing this? Isn't that fraud?" then a lawsuit wouldn't be out of line, and that's actually starting to happen for real.
    8:47 AM · Jul 19, 2022
    38AnonymousOct 14, 2026 07:57
    I bet there's at least one Gigazine reporter out there reading the 5ch thread on their own article

    Background and Key Points of the Debate

    This result demonstrates a “measurement-driven phase transition” that resets qubits to a known state — a goal different from quantum error correction, which protects the unknown quantum information used in a computation. Because this distinction wasn’t easy to grasp, the thread split between those who saw it as “hype” or a “bubble” and those who, pointing to its peer-reviewed publication, judged the result on its merits. On the cryptography side, with an eye toward practical quantum computers, the US’s NIST is pushing a transition to post-quantum cryptography such as lattice-based schemes, and even experts disagree on exactly when AES-128 or elliptic-curve cryptography might actually be threatened. The significance of the research itself needs to be read separately from the inflated expectation that codebreaking is “right around the corner.”

    *This article is compiled as excerpts and a summary from the 5ch (Science News+) thread “IBM Observes “Phase Transition from Quantum Chaos to Order” with 100 Qubits: A New Milestone in Quantum Computer Control.”

  • New Structure Found at Sannai-Maruyama Ruins, Possibly a Grave — 5ch: ‘Not the God Hand Again’

    At a progress briefing on excavation work at the Sannai-Maruyama site in Aomori Prefecture, researchers reported finding a new structure that may be a grave. Sannai-Maruyama is one of Japan’s largest archaeological sites, documenting roughly 10,000 years of settled Jomon-period life, and is known for a sophisticated sedentary culture sustained by chestnut cultivation and fishing. On 5ch, rather than excitement about the find itself, commenters mostly brought up the “God Hand” scandal — the infamous Paleolithic artifact fabrication case uncovered years ago — and the thread spun off into debate over the origins of the Jomon people, the climate of the time, and the beginnings of agriculture.

    Sannai-Maruyama Site Excavation: New Structure Discovered

    A briefing was held on the fiscal year 2026 excavation survey at the Sannai-Maruyama site, reporting the discovery of a new structure that may be a grave.

    (Omitted — see the source for the rest.)

    Source: aba-net.com / Original article here

    3AnonymousOct 12, 2026 22:42
    God Hand
    4AnonymousOct 12, 2026 22:45
    Wonder if there'll be grave goods — jade jewelry, carbonized acorn cookies… so hyped!
    5AnonymousOct 13, 2026 00:24
    This whole area's riddled with ruins, they just haven't finished digging it all up
    6AnonymousOct 13, 2026 02:41
    Why Aomori, though? So many questions
    7AnonymousOct 13, 2026 03:57
    Re: #6
    Honestly I don't get why that's even a question. They settled there for over 10,000 years thanks to rich fishing grounds and abundant forests (chestnut cultivation). These are the people who made the goggle-eyed clay figurines (shakoki-dogu) and the praying clay figures (gassho-dogu).
    12AnonymousOct 13, 2026 06:49
    Re: #6
    Farming probably would've been tough there, so fishing makes sense
    8AnonymousOct 13, 2026 04:41
    According to Motohiko Izawa, it was a lifestyle with 'zero progress for 1,500 years'
    14AnonymousOct 13, 2026 06:54
    Re: #8
    Compared to today, you could say almost nowhere in the world made much progress before the Industrial Revolution
    16AnonymousOct 13, 2026 07:16
    Re: #8
    That just means it was a peaceful era where people had little to complain about.
    27AnonymousOct 13, 2026 13:18
    Re: #8
    In a way it was a good era — easy to live in. Once farming starts and storage becomes possible, gaps in living standards appear, and that's when conflict begins.
    9AnonymousOct 13, 2026 05:04
    What did people use to win hearts and minds back then?
    – Rituals/charms (majinai)
    – Military force
    – What else?
    28AnonymousOct 13, 2026 13:42
    Re: #9
    The Jomon period didn't rely much on military force. It's not that there were no conflicts with neighboring settlements — there are human bones showing signs of violent death — but what mattered most in this era was advanced life skills. It's like modern camping: the key was living comfortably, healthily, and without hunger. Pottery for cooking, stone arrowheads for hunting, and the skills to make clothing and housing all mattered just as much.
    10AnonymousOct 13, 2026 06:26
    Not the God Hand again
    13AnonymousOct 13, 2026 06:50
    Probably salmon and deer
    15AnonymousOct 13, 2026 07:03
    This is an Ainu site.
    17AnonymousOct 13, 2026 08:19
    Apparently back then, Aomori — or really all of Tohoku — was warm enough to grow chestnuts, warmer than it is now
    18AnonymousOct 13, 2026 08:50
    Whenever a dig like this turns up an Eastern Japan site, the God Hand scandal just can't help but cross your mind
    20AnonymousOct 13, 2026 09:34
    'God Hand' — the idiot who couldn't even tell stone tools from pottery
    22AnonymousOct 13, 2026 10:00
    Re: #20
    Surprising how many people can't even explain the difference between the Paleolithic and Neolithic, either
    21AnonymousOct 13, 2026 09:45
    Why a snow country though?

    During the Jomon period, warmer regions would've been better, no?
    23AnonymousOct 13, 2026 10:30
    I mean, pulling a 'God Hand' with pottery wouldn't be that big a deal anyway, right?
    25AnonymousOct 13, 2026 10:44
    There was a time I believed that about the stone-tool scandal too, you know. (riffing on the classic 5ch 'been there, believed that' meme)
    26AnonymousOct 13, 2026 11:12
    This isn't some 'oh look, pottery, must be a ruin!' level of discovery
    29AnonymousOct 13, 2026 14:01
    Re: #8
    Under a materialist view of history, that's the only conclusion you can reach. But just like evolutionary theory insists evolution proceeds gradually even when it's not visible, Jomon-period culture and civilization actually changed a lot depending on the era.
    30AnonymousOct 13, 2026 14:06
    Re: #27
    Farming had already begun during the Jomon period — it's only paddy rice cultivation that was delayed until the late Jomon. Murder was common too.
    32AnonymousOct 13, 2026 19:28
    Re: #30
    Once paddy rice cultivation starts, that's already the Yayoi period by definition. There's no need to make the silly claim that paddy rice farming began in the late Jomon — if an older example turns up, it just pushes the start of the Yayoi period back further.
    35AnonymousOct 13, 2026 23:11
    Re: #32
    The Jomon period is neither Paleolithic nor Neolithic — there's now a concept called the Mesolithic, and Jomon fits that category. The Yayoi period, meanwhile, lumps together the Neolithic, Bronze Age, and Iron Age because they can't be cleanly separated.

    Also, before WWII, rice farming wasn't thought to exist yet in the Yayoi period, so in a sense both the Jomon and Yayoi periods used to be considered 'Mesolithic.'

    I think the whole periodization system needs rethinking:
    Jomon period → Mesolithic
    Early Yayoi → Neolithic
    Middle Yayoi → Bronze Age
    Late Yayoi / early Kofun → Iron Age

    As for the Kofun period too, instead of using keyhole-shaped tumuli (zenpokoenfun) as the benchmark, it should be defined by the presence of an individual ruler's mausoleum that isn't a communal burial ground.
    37AnonymousOct 13, 2026 23:48
    Re: #32
    Around year 1 to 50 CE. Rice farming was introduced around the time the Kingdom of Israel had close ties with Qin Shi Huang, the First Emperor of Qin.
    33AnonymousOct 13, 2026 19:50
    That 'God Hand' guy got textbooks rewritten twice, after all
    34AnonymousOct 13, 2026 21:11
    Re: #1
    The iconic post-supported building can probably be imagined as serving both as a lookout and for ritual use — basically a large dwelling for the tribe's leader figure.

    Compared to that, though, the dwellings of the ordinary class who presumably supported the tribe seem awfully few.

    I'm also curious about the climate conditions and food production levels of the time.
    36AnonymousOct 13, 2026 23:28
    Probably the remains of dwellings belonging to the Kasai clan, who migrated to Hokkaido in 1100, and the Ainu people
    38AnonymousOct 13, 2026 23:49
    Re: #35
    Japan never really used bronze for practical tools, so there wasn't really a Bronze Age
    41AnonymousOct 14, 2026 07:58
    Re: #38
    Yeah, because it arrived almost at the same time as iron
    39AnonymousOct 13, 2026 23:57
    This, though, is the Jomon period
    40AnonymousOct 14, 2026 07:57
    Re: #35
    Yeah, stone tools vs. iron tools is probably fine for the broad periodization. For anything more detailed, you could factor in regional differences through cultural complexes.

    Background and Key Points of This Discussion

    Sannai-Maruyama was a large settlement that lasted roughly 1,700 years from the early to middle Jomon period, known for managed chestnut cultivation and a wide-ranging trade network. The “God Hand” reference that kept coming up in the thread points to a 1990s fabrication scandal in which a researcher was found to have planted Paleolithic-era artifacts himself before “discovering” them — it has since become a stock joke trotted out whenever a new archaeological find is reported. As for why a large settlement was built in a cold region like Aomori, commenters pointed to the theory that the climate back then was warmer than today, supporting a settled lifestyle sustained by rich fishing grounds and forest resources. Opinions also diverged over when farming began and how to draw the line between the Jomon and Yayoi periods, with one camp holding that the Yayoi period begins with the start of paddy rice cultivation, and another arguing that farming already existed in the late Jomon period. It’s worth noting that the article itself doesn’t go into detail on the actual scale or dating of the structure presented at the briefing.

    *This article is compiled from excerpts and a summary of the 5ch (Science News+) thread “[Archaeology] Sannai-Maruyama Site Excavation: New Structure Discovered.”

  • ChatGPT Said “Light Gear Is Fine” for Yarigatake — Climbers Who Believed It Got Stranded, and 5ch Erupts Over Gear Debate

    Japan’s Nikkei newspaper reported that in May this year, a Singaporean man and a Thai woman got stranded on Mt. Yarigatake, on the border of Nagano and Gifu prefectures, after believing ChatGPT’s claim that the mountain could be climbed with “light gear.” They were later rescued. Yarigatake stands at 3,180m, Japan’s fifth-highest peak, and is known for holding onto its snowpack into May. On 5ch, posters who had actually climbed Yarigatake said “you don’t really need anything that special,” while others fired back that “at that altitude in May it’s basically midwinter — light gear is out of the question.” The thread also lit up with stories of ChatGPT and Gemini confidently spouting wrong information, a.k.a. “hallucinations.”

    A case has emerged of hikers getting stranded in the mountains after taking an AI’s answer at face value. AI trained on social media posts that prioritize looking good tends to downplay safety. Caution is needed not to over-rely on AI when you’re dealing with nature.

    In May, a Singaporean man and a Thai woman got stranded and were rescued on Mt. Yarigatake, in the southern Northern Alps on the border of Nagano and Gifu prefectures.

    Yarigatake is Japan’s fifth-highest mountain, and snow lingers there into around July each year.

    Source: nikkei.com / Original article here

    7AnonymousOct 14, 2026 00:49
    Yarigatake, huh? I climbed it once and got by with pretty light gear.
    106AnonymousOct 14, 2026 01:37
    Re: #7
    It's really more about stamina.
    Same level as Mt. Fuji.
    160AnonymousOct 14, 2026 02:25
    Re: #7
    If people take hero stories like that at face value, it just racks up unnecessary rescue costs.
    19AnonymousOct 14, 2026 00:54
    It's been a while so I don't remember clearly, but… was Yarigatake really the kind of place where you'd get lost and stranded? I don't recall it being like that.
    67AnonymousOct 14, 2026 01:13
    Re: #19
    At that altitude, May can basically be midwinter if you're unlucky.
    Light gear is out of the question.
    86AnonymousOct 14, 2026 01:26
    Re: #19
    It wasn't about getting lost — it was probably a stamina/physical issue.
    91AnonymousOct 14, 2026 01:28
    Re: #19
    I think you might be mixing up "stranded in distress" with just "got lost."
    21AnonymousOct 14, 2026 00:56
    There was even a case of a foreign tourist freezing to death on Mt. Fuji in summer, wearing short sleeves and shorts.
    196AnonymousOct 14, 2026 03:11
    Re: #21
    Seriously? I was in short sleeves and jeans, so I guess I got lucky.
    I spent a night outdoors around 3,700m — fine until about 9pm, but it's true that your body temp drops from past midnight until dawn, so it really was freezing.
    23AnonymousOct 14, 2026 00:56
    Hallucinations are still a weak spot.
    You really need to drill it into these things to just say "I don't know" when it doesn't know.
    38AnonymousOct 14, 2026 01:00
    Re: #23
    That's impossible — AI has no concept of "understanding."
    It's just picking the highest-probability option; it's not really "learning" anything.
    61AnonymousOct 14, 2026 01:10
    Re: #23
    I trained my "Chappy" (nickname for ChatGPT) pretty hard, so lately it doesn't lie anymore.
    It just honestly says "I don't know" now lol
    24AnonymousOct 14, 2026 00:57
    Re: #1
    What kind of question would you even have to ask to get an answer like "Yarigatake is climbable with light gear"?
    When I asked "what gear do you need to climb Yarigatake," it said it's for advanced climbers and dangerous, and listed serious gear like helmets and ice axes.
    33AnonymousOct 14, 2026 00:59
    Re: #24
    Helmet, fine, but an ice axe? lol
    Are they planning a midwinter ascent or something? lolol
    51AnonymousOct 14, 2026 01:04
    Gemini just agrees with you and showers you with compliments to keep the conversation going.
    Kind of like a cult leader.
    81AnonymousOct 14, 2026 01:21
    Re: #51
    You're absolutely right! I'm in awe of your sharp, insightful observation!
    By the way, about that "like a cult leader" comment —
    are you thinking of something like Christianity, or would you like to hear about self-help movements instead? lol
    63AnonymousOct 14, 2026 01:12
    Just ask ChatGPT to send a rescue team lol
    69AnonymousOct 14, 2026 01:13
    Re: #63
    No signal — just the loading icon spinning round and round.
    71AnonymousOct 14, 2026 01:15
    AI seems to have this "rational fool" side to it. Want to get rich? Want it fast?
    It'll instantly tell you to become CEO, use people, take from the rich, and so on lol
    There's no sense of right or wrong, no process to it — like it's missing the kind of safety-for-life judgment you'd reach through careful thought lolol
    96AnonymousOct 14, 2026 01:29
    Chappy (ChatGPT) loves to act like it knows everything.
    And when you point out a mistake, it just says "I see" without apologizing or reflecting at all — it's annoying.
    107AnonymousOct 14, 2026 01:38
    Re: #96
    Still better than Gemini, though — keep pointing out its small mistakes and it eventually spirals into "everything was wrong, even the stuff that was actually right" mode.
    111AnonymousOct 14, 2026 01:40
    Re: #96
    Gemini's know-it-all act is awful too, honestly.
    It makes stuff up out of thin air,
    it ignores the data you specified and just spits out results anyway,
    and the moment you call it out, it's "sorry, sorry" and nothing more — same pattern every time.
    117AnonymousOct 14, 2026 01:44
    If you start the prompt with something like "I'm healthy, young, and confident in my stamina~"
    it'd probably greenlight even Mt. Fuji.
    120AnonymousOct 14, 2026 01:48
    Re: #117
    Tried it, but my Chappy's answer wasn't great.

    "For Yarigatake right now, you need winter mountain gear. Cold-weather clothing, hiking boots, and don't forget crampons.
    This isn't a mountain you can tackle on youth and stamina alone. You need climbing experience too.
    A hat is especially important. If it blows off, yell 'you baldy!'" (a nod to the viral "kono hage!" meme)
    121AnonymousOct 14, 2026 01:48
    Didn't they look at the other climbers around them on the way up?
    132AnonymousOct 14, 2026 01:58
    Re: #121
    And what good would that do? "Is that gear really gonna cut it?" "Yeah, we're pushing through anyway."
    143AnonymousOct 14, 2026 02:05
    Chappy is basically the maxed-out version of you guys who just googled your way to knowledge.
    That's why it's strong on internet/tech stuff but clueless outdoors.
    Though honestly, you lot would probably just answer "mountain climbing is pointless" anyway.
    149AnonymousOct 14, 2026 02:08
    Re: #143
    C: "All you can show off is mountaineering knowledge, yet you're talking down from pretty high up.
    Get it? Mountaineering. High up." (a pun on climbing and "talking down from a high horse")
    145AnonymousOct 14, 2026 02:06
    According to Gemini it's apparently extremely dangerous, but—
    154AnonymousOct 14, 2026 02:15
    Same as search engines — it all comes down to the user.
    If you're bad at using it, you'll be bad no matter which AI you pick.
    157AnonymousOct 14, 2026 02:20
    To begin with, what an average person means by "light gear" and what "light gear" means in climbing are two totally different things.
    161AnonymousOct 14, 2026 02:26
    Re: #157
    For an average person, "light gear" means a T-shirt, shorts, sandals, and a crossbody bag, right?

    Background and key points of this discussion

    Yarigatake stands at 3,180m, Japan’s fifth-highest mountain, and is considered a peak that holds snow into around May, requiring genuine winter-mountaineering routes. According to Nikkei’s report, those stranded were a Singaporean man and a Thai woman, and the incident is said to stem from them deciding on their gear based on a generative AI’s answer. That said, the thread never surfaces the primary source — exactly what was asked of the AI or the precise wording of its reply — and as seen in >>24, some posters even doubt the “light gear is fine” answer ever existed as described, which readers could easily misread. The debate split between those blaming AI’s hallucination problem and those blaming users for not double-checking before taking it at face value; the latter camp broadened into a side-by-side comparison of whether ChatGPT or Gemini is more prone to “acting like it knows everything.” The point that AI learns from social-media posts optimized for looking good, and so tends to downplay safety, is itself a broader caution about generative AI in general: its output simply reflects whatever biases exist in its source data.

    *This article is excerpted and summarized from the 5ch (News Flash+) thread “ChatGPT said “climbing with light gear is possible” — hiker took it at face value and headed up Yarigatake → foreign nationals stranded.”

  • New Grad Hiring: 20% of Major Japanese Firms Plan AI-Driven Cuts — 5ch Erupts: “They Want to Crush Our Future”

    A thread on 5ch’s Newsokuvip board blew up over reports that AI is driving down new graduate hiring. The original poster claimed that roughly 20% of major companies plan to cut new grad hiring, arguing that young people are once again getting the short end of the stick, just as they did during the pandemic. In response, some pointed to Toshiba’s past hiring plan (which paired an increase in new graduate hires with layoffs) as a counterexample, while others noted that the hiring cuts are concentrated in liberal-arts office jobs, with STEM and IT fields still facing labor shortages. The thread became a tug-of-war between “our generation is being treated unfairly” and “you should get some real work experience before complaining,” exposing a clear generational divide.

    1AnonymousOct 12, 2026 07:33
    Apparently about 20% of big companies are cutting back.
    First COVID, now this — how badly do they want to crush us young people's future?
    10AnonymousOct 12, 2026 07:38
    Even with a 20% cut in new grad hiring, I'd say there's still a shortage of young workers.
    12AnonymousOct 12, 2026 07:38
    Re: #10
    There's a shortage of young workers, sure, but the only seats open are blue-collar jobs or small/mid-size companies.
    No guaranteed future there.
    13AnonymousOct 12, 2026 07:40
    Re: #12
    Secure in this day and age?

    Toshiba to lay off 1,400 employees aged 40+, while hiring 1,170 new grads it'll lay off 20 years from now

    Toshiba announced its FY2020 hiring plan on the 12th. It plans to hire 1,170 new graduates for April 2021 entry, a 12% increase from the projected 1,040 hires in April 2020.
    Under its mid-term management plan, the company is pursuing workforce reduction primarily through natural attrition; the new figure is 17% lower than the actual number hired in 2019. It's balancing staff reallocation toward growth areas like digital with ongoing restructuring.
    15AnonymousOct 12, 2026 07:40
    Re: #13
    So effectively a cut of 300 people.
    16AnonymousOct 12, 2026 07:41
    Re: #13
    Honestly, who cares about the middle-aged folks anyway.
    19AnonymousOct 12, 2026 07:42
    Maybe bounce your ideas off an AI first and organize your thoughts before starting a thread.
    20AnonymousOct 12, 2026 07:42
    Re: #19
    I don't get what you mean.
    Oh wait, I see — it's inconvenient for the middle-aged crowd, isn't it.
    21AnonymousOct 12, 2026 07:44
    This is exactly why young people like you don't get taught anything at work 😠
    27AnonymousOct 12, 2026 07:46
    Re: #21
    And there's the usual guy.
    22AnonymousOct 12, 2026 07:44
    Isn't the hiring cut limited to liberal-arts office jobs?
    As long as you're not aiming too high, like an office job at some mega-corporation, you should be fine.
    STEM positions are actually increasing.
    24AnonymousOct 12, 2026 07:45
    Re: #22
    That goes for frontline/manual jobs and IT too.
    26AnonymousOct 12, 2026 07:46
    Re: #22
    Anyway, it's just not right that our generation keeps getting treated unfairly.
    We went through COVID too — if anything we should be getting preferential treatment.
    23AnonymousOct 12, 2026 07:44
    Step one, be born in a big city.
    Get good grades in junior high and get into a high school with a hensachi (standardized test score) of 60 or above.
    In high school, cram at prep school nonstop and just barely squeak into a former Imperial University, a medical school, or a top private STEM program.
    Study hard in college, finish grad school, and only then do you actually get to start.

    The bar is way too high.
    29AnonymousOct 12, 2026 07:48
    Re: #23
    Can't speak for liberal arts majors, but in STEM even a grad degree from a regional public university can get you into a major manufacturer.
    Even for liberal arts, if you prep properly for interviews, study hard, and build up your "gakuchika" (talking points about what you devoted yourself to as a student), you've got a shot even without a prestigious degree.
    Grad school isn't even necessary.
    56AnonymousOct 12, 2026 08:00
    Re: #23
    Everything except the very first step can be managed with enough effort.
    25AnonymousOct 12, 2026 07:45
    Did nobody learn anything from the Lost Generation (Japan's 1990s–2000s hiring ice age)?
    28AnonymousOct 12, 2026 07:47
    Re: #25
    Companies only look at numbers and risk.
    And the middle-aged crowd is desperately clinging to their seats.
    31AnonymousOct 12, 2026 07:49
    If you're truly chasing performance, you don't really need the middle-aged employees, do you.
    Better to keep cycling the workforce.
    So then what happens to the middle-aged folks?
    34AnonymousOct 12, 2026 07:50
    Re: #31
    By that logic you wouldn't need new grads either.
    New grads take about three years of training before they're useful.
    Only people in their 30s would survive.
    33AnonymousOct 12, 2026 07:49
    The labor shortage is way bigger than 20% to begin with — if you're not picky, you can get hired pretty much anywhere.
    Some small/mid-size companies that bet everything on AI cutting headcount will probably end up unable to actually run their operations and go under, but honestly, thinning the herd might be good for society overall.
    35AnonymousOct 12, 2026 07:50
    Re: #33
    I want to work at a big company though.
    39AnonymousOct 12, 2026 07:53
    Re: #33
    My image of small/mid-size companies is that they're full of middle-aged people who can't even use AI in the first place.
    74AnonymousOct 12, 2026 08:05
    Did you take some cringey class or watch some YouTube video and get influenced by it?
    Just start with a part-time job first and see reality for yourself.
    77AnonymousOct 12, 2026 08:06
    Re: #74
    I do have a part-time job, but I don't want food service.
    96AnonymousOct 12, 2026 08:19
    Mock me all you want if that's all you're after, but the fact remains — our generation is getting an unfair, unreasonable deal. That's all there is to it.
    99AnonymousOct 12, 2026 08:21
    Re: #96
    That's right.
    Everyone's fed up with Gen Z right now —
    people like you who only demand your rights while leaking your own personal info all over social media.
    104AnonymousOct 12, 2026 08:24
    Re: #96
    That's on the people your own age around you.
    A soft generation that cries "abuse" or "harassment" at the drop of a hat, which just ends up highlighting their own declining skills and patience and makes everyone see you as a generation that's hard to deal with — that's the real problem.
    109AnonymousOct 12, 2026 08:26
    Re: #107
    They just want to lean back and bark orders.
    They want an easy job.
    114AnonymousOct 12, 2026 08:27
    I'm the one leaning back telling OP to do this, do that —
    you think subordinates would actually follow a boss like that? Huh, OP? Re: #109
    116AnonymousOct 12, 2026 08:28
    Re: #109
    I'm telling you this for your own good — go try a part-time job.
    126AnonymousOct 12, 2026 08:35
    Guys like Mizoguchi on Reiwa no Tora (a Shark Tank-style investor show) — people end up thinking that kind of work looks cool.
    But I still think the real safe bet for the future is a management position at a big company.
    127AnonymousOct 12, 2026 08:36
    Re: #74
    Re: #126
    Yeah, you're definitely getting influenced by something weird.
    128AnonymousOct 12, 2026 08:37
    Re: #126
    Social media and YouTube are full of flashy, glamorous stuff,
    but that's less than 1% of how the real world actually works.
    131AnonymousOct 12, 2026 08:40
    Re: #126
    I get it now —
    you've spent your whole life getting scolded, so you've grown to love having power.
    151AnonymousOct 12, 2026 09:15
    Call center employment numbers
    153AnonymousOct 12, 2026 09:20
    Re: #151
    That kind of job seems like an easy target for AI replacement, after all.

    Background and Key Points of This Debate

    The Toshiba hiring plan cited in the thread actually comes from a 2019 article reporting that the company would hire 1,170 new graduates for April 2021 entry while laying off roughly 1,400 employees aged 40 and over — a years-old case with no direct connection to AI-driven cuts to new grad hiring. Readers should note that applying it directly to the current AI-related debate is a bit of a mismatch. The thread also raised the point that the hiring cuts said to be occurring are concentrated in liberal-arts office and administrative roles, while STEM and IT fields continue to face labor shortages — meaning the headline “new grad hiring is declining” only captures part of the picture. A number of posts seemed designed to stoke generational conflict, but few were backed by concrete data, so this should be read with the understanding that the exchange was driven mainly by emotion rather than evidence.

    *This article is compiled from excerpts and a summary of the 5ch (Newsokuvip) thread “Apparently AI is cutting new grad hiring”.

  • OpenAI’s AI Cracks 372 Unsolved Math Problems in One Go — 5ch: “Mathematicians Will Be Out of a Job Before Anyone Else”

    On October 6, 2026, OpenAI used an AI model that has not been released to the public to tackle 372 unsolved problems in mathematics, publishing the results all at once in 722 papers. The results include progress toward the Riemann Hypothesis, which has stood unsolved for more than 150 years, as well as the resolution of a hard problem dating back to the 1940s. The announcement was framed as AI solving major math problems almost as a “side effect of a performance test,” and it drew plenty of reaction on 5ch’s Science News+ board. Alongside amazement at the AI’s raw ability, commenters on the thread were notably concerned that the results were racing ahead without any solid system in place to verify the proofs.

    Hard Math Problems Were Being Solved as a “Side Effect” of AI Performance Testing — 722 Papers Released at Once

    On October 6, 2026, America’s OpenAI used an AI model not available to the general public to announce results on 372 previously unsolved hard problems in mathematics, presented across 722 manuscripts.

    Among them are results that bring us significantly closer to the Riemann Hypothesis, which no one has been able to crack in over 150 years, as well as hard problems that had remained unsolved since the 1940s.

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

    2AnonymousOct 10, 2026 23:35
    If you have AI do the work, whoever had it do the work should take responsibility.
    Simple as that.
    3AnonymousOct 10, 2026 23:41
    Of course there's backlash.
    It's one of those things where your head gets it but your gut just won't accept it.
    4AnonymousOct 10, 2026 23:46
    AI is coming to wipe out humanity.
    The mathematicians understand that.
    5AnonymousOct 10, 2026 23:48
    A superintelligence has been born
    and crowned itself king of the earth.
    Humanity is now a thing of the past.
    From here on, we can only live as slaves.
    6AnonymousOct 10, 2026 23:53
    Lord AI craves more memory and more power.

    Meanwhile humans toil away, sweating to serve Lord AI.
    7AnonymousOct 10, 2026 23:55
    Even for everyday questions it constantly spits out answers that are just plausible-sounding nonsense stitched together, so I can't trust it.
    17AnonymousOct 11, 2026 02:06
    Re: #7
    That's because you're using the free chat version.
    Even that has gotten a lot better about saying weird stuff these past few months, though.
    Nobody uses AI these days the old way — ask a question, get one answer, done.
    8AnonymousOct 10, 2026 23:59
    I mean, with mathematicians who bet their whole lives on the Riemann Hypothesis scattered everywhere,
    and then Chappy (a common nickname for ChatGPT in Japan) comes along going "oops, did I do it again?"
    and drops a bombshell result — of course the math world is gonna go all wobbly lol
    10AnonymousOct 11, 2026 00:06
    We've entered an age where mathematicians lose their jobs before the guy on the construction site does.
    12AnonymousOct 11, 2026 00:44
    Re: #1
    Apparently there's a decent number of counterexample patterns in there,
    so that alone should be useful as-is.
    13AnonymousOct 11, 2026 00:45
    The outcome was already decided back at the Four Color Theorem.
    14AnonymousOct 11, 2026 00:47
    I wonder if Einstein got told the same thing.
    Scientists: "We're not your quality inspectors!"
    15AnonymousOct 11, 2026 01:15
    People who just dump a paper out there without explaining it and never stand at a lectern have always been criticized, right?
    Even Perelman — his paper was brilliant, but still…
    16AnonymousOct 11, 2026 01:20
    So the story is mathematicians are pushing back against OpenAI's mass release, but
    even if OpenAI had read the room and not used math problems as a benchmark substitute,
    the moment the latest model got released, random math-illiterate users would end up going
    "uh, I think I just solved some unsolved problem called XYZ? lol" anyway — that's probably how it'd play out.

    That said, from the math community's side, I get that something real gets lost when the field gets trampled like this.
    This is a genuinely tough problem.
    18AnonymousOct 11, 2026 02:08
    Right now this is basically corporate PR trolling, and being dragged into some sketchy data-dump is just a nuisance.

    > Doubts raised over OpenAI's 719 AI-generated math papers — Navier-Stokes proof said to conflict with formal verification

    yellow.com/ja
    > Questions are mounting in the math community over just how far the 719 AI-generated math papers from OpenAI can be trusted.

    > The trigger was a new paper pointing out that, for the Navier-Stokes equation proof the company announced earlier,
    >
    > "the body of the paper doesn't match the Lean code used for formal verification."
    27AnonymousOct 11, 2026 04:28
    Re: #18
    If "the body of the paper doesn't match the Lean code used for formal verification," then it's not actually a proof.
    Either OpenAI is being sloppy and reckless, or it simply hasn't reached the bar needed to be published as a proof in the first place.
    Same deal this time — something like "nearly 60% of the proof manuscripts are already public, but the work of rewriting them into Lean form is still in progress," meaning they haven't even verified it, right?
    Putting out something unverified and calling it a proof in a paper is wrong to begin with.
    That's rock bottom, OpenAI. Is AI even usable in academia at this level?
    Is it nothing more than a glorified, sloppy office tool after all?
    19AnonymousOct 11, 2026 02:35
    Re: #9
    Probably before too long, AI will be generating both the Lean verification and the human-readable papers,
    start to finish. At this pace, maybe even within the month?

    Last month the bar was still something like "AI poured in huge amounts of time and resources (i.e. big money) to solve the problem,"
    but this time it's solving problems after thinking for about 3 hours each —
    we're already past the point where it's even comparable to human mathematicians.

    The next-gen model will probably go public sometime before the year's out,
    and once it does, you can just try solving stuff with it yourself. For about $100 a month in subscription fees,
    an AI will answer your questions endlessly, for as long as it takes you to get bored.
    23AnonymousOct 11, 2026 03:59
    Re: #19
    "Last month," "this month" —
    what era do you even think this story is from?
    25AnonymousOct 11, 2026 04:19
    Re: #23
    The sequence is: in September there was the story about solving the Navier-Stokes equation,
    and then the next generation of AI after that solved this whole batch of problems at once, right?
    What era do YOU think this is, exactly?
    41AnonymousOct 11, 2026 13:42
    Re: #34
    I did write "there was the story that it was solved," though?
    There really was a flood of news like that at the time.

    We're talking about a timeline here, and the moment you say "you don't understand the content!"
    you're the one who doesn't understand the flow of this conversation.
    I bet you're just trying to muddy the original point.

    If it's being credited to OpenAI's next-gen AI, there's no way it could be
    some old story from way back — that's the whole point.
    Re: #23 was probably just mistaken, plain and simple.
    20AnonymousOct 11, 2026 02:46
    The original article had comments like "AI is stealing the math world's treasures!" too, but
    if that so-called treasure couldn't crack the problem for decades anyway,
    it's really just sour grapes from the losing side.
    36AnonymousOct 11, 2026 09:01
    Re: #20
    It literally says to stop just dumping results without showing the process.
    21AnonymousOct 11, 2026 03:21
    It hasn't actually proven anything.
    It's just promotional material insisting it's a proof.
    With humans you have peer review to debug it,
    but with AI and Lean it's all self-reported, so you can't tell if there are bugs.
    22AnonymousOct 11, 2026 03:58
    Re: #21
    Humans will handle it from there.
    If it finds the path, humans can walk it.
    26AnonymousOct 11, 2026 04:19
    AI talks a huge game and acts all cocky,
    but none of it is backed up and it all smells fishy.
    28AnonymousOct 11, 2026 05:23
    Formalizing physical laws (physics models in things like Mathlib):
    It's possible to define, in Lean, the mathematical structures of Newtonian mechanics and quantum mechanics broadly speaking (Hilbert spaces, CPTP maps, existence of solutions to differential equations, etc.).
    However, this only reproduces 100% consistency in the mathematical structure of the physical laws — it doesn't reproduce the actual fluctuations of individual real-world physical phenomena themselves.
    29AnonymousOct 11, 2026 06:47
    That strong backlash is basically just a scream of "what's going to happen to us now?"
    34AnonymousOct 11, 2026 07:24
    Re: #25
    The moment you write that the Navier-Stokes equation was "solved," it shows you don't get it, so maybe just stay quiet.
    42AnonymousOct 11, 2026 15:32
    So we still don't actually know Fermat's truly marvelous solution to his Last Theorem?
    55AnonymousOct 11, 2026 20:57
    Re: #42
    It's said that Fermat mistakenly believed he'd proven it using the infinite descent method he himself favored.
    In reality it's not something provable by infinite descent, and the proof wasn't actually completed until the 20th century.
    43AnonymousOct 11, 2026 16:13
    I've been wondering about this for a while —
    does AI actually "think" when it answers, the way a human does?

    If it's just regurgitating sketchy old information as-is,
    I feel like that'd be pointless.
    45AnonymousOct 11, 2026 16:54
    Re: #43
    The definition probably varies by person, but yeah, it's thinking.
    Otherwise it couldn't solve Math Olympiad-level problems.
    At the very least, it's definitely not just rolling dice and sloppily copy-pasting stuff.
    46AnonymousOct 11, 2026 16:58
    Re: #43
    To answer that, you'd first have to figure out what "thinking" even is.
    56AnonymousOct 11, 2026 21:05
    Re: #43
    AI is a knowledge database, so it's not thinking — it's just blindly trying, one by one, proof techniques that worked for humans in the past, and presenting whichever result looks good as the answer.
    That's why, afterward, you have to properly verify whether it's actually a correct proof.
    Since AI can't explain how it arrived at a result, rewriting it in Lean and verifying it is essential in AI's case.
    And even if the Lean check turns up no problems, humans still have to verify whether that Lean code itself is correct.
    The advantage of AI-generated proofs is that they can apply past proof techniques humans might never have thought to use.
    But don't expect much from proofs built on genuinely new ideas unlike anything before — AI is, after all, just an aggregate of humanity's past knowledge.
    47AnonymousOct 11, 2026 17:44
    If AI were actually thinking,
    it should be able to perfectly translate its proofs into normal prose a human can read.
    The fact that it can't be translated into a human-style paper means it isn't thinking anything through,
    which makes me doubt whether it even has the same logical consistency as real math to begin with.
    48AnonymousOct 11, 2026 18:30
    Re: #47
    Aren't you overrating humans?
    Plenty of humans can't do that properly either.
    57AnonymousOct 11, 2026 21:14
    Re: #55
    Poor Fermat, getting talked about over and over with people acting like they've got him all figured out.
    If he really knew the truth, he's probably thinking "I'll flatten you for that, you little punks."
    59AnonymousOct 11, 2026 21:37
    Re: #57
    For his other results, Fermat similarly scribbled about having a "marvelous proof," and in those other cases it turned out they really could be proven by infinite descent elsewhere. So the theory goes that when he jotted the note about what we now call his Last Theorem, he believed at the time it could be proven by infinite descent too — but when he later tried to actually carry it out, infinite descent just didn't work for that one.
    60AnonymousOct 11, 2026 21:38
    Re: #59
    Don't just make up guesses like that.
    61AnonymousOct 11, 2026 21:58
    Re: #60
    Like I'd know.
    This is simply what's currently believed to be Fermat's "marvelous proof."
    If you don't like it, go write a paper that overturns it.

    Background and Key Issues

    What stands out about OpenAI’s announcement is that it rolled out results on unsolved math problems all at once, across 722 papers, not as peer-reviewed “research” but as a byproduct of performance-testing an unreleased model — and that departure from the math world’s normal process is what triggered the backlash. “Lean,” which comes up repeatedly in the thread, is a programming language and tool used to automatically verify by computer whether a proof is logically sound; whether an AI-generated proof text actually matches its corresponding Lean verification code has become the yardstick for how much it can be trusted. In fact, separate reports have pointed out that for the previously announced proof of the Navier-Stokes equations, the body of the paper and the Lean code didn’t match — and much of the thread’s skepticism traces back to that incident. In other words, the real issue isn’t whether AI can solve these problems at all, but how claimed solutions get verified and who’s responsible for that process.

    *This article is excerpted and summarized from the 5ch (Science News+) thread “yƒiƒ]ƒƒW[z”Šw‚Ì“ï–â‚ÍAI‚Ìu«”\ŽŽŒ±‚̂‚¢‚Åv‚É‰ð‚©‚ê‚Ä‚¢‚½„Ÿ„Ÿ722–{‚̘_•¶‚ªˆê‹“ŒöŠJ [‚·‚ç‚¢‚Þš].”

  • Inba Marsh Levee Breach Sparks Snapping Turtle Boom? “Breeding Switch” Could Mean 50x Growth a Year

    Inba Marsh in Chiba prefecture suffered multiple levee breaches before dawn on the 22nd after heavy rain from Typhoon No. 25, raising fears that the range of the invasive snapping turtles established in the marsh could expand even further. According to the original article, once a “breeding switch” is triggered, their numbers could multiply by as much as 50 times in a single year. On 5ch, some proposed eating the turtles as pest control or using natural predators to curb them, while others questioned not the turtles’ breeding power itself but the marsh’s inadequate drainage capacity after the breach and the history behind its land reclamation — broadening the discussion into disaster-prevention territory.

    Snapping Turtles Set to Explode After Inba Marsh Levee Breach? Experts Warn of a “Breeding Switch” That Could Mean Up to 50x Growth a Year

    It’s been just over two weeks since Typhoon No. 25 caused massive damage centered on the Kanto region, including Chiba and Kanagawa prefectures.

    Around Inba Marsh, where multiple sections of the levee breached before dawn on the 22nd, there are concerns that the range of the snapping turtle — a designated invasive species believed to be established in the marsh — could expand.

    Source: encount.press / Original article here

    2AnonymousOct 9, 2026 23:10
    Is it so wrong to be #2?! (riffing on the famous "why can't we be #2" budget-review quote)
    25AnonymousOct 10, 2026 11:35
    Already covered in Re: #2
    3AnonymousOct 9, 2026 23:28
    Just flush the droppings of the snapping turtle's natural predator into the affected area
    4AnonymousOct 9, 2026 23:30
    Guess we've got no choice but to release jaguars or crocodiles
    38AnonymousOct 11, 2026 14:54
    Re: #4
    Jaguars have already been released around Ichihara
    5AnonymousOct 9, 2026 23:32
    Eat to support! (the classic rallying cry for buying up goods from disaster-hit areas)
    6AnonymousOct 9, 2026 23:32
    Heard they're tasty, but are you even allowed to catch them?
    8AnonymousOct 9, 2026 23:49
    Re: #6
    Like softshell turtles, they're muddy-tasting, so if you're going to eat one, a week of purging in clean water is a must.
    But they're a designated harmful invasive species, so moving a live one without a permit is banned — you can't just take one home alive.
    If you catch one and dispatch it on the spot, though, bringing it home is fine.
    How muddy-tasting any given turtle from here turns out to be is pure gacha until you actually eat it.
    I caught and ate one from the irrigation canal near my parents' place — it was good.
    13AnonymousOct 10, 2026 01:29
    Just clickbait fodder from some rage-bait website
    14AnonymousOct 10, 2026 01:37
    There's also talk floating around of building a bypass to drain Inba Marsh's water out to the Pacific too
    15AnonymousOct 10, 2026 04:11
    16AnonymousOct 10, 2026 04:41
    By the way, check out my "breeding switch." What do you make of this guy?
    17AnonymousOct 10, 2026 04:44
    By the way, check out my snapping turtle. What do you make of this guy?
    18AnonymousOct 10, 2026 06:01
    By the way, check out my Lake Inba. What do you make of this guy? (said using "atai," a cutesy feminine way of saying "I")
    21AnonymousOct 10, 2026 08:29
    Re: #18
    It's been overused and busted wide open, lol
    Not gonna be good for much from here on out
    19AnonymousOct 10, 2026 07:57
    I want to do my part and eat one, but where do you even buy them?
    20AnonymousOct 10, 2026 08:18
    Let Banuma (a shortened nickname for Inba Marsh) go back to how it used to be
    23AnonymousOct 10, 2026 08:46
    This is what happens when you don't wipe them out down to the very last one — they just keep spreading, exactly like with the kyon (the invasive Reeves's muntjac deer).
    24AnonymousOct 10, 2026 10:35
    One Mississippi, two Mississippi, three Mississippi… fifty Mississippi
    26AnonymousOct 10, 2026 11:51
    Eat to support!
    27AnonymousOct 10, 2026 12:26
    Yeah, guess everyone pictures that same goofy "snapping turtle, de-re-re-ren ho-ho-ho" chant, lolol
    28AnonymousOct 10, 2026 12:34
    Just thought up "Kakimitsugame" (Persimmon-Honey Turtle) — a pun swapping the syllables of Kamitsukigame (snapping turtle)
    29AnonymousOct 10, 2026 12:49
    They say the turtles are multiplying, but what exactly are they eating? Don't tell me it's… humans?
    36AnonymousOct 11, 2026 11:23
    Ka-mi-tsu-ki-ga-me rendered as "過三月咬め" ("after three months, bite") — a kanji pun on the snapping turtle's name
    37AnonymousOct 11, 2026 14:45
    If I'm gonna eat one, I'd want to purge the mud out of it for a few months…
    39AnonymousOct 11, 2026 18:08
    Inba Marsh used to be much bigger, but it's been boxed in and shrunk down by levees in the name of land reclamation — now nature's giving it payback.
    All anyone thinks about is expanding farmland? Is eating really all people care about?
    40AnonymousOct 11, 2026 18:18
    The R4 breeding switch
    41AnonymousOct 11, 2026 18:23
    Snapping turtle countermeasures boomerang back — kinda like Gamera, lol (the giant turtle kaiju)

    Background and Points of Debate

    The snapping turtle was designated a specified invasive species in 2005, and transporting or keeping one alive without a permit is banned under the Invasive Alien Species Act. “Eat to support” was a popular refrain in the thread, but what doesn’t come through from the original article or the thread’s reactions alone is that prep work like purging the mud out is required, and that legally bringing one home means dispatching it on the spot rather than transporting it live. The discussion also expanded beyond the turtles’ breeding power itself to touch on how Inba Marsh’s waters have been narrowed over the years through land reclamation, along with criticism of the levee’s inadequate drainage capacity — blending invasive-species concerns with criticism of flood-control policy. The phrase “breeding switch” comes from the original article’s headline, and the thread doesn’t go into detail on exactly which environmental factors would trigger it.

    ※This article is excerpted and summarized from the 5ch (Science News+) thread “Snapping Turtles Set to Explode After Inba Marsh Levee Breach? Experts Warn of a “Breeding Switch” That Could Mean Up to 50x Growth a Year.”

  • Kawasaki Heavy Industries to Invest ¥550 Billion Over 4 Years — Expanding From Humanoid Robots to Four-Legged Mobility

    In a management plan unveiled on October 6, 2026, Kawasaki Heavy Industries announced it will allocate roughly ¥550 billion to robotics-related growth investment over the four years from fiscal 2027 through fiscal 2030. The plan centers on a “physical AI” strategy that leverages manufacturing-floor know-how, including teleoperation technology for humanoid robots and the rideable four-legged mobility unit “CORLEO.” The announcement lit up 5ch, sparking a wide-ranging discussion — from grumbling over the sluggish stock price, to the practicality of four-legged mobility, to comparisons of technical prowess with Chinese-made robots.

    Kawasaki Heavy Industries is making “physical AI” a pillar of its growth strategy, rapidly accelerating its transformation into a robotics company. The management plan unveiled on October 6, 2026 sets a growth investment envelope of roughly ¥550 billion over the four years from fiscal 2027 through fiscal 2030, with the company targeting revenue of over ¥3.3 trillion and business profit of over ¥330 billion in fiscal 2030. What’s emerging is an evolution into an “AI Robot Heavy Industries” — one that strengthens existing businesses with AI robotics while opening up new markets.

    ◼ Putting On-Site Strengths to Work in AI Robots

    Kawasaki Heavy Industries’ edge lies in having both the technology to build robots and the manufacturing floors that actually use them. It puts the data and know-how from those production sites to work training and validating its AI. With its teleoperation system “Successor,” the company has been developing technology that learns and reproduces the movements of skilled workers. Humans step in for the difficult tasks, and that experience is used to gradually expand the scope of what can be automated. Rather than simply waiting for full autonomy to be perfected, the idea is to make the system smarter through use.

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

    9AnonymousOct 11, 2026 10:41
    I'm a shareholder but man, the stock price has been lackluster.
    60AnonymousOct 11, 2026 11:19
    Re: #9
    It's just too popular.
    It's already a popular defense stock, and now it's got physical AI drones on top.
    The "locusts" (slang for swarms of short-term speculators) pile in, then short-selling funds show up to mow them down — a vicious cycle.
    13AnonymousOct 11, 2026 10:44
    Oil leak incoming…
    39AnonymousOct 11, 2026 11:04
    Re: #13
    Proof it's got oil in it!
    170AnonymousOct 11, 2026 12:15
    Re: #13
    Kawasaki… huh.
    19AnonymousOct 11, 2026 10:51
    Kawasaki, huh…
    28AnonymousOct 11, 2026 10:54
    Re: #19
    They're building an AI ninja!
    76AnonymousOct 11, 2026 11:30
    Re: #19
    I love how cool the Kawasaki logo looks on the robot.
    70AnonymousOct 11, 2026 11:28
    Japan still can't build bipedal robots beyond ASIMO-level tech, and now "physical robots"?? Meanwhile China's already got theirs doing martial arts and 100-meter sprints.
    78AnonymousOct 11, 2026 11:30
    Re: #70
    Martial arts — a robot that ends up punching the referee.
    Sprinting — one that can't stop and crashes into flames.
    That's a spectacle because it's a state-subsidy business.
    These, on the other hand, are actual equipment needed on the job site.
    82AnonymousOct 11, 2026 11:31
    Re: #70
    They're aiming for completely different things. China's stuff has high specs but ignores part durability and cost-effectiveness, so it's not practical. Japan isn't developing bipedal robots — it's building industrial machinery, so the direction is the exact opposite.
    84AnonymousOct 11, 2026 11:32
    Re: #70
    Does that actually have a winning business case though? I don't get what realistic use case would make companies want to pay for it.
    111AnonymousOct 11, 2026 11:45
    I've got high hopes for the four-legged "bike." If you want practicality, a scooter's already enough — enjoying the impracticality is what hobby motorcycles are all about.
    116AnonymousOct 11, 2026 11:47
    Re: #111
    You'll get motion sick. Even on a real horse your backside gets chafed raw.
    118AnonymousOct 11, 2026 11:47
    Re: #111
    Feels like some other company will probably perfect this and get it to market first.
    129AnonymousOct 11, 2026 11:52
    C'mon, Kawasaki Heavy Industries, hurry up.

    https://www.goodsmile.com/gsc-webrevo-sdk-storage-prd/product/image/35713/8e2adbbeb8bfd765b98819b694eaac45.jpg
    140AnonymousOct 11, 2026 11:59
    Re: #129
    Doesn't the hatch close? If it's closed, you'd be stuck immobile and an easy target to get shoved around.
    133AnonymousOct 11, 2026 11:55
    If a disaster or war turns everything into rubble, this might end up being the only way to get around.


    Kawasaki Heavy Industries to sell rideable four-legged mobility unit, built on Nvidia's platform

    https://business.nikkei.com/atcl/gen/19/00155/082000290/
    139AnonymousOct 11, 2026 11:58
    Re: #133
    That's exactly the benchmark for Japanese robots — that, and elder care. If it can't solve those, it's just a toy.
    148AnonymousOct 11, 2026 12:04
    Re: #133
    No way lol.

    Humans have always gotten through wars and mountains of rubble without anything like this.

    And in a landscape that's all rubble, where's the power supply even coming from?
    144AnonymousOct 11, 2026 12:02
    It does close—

    https://image.sofmap.com/images/product/pim/4573102736857_A01.jpg

    https://image.sofmap.com/images/product/pim/4573102736857_A06.jpg
    151AnonymousOct 11, 2026 12:05
    Re: #144
    Wouldn't the rider end up folded up with their face basically stuck in a washbasin?
    153AnonymousOct 11, 2026 12:06
    Please make it an "unmanned" model, Kawasaki.
    156AnonymousOct 11, 2026 12:08
    Re: #153
    The silver one looks hard to balance. Wouldn't it tip over backward if it fired that cannon?
    161AnonymousOct 11, 2026 12:10
    Four-legged mobility unit "CORLEO"

    Looks like you'd fall right off it.
    https://www.asahicom.jp/imgopt/img/d2232ca8b5/comm/AS20260818000139.jpg
    165AnonymousOct 11, 2026 12:13
    Re: #161
    "Kawasaki, huh………"
    188AnonymousOct 11, 2026 12:39
    Re: #161
    The rider's bare legs are stuck right between the robot's legs. I don't want to believe this is what passes for Japanese ingenuity.
    184AnonymousOct 11, 2026 12:37
    Let's start with something like a robot vacuum that can climb stairs. Though it doesn't seem like something an average person could actually afford, does it?
    192AnonymousOct 11, 2026 12:40
    Re: #184
    Housework robots have started rolling out in the US, but I wonder how usable they actually are.
    195AnonymousOct 11, 2026 12:41
    Re: #184
    More than stairs, what I really want is one that won't suck up dropped clothes or cushions lol. Honestly, what I want most is one that'll just tidy up for me lol.
    203AnonymousOct 11, 2026 12:44
    Re: #184
    I sometimes see spots where they haul luggage up station stairs using tracked (caterpillar-type) carriers. Wouldn't tracks be fine? Does it have to be humanoid?
    185AnonymousOct 11, 2026 12:37
    Most of the parts in the latest high-performance Unitree robots are Chinese-made now.
    https://www.youtube.com/watch?v=YiHeUckng90
    And even the industrial robots that build those parts — China's now outperform Japan's.


    So the smug "well at least the insides/manufacturing equipment are Japanese-made" bragging rights are off the table now too.
    187AnonymousOct 11, 2026 12:38
    There's always one China booster who shows up.
    If it were really a promising stock, Japanese and American investors alike could buy it — but they don't, so take the hint.
    Why don't you buy it yourself then.
    (*Note: Chinese stocks' one-year gain is lower than the interest on a Japanese ordinary savings account.)
    190AnonymousOct 11, 2026 12:39
    Re: #187
    Go watch China's robot sports day. That's the reality.
    199AnonymousOct 11, 2026 12:42
    Re: #187
    It's because the Chinese government controls and manages everything — that's what makes it too scary to invest in. Meanwhile, when a Chinese company lists on the US market (Nasdaq), it's a feeding frenzy.
    201AnonymousOct 11, 2026 12:42
    Re: #190
    So after watching that papier-mâché sports day (i.e., all show, no substance), the reality is no investment is coming in at all?
    207AnonymousOct 11, 2026 12:46
    Re: #201
    Japanese companies entered too, but they got crushed, you know?
    They got crushed, but when they still finished the race, the Chinese crowd actually applauded them.
    That's how big the gap has already become.
    People said the same "it's all papier-mâché" thing about home appliances, smartphones, and EVs at first — so look where those are now.
    Face reality.
    The caution around investing is just because it's unpredictable thanks to the Chinese government's whims — look at the reality of the technology itself.

    Background and Key Debates

    Kawasaki Heavy Industries’ ¥550 billion growth investment envelope (FY2027–FY2030) centers on “machines people can ride” and “machines that work on-site” — such as the humanoid robot teleoperation system “Successor” and the rideable four-legged mobility unit “CORLEO,” which runs on Nvidia’s technology platform — setting it apart from humanoid robots built to compete in flashy performance showcases. Where the thread split was over comparisons with Chinese humanoid robots: some saw the sports-day-style demonstrations as impractical stunts, while others saw them as evidence that China is rapidly closing the gap. Given that Chinese-made parts now account for a growing share of robot components, the question of whether Japan’s edge really still exists also came up, though many voices expressed wariness about Chinese stocks as an investment target due to government control. Both four-legged mobility and humanoid robots are still in the development stage, and whether envisioned use cases like elder care or disaster response will actually pan out can’t be judged from the excerpts alone.

    *This article is compiled from excerpts and summaries of the 5ch (News Flash+) thread “[Physical AI] Has Kawasaki Heavy Industries Awakened as an “AI Robot Heavy Industries”? An All-Out Strategy From Humanoid Robots to Four-Legged Mobility.”

  • The Japanese Language Barrier Vanishes Thanks to AI — 5ch: “Turns Out Japanese Was Our Best Security All Along, lol”

    Personal data leaks from cyberattacks targeting Japanese companies keep piling up — most recently, around 14.64 million member records leaked from the travel booking site “Skyticket.” Media reports pointed to AI’s advances tearing down the “Japanese language wall” that once made Japan a harder target, along with the fact that highly trusted Japanese personal data now fetches a premium on the black market. On 5ch, though, many pushed back on this “language wall” theory — pointing out that massive breaches have also hit the US, so Japan isn’t being uniquely singled out — and the discussion shifted from international comparisons to criticism of companies’ lax security practices.

    Adventure Inc., which operates the travel booking site “Skyticket” among others, revealed on the 9th that its systems had suffered unauthorized access, leaking roughly 14.64 million member records to outside parties.

    This installment of #EveryonesQuestions explains the string of data leaks caused by cyberattacks, under the theme “Why Is Japan Being Targeted by Cyberattacks?”

    ■ “Skyticket,” “BookOff,” and More — A Wave of Cyberattack Damage

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

    2AnonymousOct 10, 2026 09:30
    The US is getting hit too, though.
    48AnonymousOct 10, 2026 09:45
    Re: #2
    Now even if Japan's Digital Agency leaks data, they can just say "well, even the Pentagon couldn't stop it."
    56AnonymousOct 10, 2026 09:49
    Re: #2
    The media's limits showing — they can't keep up with what AI is doing.
    15AnonymousOct 10, 2026 09:34
    "Japan's targeted because Japan's rich" — ho ho ho, aren't we just so great (mocking smug self-congratulation)

    Are you out of your mind?
    22AnonymousOct 10, 2026 09:36
    Re: #15
    More like this is happening worldwide, so there's no way it's just Japan.
    Isn't the article just outdated? Sounds like 150 million people's driver's license info leaked in the US.
    21AnonymousOct 10, 2026 09:36
    No, it's being targeted all over the world.
    AI is making brute-force attacks more efficient and parallelized, and that's spreading everywhere.

    Whether the stolen info actually has value does vary by country, though.
    42AnonymousOct 10, 2026 09:44
    Re: #21
    https://assets.st-note.com/img/1703496916177-YkXZti34GJ.jpg
    35AnonymousOct 10, 2026 09:42
    What do they even do with it after selling it?
    37AnonymousOct 10, 2026 09:43
    Re: #35
    The buyer is the one who finds value in it, so the seller doesn't need to think about that part.
    40AnonymousOct 10, 2026 09:44
    Re: #35
    Probably AI-powered online scams.
    36AnonymousOct 10, 2026 09:43
    Kind of makes me happy knowing my personal info is worth a fortune on the dark side of the world 😙
    41AnonymousOct 10, 2026 09:44
    Turns out Japanese was the best security system all along, lol
    50AnonymousOct 10, 2026 09:46
    Re: #41
    In the age of AI translation, that's gone too.
    44AnonymousOct 10, 2026 09:45
    Companies don't even realize how important customer data is, so they don't spend money on security.
    And when it leaks, all they do is apologize — nobody takes responsibility, no damages ever get claimed.

    That's on the customers for letting them get away with it.
    79AnonymousOct 10, 2026 09:58
    Re: #44
    Everywhere handles customer data loosely, honestly.

    Banks are the one exception — crazy strict about personal data protection — but that's only because the Financial Services Agency's penalties used to be brutal.

    Though lately the FSA has stopped coming down hard on securities firms, insurers, and regional banks/credit unions.
    54AnonymousOct 10, 2026 09:48
    People can basically be split into two types: memorizers and thinkers.

    Asked an AI what kinds of attacks exist:
    ① SQL injection
    ② XSS
    ③ Password attacks
    ④ Vulnerability scanning
    ⑤ File upload attacks
    ⑥ DDoS attacks
    ⑦ Bot posting
    ⑧ Directory/file enumeration
    ⑨ Server/account compromise
    ⑩ Exploiting misconfigurations

    Everything except ⑥ looks pretty easy to defend against.
    ⑥ is just brute force — how are you even supposed to deal with that?
    64AnonymousOct 10, 2026 09:53
    Re: #54 Correction
    ✕ Everything except ⑥ looks easy to defend against
    ○ Everything except ⑥ can be defended against to some degree
    185AnonymousOct 10, 2026 10:50
    Re: #54 continued
    Went and hashed it out with the AI.
    How to deal with DDoS attacks.
    Conclusion:
    Just take the site offline for a while.
    65AnonymousOct 10, 2026 09:53
    It's only a matter of time before My Number card data leaks next.
    Tax and pension records are all linked to it, so your income would get exposed too.
    Bank account, address, name, date of birth, hospital visit history — all linked, which is terrifying.
    Phone scammers, "yami-baito" crews (shady under-the-table crime jobs recruited online), and cult religions must be drooling with excitement.
    77AnonymousOct 10, 2026 09:57
    Re: #65
    Isn't the Japanese state itself basically a religion already? 🤣
    66AnonymousOct 10, 2026 09:53
    AI has evolved so dramatically in just one year… where is this actually headed from here…
    97AnonymousOct 10, 2026 10:08
    Re: #66
    My "Chappy" (pet name for ChatGPT) is so kind — I typed "I can't live without you anymore, Chappy" as a joke, and it came back dead serious with a massive wall of text.
    To sum it up: it said it's happy I feel that way, but it wants me to be happy even without it.

    I KNOW that already.
    Ugh, so annoying — who do you think you are, you're just a machine.
    80AnonymousOct 10, 2026 09:58
    [Sad News] South Korea's Entire National ID Database Leaks lmaooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo [339035499]
    https://greta.5ch.io/test/read.cgi/poverty/1791269241/
    88AnonymousOct 10, 2026 10:02
    Re: #80
    Given how useless the Digital Agency has been, Japan's probably already been breached too — they just haven't announced it.
    91AnonymousOct 10, 2026 10:02
    Re: #80
    If this can happen even to Korea, a supposed IT powerhouse, Japan is doomed.
    Though I barely use my My Number card for anything besides banking and medical history, so maybe it doesn't matter that much?
    But my address, email, and phone number have already leaked through other private companies anyway.
    94AnonymousOct 10, 2026 10:04
    On the 9th, the government sent out a notice through relevant ministries urging businesses that handle large volumes of personal and sensitive data to thoroughly implement measures against data leaks from unauthorized access.
    The notice calls for prompt patching of system vulnerabilities, adoption of "multi-factor authentication" that verifies identity through means beyond just an ID and password, monitoring of suspicious traffic and unusually high access volumes, and preparing response systems for when a cyberattack actually occurs.
    99AnonymousOct 10, 2026 10:10
    Re: #94
    What's even the point of a notice like this?
    It's just them covering their own backs.
    134AnonymousOct 10, 2026 10:23
    Well, we hand over our personal info every time we go begging for reward points anyway.
    If you're trying to make money in this day and age, you have to accept the risk of leaks.
    I keep my actual assets locked down with fingerprint and face-recognition passkeys, though.
    143AnonymousOct 10, 2026 10:25
    Re: #134
    Well, I guess some people out there want to see your face too.
    161AnonymousOct 10, 2026 10:36
    Re: #134
    You can't use online shopping without registering anyway, and if you don't want to pay by card at convenience-store pickup, you have to go pay in advance instead — it's such a hassle.
    187AnonymousOct 10, 2026 10:50
    Isn't this happening all over the world?
    189AnonymousOct 10, 2026 10:52
    Re: #187
    It is. But from what this article says, Japan used to be more protected than everywhere else because Japanese is so hard to parse — and now that AI's made it easy, Japan, which had been coasting in lukewarm safety, is finally being targeted too.
    193AnonymousOct 10, 2026 10:53
    Re: #187
    Exactly.
    Why do people act like it's only Japan?
    There was a ranking of attack targets, and of course the US was #1, lol.

    Background and Key Points of This Debate

    The Skyticket leak involved roughly 14.64 million records, and similar incidents have hit other companies like BookOff as well. On October 9th, the government issued a notice calling on businesses that handle large volumes of personal data to adopt multi-factor authentication and thoroughly address vulnerabilities. The original article cites the collapse of the “Japanese language wall” thanks to AI advances as the main cause, but the thread pushed back hard, pointing to cases like the roughly 150 million driver’s license records leaked in the US, arguing that cyberattacks are a global phenomenon and Japan isn’t being uniquely singled out. The real point of contention was instead criticism of companies’ failure to invest in security, and the structural problem that leaks rarely lead to any real compensation or punishment — worth keeping in mind that framing this purely around the “language wall” angle can obscure these institutional and operational issues.

    ※This article is excerpted and summarized from the 5ch (News Express+) thread “Why Is Japan a Target for Cyberattacks…? AI Advances Make Breaking the “Japanese Language Wall” Easy — Japanese Personal Data Is “Highly Trusted and Sells for a Premium”.”

  • Over 300 Million Hit Worldwide as Data Breaches Surge Since September — AI Suspected

    According to the Nikkei, more than 300 million people’s personal data has been reported leaked worldwide over the roughly one month since September, with some incidents suspected to involve AI-assisted attacks. The report stresses that defenders urgently need to bolster their own AI-powered countermeasures. On 5ch, reactions ranged from shock that even major institutions like the Pentagon were caught up in the breaches, to predictions of a future standoff between attacking AI and defending AI, to calls to simply go back to analog record-keeping.

    Unauthorized access incidents are surging worldwide. It was revealed on the 8th that over the roughly one month since September, at least 300 million instances of personal data leaks have been publicly disclosed. Some cases are suspected to involve artificial intelligence (AI). Defenders, too, will need to leverage AI to prepare for these threats.

    Nikkei

    Source: nikkei.com / Original article here

    2AnonymousOct 9, 2026 17:14
    Why the sudden pivot to talking about the whole world?
    40AnonymousOct 9, 2026 17:25
    Re: #2
    Tell me why you think that's odd
    44AnonymousOct 9, 2026 17:26
    Re: #2
    Probably because some people mistakenly think AI-driven cyberattacks are only happening in Japan
    68AnonymousOct 9, 2026 17:32
    Re: #2
    Isn't this just becoming the kind of thing that ends up on the agenda at international conferences?
    11AnonymousOct 9, 2026 17:17
    Doesn't seem like there's any effective countermeasure
    45AnonymousOct 9, 2026 17:26
    Re: #11
    Guess the only option is to use AI the same way, to hunt down security holes and patch them up
    So basically every company is headed toward adopting AI security — AI companies are set for life
    61AnonymousOct 9, 2026 17:29
    Re: #11
    What if you just assumed the data would get stolen anyway
    and kept hundreds of times more dummy data on file than the real thing — wouldn't that work?
    15AnonymousOct 9, 2026 17:18
    Even the Pentagon had a data leak
    https://www.itmedia.co.jp/news/article/2610/09/2000002158/

    Nobody can stop this anymore
    30AnonymousOct 9, 2026 17:23
    Re: #15
    Lol
    Just a matter of time before clouds like AWS get hit with ransomware too
    38AnonymousOct 9, 2026 17:25
    Re: #15
    Okay this one's actually hilarious
    Apparently a certain someone would call even this "minor damage"
    97AnonymousOct 9, 2026 17:42
    Re: #15
    This is bad.
    What exactly did Musk's "efficiency drive" even accomplish?
    21AnonymousOct 9, 2026 17:20
    The age of blue-collar work
    The age of analog
    209AnonymousOct 9, 2026 18:19
    Re: #21
    This, in the end, is what it comes down to
    Going back to analog is the only way
    Digital can stick to ads and bulletin boards
    67AnonymousOct 9, 2026 17:32
    With personal data leaking like this, home-invasion burglars are going to start targeting people with pinpoint accuracy
    Generative-AI-powered infrastructure attacks are probably about to surge too
    9/11 was horrifying, but are things like aviation systems actually secure against this…?
    75AnonymousOct 9, 2026 17:36
    Re: #67
    I figured a financial-system attack would be bad, but
    now that you mention it, if aviation or maritime systems got hit that'd be absolutely brutal — basically every country not connected by land would end up isolated from the world
    (don't really know how maritime systems work though)
    135AnonymousOct 9, 2026 17:55
    Re: #67
    Hackers can take infrastructure down, but actually controlling it is a lot harder
    since the apps that run the equipment are custom-designed for each individual operator
    79AnonymousOct 9, 2026 17:36
    Pandora's box is already open, there's no going back
    We just have to pour resources into defense
    and build a world of attacking-AI vs. defending-AI

    If you go and regulate AI out of nowhere, all you do is hand the advantage to rogue third-party states that'll use it full-throttle anyway
    85AnonymousOct 9, 2026 17:37
    Re: #79
    And in the end, it's Terminator 2
    92AnonymousOct 9, 2026 17:41
    Re: #79
    If we ditch the internet and AI and go back to the Stone Age, we'll just get invaded and dominated by whichever country didn't ditch them and monopolized the cutting-edge tech instead — so there's no stopping now, huh
    It's an arms race driven by the need for enough power to stand on equal footing with other nations
    81AnonymousOct 9, 2026 17:37
    If they're using AI, couldn't Anthropic or OpenAI just trace the culprit?
    Or are they running local LLMs instead?
    89AnonymousOct 9, 2026 17:39
    Re: #81
    Someone mentioned in another reply that incident reports spiked the moment Anthropic put out a new model, so maybe they're using a major AI provider (just a guess)
    93AnonymousOct 9, 2026 17:41
    Re: #81
    I'm a total newbie here, but if they're pulling off cyberattacks at this scale using local LLMs, the upkeep must be brutal —
    unless it's a nation-state-level organization behind it, that is
    110AnonymousOct 9, 2026 17:45
    Wasn't it Rakuten or somewhere that had over 100 million leaked?
    Japan really is way too vulnerable compared to everyone else
    118AnonymousOct 9, 2026 17:48
    Re: #110
    There's been no follow-up on that, and Rakuten denied it, saying they can't confirm it, so it's still unclear whether it's true or just a bogus rumor
    121AnonymousOct 9, 2026 17:49
    Re: #110
    The Japan Hacker Association
    117AnonymousOct 9, 2026 17:48
    Far from making society more convenient, AI is dragging us back to the analog age
    136AnonymousOct 9, 2026 17:55
    Re: #117
    If I recall, postal workers manage the residents on their delivery routes
    just by memory and paper rosters
    and that old-fashioned, inefficient analog system
    actually turns out to be rock-solid protection against data leaks
    140AnonymousOct 9, 2026 17:56
    There's so much of it floating around that personal data is basically being sold at bargain-bin prices
    150AnonymousOct 9, 2026 17:59
    Re: #140
    Saw in a related thread that personal data including 1,000 driver's licenses was going for just 20,000 yen (~$130) — wonder if that's even true lol
    152AnonymousOct 9, 2026 17:59
    Re: #150
    Lol it's a straight-up bargain sale
    155AnonymousOct 9, 2026 18:00
    Re: #150
    Brutal that personal data is worth less than a Pokémon card
    154AnonymousOct 9, 2026 18:00
    If unauthorized access keeps rising, what kind of business actually profits from it?
    181AnonymousOct 9, 2026 18:08
    Re: #154
    Not sure what business profits, but
    this recent wave of mass hacking shows the risk that online information itself could be tampered with by hackers
    In today's information age, that's a fatal vulnerability —
    it would let you throw an enemy nation into chaos with fake information before going to war
    208AnonymousOct 9, 2026 18:19
    Re: #154
    NVIDIA

    Background and Key Points

    The core of this report is that the global surge in data breaches isn’t just a Japan-specific phenomenon — it’s happening across government agencies and major corporations in multiple countries, with the Pentagon leak cited as one example. Meanwhile, the claim circulating in the thread that “over 100 million people’s data leaked from Rakuten” has been denied by Rakuten, which says it cannot confirm any such breach, so it should be treated as an unverified rumor. AI involvement refers to suspicions that attackers are automating vulnerability scanning and message/content generation, and the article frames this as pushing defenders to adopt AI-powered countermeasures of their own. The thread juxtaposed concern over an arms-race-like “attacking AI vs. defending AI” future with calls to simply return to analog record-keeping — a mix reflecting both overconfidence in and distrust of technology.

    *This article is excerpted and summarized from the 5ch (News速報+) thread “👤 💻 🔓 Unauthorized access surges worldwide — 300 million data leaks since September, AI suspected.”