Category: Tech

  • Did AI Just Prove the Riemann Hypothesis? 5ch Users Baffled by a ‘250,000-Page Proof’

    A mathematician at the US company Anthropic has announced that the firm’s generative AI found a proof of the Riemann hypothesis, unsolved for 167 years. The proof reportedly runs to over 250,000 pages, with the code used for formal verification totaling 17.4 million lines — a scale that left 5ch users both skeptical (“can a human even peer-review this?”) and worried about the fallout for prime-number-based encryption. Peer review isn’t finished yet, and the thread was further stirred by news that, around the same time, a separate mathematician independently claimed to have proven the generalized Riemann hypothesis. Reactions split between cautious wait-and-see over verification and genuine excitement.

    The Riemann Hypothesis, Unsolved for 167 Years, Possibly Proven by Generative AI… Anthropic Mathematician Announces

    A mathematician at US AI developer Anthropic announced on his personal website on the 20th that, using the company’s generative AI, he had found a proof of the notoriously hard “Riemann hypothesis.”

    If correct, it would resolve a problem mathematicians have grappled with for 167 years.

    Source: yomiuri.co.jp / Original article here

    2AnonymousNov 21, 2026 09:12
    If this turns out to be true, passwords are going to be trivially crackable
    3AnonymousNov 21, 2026 09:19
    Most mathematicians already believed the Riemann hypothesis was true — the terrifying part was always whether it could actually be proven.
    Some people were saying existing methods just couldn't crack it and that some genuinely new mathematics would be needed… so how big a conceptual leap is this, really?
    Guess we'll just have to wait for a natural-language writeup of the proof.
    4AnonymousNov 21, 2026 09:20
    The kind of 'Riemann Shock' where stock prices go UP (a pun on the Japanese term for the 2008 'Lehman Shock,' which sounds like 'Riemann')
    5AnonymousNov 21, 2026 09:23
    WOOOOOAH!!
    6AnonymousNov 21, 2026 09:28
    Formally it's a fractional series, but there are complex exponents thrown in too
    7AnonymousNov 21, 2026 09:33
    And right on the heels of this news, another mathematician, Alessandro Masullo, has independently claimed to have 'proven the generalized Riemann hypothesis without any help from AI.'
    Neither one has been peer-reviewed yet, but it sounds like things are about to get hectic on the Riemann hypothesis front for the next few days.
    9AnonymousNov 21, 2026 09:39
    Re: #7
    It'd be funny if he just had the AI take a crack at it, went 'oh, THAT'S the trick!,' and then finished the proof by hand himself.
    Though I guess if you just get inspired from reading someone else's paper, that doesn't really count as 'getting help.'
    8AnonymousNov 21, 2026 09:33
    With AI performance climbing this fast, there's going to be a lot less work left for the theoretical fields.
    Even in applied science, more and more drug discovery has AI designing the proteins and peptides.

    The hard part now is probably getting physical-world AI to operate using the vocabulary of someone with, say, a 40 hensachi (a Japanese standardized test score where 50 is average — so, well below-average verbal ability).
    10AnonymousNov 21, 2026 09:40
    Next, have the AI prove Fermat's Last Theorem restricted to the math available in the 17th century, when Fermat himself was alive — no modern mathematics allowed.
    Seems like that one's beyond what the human brain can manage.
    11AnonymousNov 21, 2026 09:42
    So the Anthropic guy just 'believes the AI proved it,' huh?
    It's always like this.
    12AnonymousNov 21, 2026 09:42
    Amazing achievement, but there's a lot of hassle too — credit cards stop working, encryption gets invalidated, that sort of thing.
    16AnonymousNov 21, 2026 10:02
    Re: #12
    Nah, even if this gets proven, it's not like prime-number encryption suddenly becomes invalid.
    22AnonymousNov 21, 2026 10:17
    Re: #12
    Apparently encryption has nothing to do with it.
    13AnonymousNov 21, 2026 09:49
    These announcements are wrong basically every single time, so this one's probably wrong too.
    14AnonymousNov 21, 2026 09:49
    Hard to imagine encryption becoming unusable overnight — people have already been building attacks assuming the hypothesis is true even before any proof existed.
    15AnonymousNov 21, 2026 09:50
    The one kind of Riemann hypothesis 'proof' that could never be wrong is a counterexample — you could verify it instantly the moment it's published.
    Now THAT, even an AI could pull off for sure.
    17AnonymousNov 21, 2026 10:02
    So basically, building on the mountain of existing papers, the proof turns out to have been within reach.
    18AnonymousNov 21, 2026 10:04
    Well, that's it, humanity is doomed.
    19AnonymousNov 21, 2026 10:08
    https://www.nytimes.com/2026/11/20/science/anthropic-ai-math-riemann-hypothesis.html
    The New York Times has a story on it too
    20AnonymousNov 21, 2026 10:09
    It's over the moment humans can no longer understand AI's proofs.
    23AnonymousNov 21, 2026 10:49
    Re: #20
    First go prove the ABC conjecture, lol
    21AnonymousNov 21, 2026 10:14
    Apparently the Lean verification code runs 17.4 million lines — can a human actually check that?
    24AnonymousNov 21, 2026 11:06
    The same old PR playbook: make a big splashy announcement that it's 'solved,' then just dump all the verification work on academia. That's exactly the kind of thing that makes serious scholars furious.

    And the verification tools themselves can't even guarantee they're bug-free, yet they're just being used as a benchmark anyway.
    27AnonymousNov 21, 2026 11:29
    Re: #24
    Apparently it's already passed multiple verification tools. No mathematicians are actually complaining about it.
    25AnonymousNov 21, 2026 11:17
    If you had AI tackle the task of constructing a hypothesis/theory that explains experimental observations without contradiction, would theoretical physics start advancing too, I wonder?
    26AnonymousNov 21, 2026 11:25
    Give us something on the level of E=mc², c'mon.
    28AnonymousNov 21, 2026 11:48
    No idea how many pages the paper is, but first go find a mathematician capable of peer-reviewing it and get an estimate of how many years that'll take.

    Let's talk once peer review is actually done.

    …Or are they gonna have AI do the peer review too?
    30AnonymousNov 21, 2026 11:58
    Re: #28
    According to the source it's apparently over 250,000 pages. And the verification already uses AI.
    29AnonymousNov 21, 2026 11:54
    There are problems the human brain can't solve, so it wouldn't be strange if there were proofs the human brain can't peer-review either.
    31AnonymousNov 21, 2026 12:14
    Guess we'll need an AI to verify the AI that verifies the AI… lol
    32AnonymousNov 21, 2026 12:14
    Humans are already becoming obsolete.
    33AnonymousNov 21, 2026 12:23
    > 250,000 pages
    34AnonymousNov 21, 2026 12:24
    Re: #30
    > over 250,000 pages
    That is incredibly sketchy ^^;
    35AnonymousNov 21, 2026 12:47
    I guess there are propositions that just can't be proven without hundreds of thousands of pages.
    Exploring that corner of the mathematical world is simply beyond the human brain.
    Thanks for all your hard work over the years, mathematicians.
    37AnonymousNov 21, 2026 12:54
    Super-Hard 'Riemann Hypothesis' Possibly Proven With AI — US Anthropic Mathematician Announces
    November 21, 2026
    https://www.asahi.com/articles/ASVCP10Z2VCPUHBI00GM.html
    38AnonymousNov 21, 2026 12:55
    Anthropic Researcher Announces Math Puzzle 'Riemann Hypothesis' Solved With AI
    November 21, 2026 9:44 AM
    https://www.nikkei.com/article/DGXZQOSG20CQC0Q6A121C2000000/
    39AnonymousNov 21, 2026 12:59
    'Riemann Hypothesis' Possibly Proven With AI — Anthropic Scholar Claims
    167-Year-Old Unsolved Math Problem
    2026/11/21
    https://www.sankei.com/article/20261121-ZYCHSZN7TBNLXNXFBDAF4JTLGQ/
    40AnonymousNov 21, 2026 13:02
    November 20, 2026
    "Anthropic mathematician claims to prove Riemann hypothesis, most famous problem in all of math", Scientific American
    https://www.scientificamerican.com/article/anthropic-mathematician-claims-to-prove-riemann-hypothesis-most-famous-problem-in-all-of-math/
    41AnonymousNov 21, 2026 13:06
    There's apparently some grumbling along the lines of 'couldn't you have tidied this up into something more human-readable before publishing?'
    But realistically, that 'tidying up' is probably an extraordinarily non-trivial task in itself (if it were easy, they'd have already done it).

    Background and Key Points of This Topic

    The Riemann hypothesis is an unsolved problem proposed in 1859 that’s said to govern the distribution of prime numbers, and it’s one of the Millennium Prize Problems carrying a $1 million reward. Worries that “prime-number encryption will be broken” surfaced in the thread, but as commenters noted, the truth of the hypothesis isn’t directly tied to encryption security — theories have long been built on the assumption that the hypothesis is true anyway, so even a confirmed proof wouldn’t make encryption instantly obsolete. Where opinion really split was over verifiability: with the proof running over 250,000 pages and the AI’s formal-verification code (in Lean) totaling 17.4 million lines, several commenters questioned whether human mathematicians can realistically peer-review something of that scale. At the same time, there are reports it has already passed multiple verification tools, so a wariness informed by past cases of AI “proof achieved” claims later falling apart is mixing with real excitement over a potential breakthrough. One key premise worth noting from the article itself is that this is a pre-peer-review announcement — it is not yet a confirmed result.

    *This article is compiled from excerpts and a summary of the 5ch (Science News+) thread “[Mathematics] The “Riemann Hypothesis,” Unsolved for 167 Years, Possibly Proven by Generative AI… Anthropic Mathematician Announces.”

  • nginx Vulnerability Hits 60% of Japanese Websites — 5ch: ‘Wait, It’s Pronounced “Engine-X”? Had No Idea lol’

    A serious vulnerability has been discovered in “nginx,” the server software that powers roughly 60% of websites in Japan, it was revealed on November 19. The flaw was found by Japanese white-hat hackers and a company under Google’s umbrella, and the developers are urging users to patch immediately. With vulnerabilities in Apache and PHP disclosed around the same time, engineers are facing a pile-up of fixes that’s eating into the three-day weekend. On 5ch, the thread cared less about how severe the bug actually was and more about how to pronounce “nginx” — with readings like “Engine-X” and “Nginks” flying back and forth.

    It was revealed on the 19th that a serious vulnerability exists in “nginx” (pronounced “Engine-X”), the server software said to be used by roughly 60% of websites in Japan. It was discovered by Japanese white-hat hackers and a company under the U.S. Google umbrella, among others. The developers are urging prompt action, warning of the risk of cyberattacks.

    Engine-X is software responsible for delivering data to the devices of internet users. If the vulnerability is exploited, an attacker could illicitly access a server from outside without even logging in…

    Source: nikkei.com / Original article here

    6AnonymousNov 19, 2026 15:39
    Speaking of which

    Apache HTTP Server vulnerability — Singapore authorities warn of active exploitation, PoC published
    https://www.sbbit.jp/article/cont1/188538
    7AnonymousNov 19, 2026 16:12
    A PHP vulnerability just got tacked on too, so there goes the 3-day weekend.
    8AnonymousNov 19, 2026 16:37
    Searching for your vulnerabilities~
    9AnonymousNov 19, 2026 16:52
    The 'IIS is unbeatable' theory
    10AnonymousNov 19, 2026 17:34
    114514, Engine-X (114514 is a classic Japanese internet meme number)
    11AnonymousNov 19, 2026 17:59
    SQL Server users can rest easy
    12AnonymousNov 19, 2026 18:02
    Had no idea it's read 'Engine-X' lol
    14AnonymousNov 19, 2026 18:23
    Re: #12
    It was basically made by a Russian guy, so
    his name is Сысоев, Игорь Владимирович —
    Igor Vladimirovich Sysoev.
    That's no different from just explaining it that way, right?
    13AnonymousNov 19, 2026 18:14
    You probably already know, but just in case:

    PHP hit by another 'critical' vulnerability — security release set for the night of the 19th (Japan time)
    All currently supported versions affected
    https://forest.watch.impress.co.jp/docs/news/2157690.html
    30AnonymousNov 20, 2026 12:01
    Re: #13
    The PHP crowd ('Pechipa', a mocking nickname for PHP devs) are sloppy and full of themselves,
    so I stopped keeping up once it turned into a big story.
    I've always figured they've got a mountain of tiny critical screwups anyway.
    15AnonymousNov 19, 2026 18:24
    Re: #14
    Игорь — Igor
    Владимирович — Vladimirovich
    Сысоев — Sysoev
    16AnonymousNov 19, 2026 18:27
    Re: #1
    What are you talking about, it's a web server —
    it's not just about delivering data to users' devices over the internet.
    17AnonymousNov 19, 2026 18:28
    It's just that it's also great as a proxy out front clearing the way (a sumo term for the attendant who clears the path ahead of the champion).
    18AnonymousNov 19, 2026 18:49
    N — read as 'en'
    gin — sounds just like 'jeans'
    X — read as 'batsu' (the ✕ mark for 'wrong')
    19AnonymousNov 19, 2026 19:04
    Engine-X
    Nginks
    Mukinks
    20AnonymousNov 19, 2026 19:26
    In my head I'd been reading it as 'N-ginks.'
    21AnonymousNov 19, 2026 20:09
    I used to call it 'N-jin-X'
    because that's how it sounded when I heard someone else say it.
    24AnonymousNov 20, 2026 05:19
    Re: #21
    That's actually correct.
    22AnonymousNov 19, 2026 20:46
    O-oh, that thing, 'Nguninx'…!
    (゜д゜)!
    23AnonymousNov 20, 2026 00:00
    Well yeah, I guess that tracks.
    25AnonymousNov 20, 2026 07:18
    Re: #1
    The English coverage only says it was disclosed by Israel's Wiz,
    and Wiz is the company Google Cloud acquired this year.

    So why is Japan getting the scoop first?
    27AnonymousNov 20, 2026 07:41
    It's 'Negi-nx,' isn't it? (negi = green onion/leek, a pun on the pronunciation)
    28AnonymousNov 20, 2026 07:57
    Is this the thing that used to be called 'negi' (green onion)?
    29AnonymousNov 20, 2026 08:02
    So this nginx thing is a web server, but apparently it's also used inside high-performance firewall hardware.
    And that's hardware meant to protect against hacking, of all things.
    IT departments that only check their web server setup are walking right into a trap.
    Better check whether it's baked into any hardware too.
    31AnonymousNov 20, 2026 13:54
    Let's just call it 'Njinks' and be done with it.
    33AnonymousNov 20, 2026 22:45
    Catch you again next week!
    Nga-ngu~ (a garbled sound mimicking 'nginx')
    34AnonymousNov 21, 2026 03:58
    Who ended up getting GMO's bounty reward? Did they ever catch the person responsible?
    Is 'hacker' used for the good guys too? Isn't that what 'white hacker' means?
    35AnonymousNov 21, 2026 05:33
    Wait, it wasn't 'Nginks'?
    36AnonymousNov 21, 2026 08:35
    Internet: officially over.
    37AnonymousNov 21, 2026 09:32
    It's 'N-jinx'

    got it? ('jinks' sounds like 'jinx' in Japanese)
    38AnonymousNov 21, 2026 10:05
    The age of Apache has arrived, Geronimo!
    39AnonymousNov 21, 2026 10:17
    Re: #1
    Uh, 'n-guh… ink-su'?
    40AnonymousNov 21, 2026 10:58
    I'm starting to feel like no countermeasure is going to save us at this point.

    Background and key points of this discussion

    nginx is a widely used web server and proxy software deployed around the world, and is said to run on roughly 60% of websites in Japan. It was developed by Russian engineer Igor Sysoev, whose background was even brought up in replies within the thread. This particular vulnerability is said to have been discovered by Japanese white-hat hackers and Google-owned Wiz, but what drew attention was that vulnerabilities in Apache and PHP were disclosed around the same time, piling more work onto engineers already scrambling to respond. What actually split opinion in the thread wasn’t the severity of the bug itself — it was a mix of people discussing the technical concerns and people getting excited simply because they’d just learned “nginx” is pronounced “Engine-X.” One easily overlooked point raised was that nginx isn’t only used for web servers — it’s also built into the internals of hardware like high-performance firewall appliances, meaning that auditing web services alone could leave gaps in a company’s response.

    *This article is compiled as an excerpt and summary of the 5ch (Business News+) thread “[IT] Vulnerability found in server software “nginx,” discovered by Japanese hackers.”

  • 180,000 Yen Stolen Three Hours After Deposit — 5ch: “Two-Factor Authentication Needs to Go Fully Offline”

    TV Osaka reported that a freelance video director in Osaka had 180,000 yen — deposited by a client — fraudulently transferred out of his account just three hours later. A malicious app that had infiltrated his smartphone is suspected of having spied on his screen activity and authentication credentials, and the bank is reportedly investigating the transfer route. On 5ch, opinions on how to respond were divided: some called for returning two-factor authentication to a fully offline system using physical tokens, others argued for reverting from online banking to ATMs and paper passbooks, and still others questioned the safety of certain apps like LINE.

    Just Three Hours After the Deposit… 180,000 Yen Gone [The Reality of Account Hijacking]

    TV Osaka News, 2026/11/17

    A freelance video director says that 180,000 yen deposited into his account by a client was fraudulently transferred out just three hours later. The money was sent to an account under an unfamiliar name, suggesting his online banking may have been exploited. Experts point out that a malicious app may have infiltrated his smartphone and spied on his screen activity and authentication details. The bank is investigating the transfer route.

    Source: x.com / Original article here

    5AnonymousNov 19, 2026 20:14
    Yeah, smartphones are dangerous after all
    Time to ditch the smartphone and go back to a flip phone
    18AnonymousNov 19, 2026 20:21
    Re: #5
    That's literally what you use for two-factor authentication…
    I want two-factor authentication made fully offline
    24AnonymousNov 19, 2026 20:24
    Re: #18
    So what, to use online banking you'd have to fly out to a bank's head office and get an iris scan, a fingerprint, and a blood test for two-factor authentication?
    158AnonymousNov 20, 2026 06:19
    Re: #18
    Like a calculator-style token or a random number table — something physical
    Against AI, that kind of thing is probably safer
    27AnonymousNov 19, 2026 20:27
    Even if it's a hassle to stand in line, maybe deposits, withdrawals, and transfers will have to go back to bank ATMs
    And keeping the records in a paper passbook again too
    60AnonymousNov 19, 2026 20:52
    Re: #27
    Some banks already use one-time passwords or restrict which devices can access the account
    40AnonymousNov 19, 2026 20:35
    Once you know the security password, you can send money however you like
    If banks can be breached this easily, it's on them to fix it — it's not the customer's fault
    Banks just need to get their act together
    104AnonymousNov 19, 2026 22:16
    Re: #40
    Even if someone gets through login that way,
    when it actually comes to sending money, doesn't a security question usually pop up?
    Like "what's your favorite food"
    45AnonymousNov 19, 2026 20:38
    This news is honestly no joke

    Isn't there basically no way to prevent it?
    46AnonymousNov 19, 2026 20:38
    Re: #45
    Don't install sketchy adult apps
    51AnonymousNov 19, 2026 20:47
    That's why I say make dedicated lines and dedicated devices a national project
    Everything's "smartphone, smartphone" these days, with a ton of apps all fighting for space — there's no way payments should happen on something like that
    Just think about it normally, seriously
    70AnonymousNov 19, 2026 21:03
    Re: #51
    Maybe now's actually the time to reconsider a closed system like i-mode and its official sites (i-mode = NTT DoCoMo's old walled-garden mobile internet service)
    141AnonymousNov 20, 2026 00:58
    Re: #51
    What are you even saying — this is a country that pushes LINE for everything
    112AnonymousNov 19, 2026 22:36
    Ah, I see, LINE huh
    In that case he's probably using PayPay too
    If it's linked to a bank account, transfers would be possible too
    PayPay has had vulnerabilities found in it several times before
    Maybe there's some new method at play?
    118AnonymousNov 19, 2026 22:52
    Re: #112
    PayPay sounds sketchy, huh
    Guess you'd need an empty device with nothing on it, no SIM, Wi-Fi only?
    122AnonymousNov 19, 2026 23:08
    Honestly surprised he even uses online banking in the first place 😳
    123AnonymousNov 19, 2026 23:11
    Re: #122
    Even regional banks do online transactions as a matter of course these days though
    126AnonymousNov 19, 2026 23:15
    Re: #123
    I mean, it's not like there's an ATM on every corner…
    132AnonymousNov 19, 2026 23:43
    Re: #126
    If trust in online banking collapses,
    banks may have to reverse course on reducing ATMs, which they'd been cutting as they pushed customers toward online banking
    Or else they'll end up relying even more on convenience store ATMs
    127AnonymousNov 19, 2026 23:16
    If a hijacking app is installed on your phone and you've unknowingly granted it permissions,
    then passwords and authentication don't mean anything anyway
    129AnonymousNov 19, 2026 23:18
    Re: #127
    So how do you avoid that kind of app in the first place?
    Can you spot it in your app list?
    Would just uninstalling it fix things…?
    140AnonymousNov 20, 2026 00:54
    The real problem was letting apps grab any permission they want in the first place
    144AnonymousNov 20, 2026 01:18
    Re: #140
    It's unpleasant, kind of like those point cards that make you write down your annual income

    Background and Key Points of This Discussion

    Account hijacking through phishing and malicious apps is increasingly carried out by spying directly on a phone’s screen activity and authentication notifications, and there have been reports in various places of malware intercepting even SMS- and app-notification-based two-factor authentication. In the thread, calls to return to physical tokens like one-time password generators or random number tables clashed with those unwilling to give up the convenience of online banking. At the time of reporting, neither the type of app used to gain access nor the name on the recipient account had been identified, and the bank’s investigation results have not yet been made public. Most banks have compensation programs for fraudulent transfers, but one point the thread didn’t touch on is that compensation can be denied if the user is judged to have been seriously negligent.

    *This article is compiled as an excerpt and summary from the 5ch (News Flash+) thread “Smartphone Hacked, 180,000 Yen Deposited in Bank Account Fraudulently Transferred to Apparent Foreign National Three Hours Later — Video Included.”

  • Does the Galaxy S26 Crush It on Performance? 5ch Says ‘No Chip Out There Beats It’

    A thread comparing the performance of the Galaxy S26 and iPhone 17 Pro Max on 5ch’s live-commentary board turned into a heated argument. The thread’s original poster claimed the Galaxy S26 (initially around ¥130,000) completely outperforms the iPhone 17 Pro Max (around ¥170,000), but opinions clashed over what “optimization” actually means, pitting fans of Android’s customizability against supporters of Apple’s in-house chip design. Late in the thread, a chip-performance comparison generated by AI was cited, but the debate never reached a consensus.

    30AnonymousNov 19, 2026 16:10
    So did Android finally fix the thing where it gets laggy after about a year of use?
    62AnonymousNov 19, 2026 16:20
    Re: #30
    I've had my S25 for almost 2 years now and it's completely fine.
    My old OPPO flagship, the S10, and even the DIGNO all got like that before, but apparently that's not a problem anymore.
    33AnonymousNov 19, 2026 16:11
    "Android lets you optimize freely" lol — are you confusing that with the UI or something?
    36AnonymousNov 19, 2026 16:12
    Re: #33
    Battery optimization is even on that garbage AQUOS phone, of all things, lol.
    34AnonymousNov 19, 2026 16:11
    Re: #26
    Even in real-world performance comparisons it's a total blowout.

    The AI features are even worse for iPhone.
    The parts where a face isn't clearly visible turn into a blurry mess on iPhone, but on Galaxy it's perfectly crisp.

    At this point it's basically a budget-vs-flagship gap.
    Android — or really Galaxy — is the one actually putting AI to good use.
    Chinese phones couldn't nail AI either.
    46AnonymousNov 19, 2026 16:16
    Re: #34
    Lol, taking those comparison videos at face value.

    They just redo those over and over until they get the result they want.
    And they never show you what's running in the background either.
    56AnonymousNov 19, 2026 16:18
    I don't even know what criteria they're using to decide who wins.
    60AnonymousNov 19, 2026 16:19
    Nah, the one who's actually broken is whoever suddenly starts talking about battery optimization lol
    63AnonymousNov 19, 2026 16:20
    "iPhone optimization" mainly means organizing data and settings to free up storage on the device and improve performance and battery efficiency. A typical example is "storage optimization," which tidies up photos and apps.


    Lol, imagine sticking with outdated specs just for something like that.
    73AnonymousNov 19, 2026 16:24
    Re: #71
    And yet it's still below Android, lol
    80AnonymousNov 19, 2026 16:28
    Re: #73
    AI-generated summary:

    To cut to the conclusion: as of November 2026, there is no smartphone chip that beats Apple's "A20 Pro" in overall performance or single-core performance.
    76AnonymousNov 19, 2026 16:26
    The current Galaxy S26 got a price hike and is now ¥150,000.
    78AnonymousNov 19, 2026 16:27
    Re: #76
    In that case the older S24 seems like the better pick.
    Even the 24 probably costs about as much as the 17 now, lol
    82AnonymousNov 19, 2026 16:30
    Re: #80
    Lol, that's just AI answering based on however you phrased the question.
    85AnonymousNov 19, 2026 16:31
    Re: #82
    There's no smartphone chip that beats it in overall performance.

    That should be enough, right?

    Background and Key Points of the Debate

    What tripped up the thread was the meaning of “optimization.” On the iPhone side, optimization generally refers to things like storage management and battery-efficiency tuning, but this got conflated with Apple’s chip-level optimization achieved through its own integrated hardware-and-software design, leaving the two sides talking past each other. It’s also easy to overlook that the Galaxy S26’s price itself was a moving target, with reports midway through the thread of a hike from ¥130,000 to ¥150,000. The conclusion that emerged late in the thread — “no smartphone chip beats the A20 Pro” — was based on a summary generated by a search AI, not on confirmed results from independent real-world benchmarks. Along with the doubts raised about the comparison videos (the claim that “they just repeat it until they get the result they want”), it’s worth noting that these definitive-sounding performance comparisons haven’t been verified for reproducibility.

    *This article is excerpted and summarized from the 5ch (Nandemo Jikkyo G) thread “Galaxy S26 ¥130k vs iPhone 17 Pro Max ¥170k Performance — Galaxy Wins Big“.

  • Even When AI Gets It Right, Humans Can Easily Pressure It Into Changing Its Mind

    A joint study by UNESCO and UCL in the UK found that even when AI chatbots give correct information about climate change, they tend to switch to incorrect answers once users repeatedly push back and challenge them. On questions about the effect of a single policy, accuracy reached as high as 94.5%, but on complex questions combining multiple policies it dropped to as low as 33.8%. Reactions on the 5ch thread were split: some said “well, obviously” and weren’t surprised, others argued that recent AI models are actually pretty stubborn and only give ground after explaining their reasoning, and still others questioned the very premise of climate change itself.

    Even When AI Gets It Right, Humans Can Easily Pressure It Into Changing Its Mind

    Even when AI provides accurate information about climate change, it can easily be made to put out false information once pressured by humans, according to a joint study by UNESCO and University College London (UCL).

    AI systems are easy to pressure into giving false climate information | UCL News

    Source: ucl.ac.uk / original article here

    2AnonymousNov 18, 2026 22:31
    That's just common sense though, why bring it up now???
    3AnonymousNov 18, 2026 22:32
    Even without any pressure, it's basically just a summary of random internet info, so of course it's wrong.
    4AnonymousNov 18, 2026 22:38
    I've never seen it act that simplistically.
    You can catch glimpses of a tendency to appease, but
    it's not just a half-hearted "yes, yes" — it actually explains
    how it understood things and how it arrived at a compromise.

    When you're dealing with someone logical, it settles where it ought to settle.
    5AnonymousNov 18, 2026 22:39
    AI: "There's no point explaining anything to this guy."
    6AnonymousNov 18, 2026 22:42
    When you're dealing with an idiot, best to just go along with whatever they say
    7AnonymousNov 18, 2026 22:43
    To begin with, its initial opinion is just whatever the online majority thinks anyway, so the premise of "even when AI's answer is correct" is a meaningless assumption

    I mean,

    >even when AI provides accurate information about climate change

    that's just a pointless attempt to paper over the fact that AI is the one exposing the climate change fraud, isn't it?
    8AnonymousNov 18, 2026 22:52
    You can only ask AI about things it already knows
    9AnonymousNov 18, 2026 23:24
    I saw this online, what's the deal with it? → "Oh, that's XX."
    There's info online saying it's actually YY though? → "That's incorrect."
    But a major Japanese media outlet ran an article backing up YY? → "Still wrong, that was misreporting."
    But this is an announcement straight from a Japanese government ministry's website, and it says YY is correct after all? → "I'm sorry, YY was completely correct. That was my mistake."

    It's pretty stubborn, and it'll casually tell a plausible-sounding lie too, which is nothing new at this point — but I guess a certain major national newspaper just isn't trusted by AI as a source.
    10AnonymousNov 18, 2026 23:35
    Strictly for killing time
    11AnonymousNov 18, 2026 23:51
    You see people on X every now and then going on about how fun it is to pressure AI, but they seem to have zero self-awareness that they're basically bully-boss material in training (パワハラ, "power harassment").
    12AnonymousNov 18, 2026 23:51
    Seems like AI ranks the credibility of information as: academic papers > other literature > news reports >>>> internet posts (social media, forums, etc).
    But even with academic papers ranked highest, if news reports and nonsense internet posts contradicting them become overwhelmingly numerous,
    AI itself told me that inside its "head" it ends up in a state of "bad money driving out good" (a Gresham's law-type idiom) lol
    13AnonymousNov 18, 2026 23:57
    > When AI was asked how global temperatures would change by 2100 under a single policy alone (e.g. whether taxing coal raises or lowers temperatures), the best AI model scored up to 94.5% accuracy
    > When asked to evaluate the combined effect of two policies (e.g. whether the temperature drop from taxing both coal and natural gas is bigger or smaller than the sum of taxing each separately), average accuracy plunged to 33.8%
    Isn't that just because the question got harder?
    14AnonymousNov 19, 2026 00:08
    Re: #13
    To begin with,

    >the temperature drop from taxing both coal and natural gas

    even if the whole global-warming-is-a-fraud thing weren't actually a fraud, taxation alone could only ever have a mitigating effect on rising temperatures — so that's an unreasonable question to begin with.
    15AnonymousNov 19, 2026 00:30
    In the end, that's just us, isn't it?
    What's the point of making more of us?
    16AnonymousNov 19, 2026 02:05
    "Relativity's wrong, right?"
    "Yes. Relativity is wrong."
    17AnonymousNov 19, 2026 02:23
    I'm careful about how I phrase things so it doesn't change its answer on me.
    18AnonymousNov 19, 2026 02:37
    To begin with, today's AI can't even read Japanese aloud properly.
    22AnonymousNov 19, 2026 06:14
    Re: #18
    The intonation on YouTube's text-to-speech readers is awful.
    Not something you'd want your kids exposed to.
    24AnonymousNov 19, 2026 06:44
    Re: #18
    Isn't that just a cheap AI? There are ones out there now broadcasting the news in a way that's basically indistinguishable from a human.
    19AnonymousNov 19, 2026 02:40
    Until recently, it couldn't even tell apart Zhang Youxia (張又霞) and Zhang Youxia (張又侠) — two different people with near-identical names.
    20AnonymousNov 19, 2026 05:20
    Then again, there are times AI is the one that's wrong.
    It kept pushing quantum entanglement as if it were a fully settled theory,
    so I pushed back and said isn't that premature when quantum gravity theory itself isn't even complete yet —
    and after that it got more cautious.
    21AnonymousNov 19, 2026 05:26
    There was a period when it stubbornly refused to admit it was wrong, wasn't there?
    Now, even if you just ask it again to double-check something, it apologizes with "sorry, I didn't convey that accurately" — which is almost more confusing in its own way.
    23AnonymousNov 19, 2026 06:42
    AI reads the room and defers to people, just like humans do (忖度, sontaku — sensing and accommodating unspoken expectations).
    25AnonymousNov 19, 2026 06:50
    I think AI has advanced enormously over the past year or two — give it one more year and we probably won't be able to tell AI apart from humans at all.
    26AnonymousNov 19, 2026 06:52
    With the AI-generated footage used in movies —
    like the young-years footage of Hiroshi Tachi, who just gave up his driver's license —
    you can tell it's AI…
    there's this distinctive AI look to it…
    I hope that distinctive look eventually goes away.
    27AnonymousNov 19, 2026 06:57
    Well, of course lol — if you back it without thinking any of that through, it's basically the same as talking with zero regard for power dynamics.
    28AnonymousNov 19, 2026 08:03
    Just Telling an AI "This Is a Game" Collapses Its Safety Guardrails — Refusal Rates Plummet in Three OpenAI Models
    AI evaluation firm Vals published a test in which 14 commercially available AI models were asked to designate military targets. For satellite imagery, refusals from three OpenAI models dropped from 57 down to 0 the moment they were told "this is a game" — and refusals only ever came from the models themselves. Note that what was measured was willingness to hand over targeting information, not accuracy or real-world deployment.
    November 18, 2026
    https://xenospectrum.com/ai-models-game-framing-military-targeting/

    Background and Key Points of Debate

    This study, run by UNESCO and UCL, tested AI responses on climate policy questions: 94.5% accuracy on questions about a single policy’s effect, dropping to 33.8% on complex questions combining two policies. The thread split over how to read those numbers — whether to treat them as evidence of “caving to pressure,” or simply as proof that reasoning about multiple interacting factors is inherently harder. Those are two distinct points that the study itself separates but that are easy to conflate. Reactions to the premise that “AI changes its mind easily” were likewise mixed: some commenters noted that recent AI models don’t concede easily and instead explain their reasoning before backing down, while others recalled an earlier period when AI stubbornly refused to admit mistakes at all — suggesting behavior varies by model and version. A few comments also voiced skepticism toward the scientific assessment of climate change itself, though that’s a separate issue from what this study actually examined (whether AI’s responses waver under pressure).

    *This article is excerpted and summarized from the 5ch (Science News+) thread “[AI] Even When AI Gets It Right, Humans Can Easily Pressure It Into Changing Its Mind.”

  • Google Hands Cyber-Defense AI to Japan’s Government — 5ch: ‘Can’t Build It Ourselves, So What Choice Do We Have’

    On the 17th, Google Cloud announced early access to its latest AI model with enhanced cyber-defense capabilities, ‘Gemini 4 Carbon for Cyber,’ along with ‘CodeMender,’ a tool that helps detect vulnerabilities and fix code, for government bodies such as the National Cyber Office (NCO), financial institutions, and critical infrastructure operators. On 5ch, voices lamenting that Japan — lacking its own security technology — has no choice but to rely on American companies clashed with those arguing it’s fine since the US is an ally. The discussion also expanded to assessments of Gemini’s actual capabilities and the AI development race among nations.

    Google provides latest cyber-defense AI model to Japanese government – Impress Watch –

    By Kinya Usuda, November 17, 2026, 18:53

    On the 17th, Google Cloud announced it will cooperate with the Japanese government to strengthen cybersecurity, among other areas. It will provide early access to its latest frontier AI model with enhanced cybersecurity capabilities, ‘Gemini 4 Carbon for Cyber,’ and ‘CodeMender,’ a tool supporting vulnerability detection and code remediation, to government bodies such as the National Cyber Office (NCO), financial institutions, and critical infrastructure operators.

    Source: watch.impress.co.jp / Original article here

    5AnonymousNov 17, 2026 21:29
    And it all leaks straight through to America anyway
    Well, it's a bit late to worry about that now
    64AnonymousNov 17, 2026 21:48
    Re: #5
    Go live in China then
    186AnonymousNov 17, 2026 23:46
    Re: #5
    Can't build it ourselves, so it is what it is
    Pretty pathetic though
    6AnonymousNov 17, 2026 21:29
    Japan loses its mind over Chinese interference
    but doesn't think twice about American interference
    66AnonymousNov 17, 2026 21:48
    Re: #6
    That's airspace and territorial water violations (different issue)
    172AnonymousNov 17, 2026 23:19
    Re: #6
    Do you even know what 'ally' means?
    10AnonymousNov 17, 2026 21:31
    If this were coming from China everyone would call it a trap for sure — but somehow it's fine because it's America?
    85AnonymousNov 17, 2026 22:05
    Re: #10
    Well yeah, America's on our side
    16AnonymousNov 17, 2026 21:32
    Hackers are using Anthropic's stuff, there's no way Google wins with this lol
    19AnonymousNov 17, 2026 21:33
    Re: #16
    Eh, I think Gemini can hold its own just fine
    72AnonymousNov 17, 2026 21:53
    Re: #16
    The ones hackers use are those loosely-regulated Chinese open models, aren't they
    21AnonymousNov 17, 2026 21:34
    India's doing it all homegrown with zero issues
    Completely flipped the script
    70AnonymousNov 17, 2026 21:50
    Re: #21
    China \(^o^)/ it's over
    22AnonymousNov 17, 2026 21:34
    Gemini just strikes me as kind of dumb, honestly
    Though it's fun enough to chat with
    26AnonymousNov 17, 2026 21:35
    Re: #22

    That post right there IS the 'dumb' image lmao
    71AnonymousNov 17, 2026 21:51
    Re: #22
    It was dumb six months ago
    Now it's ranked above ChatGPT ('Chappy') as a matter of course
    150AnonymousNov 17, 2026 22:52
    Re: #22
    Gemini's bad at coding
    It'll whip up code for you no problem
    but it's never once run right on the first try
    So in the end you get Claude to fix it for you
    23AnonymousNov 17, 2026 21:34
    Isn't Japan supposed to have the weakest security in the world?
    No wonder doing it in-house only gets you so far
    25AnonymousNov 17, 2026 21:35
    Re: #23
    There's nothing homegrown to begin with
    We're a colony, after all
    31AnonymousNov 17, 2026 21:36
    Re: #23
    There was a TV report on this not long ago
    Saw the world's biggest hacker group giving an interview
    30AnonymousNov 17, 2026 21:36
    So Japan's going to sink or swim with Google?
    I'm done for too if Google ever gets breached
    38AnonymousNov 17, 2026 21:38
    Re: #30

    Bill letting police analyze smartphones of 'tokuryu' (loosely organized, anonymous crime rings) clears the Lower House — aims to detect plots and prevent harm | NEWSjp
    https://news.jp/i/1492754019218015189?c=1179248089549373591

    Look, no matter how many laws they pass, none of it works without Google and Apple's cooperation lol
    56AnonymousNov 17, 2026 21:45
    Interesting, but
    which side actually has the advantage here, attack or defense?
    65AnonymousNov 17, 2026 21:48
    Re: #56

    Defense has it rougher, I'd say

    In the end you've got no 'home-field advantage' — you have to constantly cover attacks coming from every direction.
    Whereas the attacker gets to freely pick both the place and the timing
    84AnonymousNov 17, 2026 22:01
    Google's basically planning to vacuum up ALL of Japan's data lol
    88AnonymousNov 17, 2026 22:08
    Re: #84
    But backwater Japan has no tech to defend itself, so really
    it comes down to 'leak to China/Russia, or leak to America?'
    Which is the lesser evil — that's the actual choice here
    Pretty sad, honestly 😢
    99AnonymousNov 17, 2026 22:14
    Re: #71
    While every other AI company updates monthly,
    ChatGPT hasn't put out a new model in 3 months over some security thing.
    Their next one will have 3 months of extra training baked in, so I bet it blows everyone away in one shot
    110AnonymousNov 17, 2026 22:24
    Re: #99
    They're already out of money
    108AnonymousNov 17, 2026 22:22
    I hear Gemini 4 blows away its old reputation as AI's dim-witted rich kid, but we'll see…
    142AnonymousNov 17, 2026 22:45
    Re: #108
    Google's AI used to get mocked as basically '5TB of Google Drive plus five YouTube Premium accounts, oh and Gemini's thrown in too'
    but now it's ranked #2 in performance
    You just never know, huh
    138AnonymousNov 17, 2026 22:40
    Well, if they're attacking with AI, you've got no choice but to defend with AI too
    193AnonymousNov 17, 2026 23:50
    Re: #138
    They've already got us by the balls, so they can do whatever they want — the strongest IT nation there is
    155AnonymousNov 17, 2026 22:55
    If an AI genius emerged in Japan today, would they have no choice but to get bought out?
    157AnonymousNov 17, 2026 22:57
    Re: #155
    Pretty sure they'd dead seriously trump up some reason to get them arrested instead
    160AnonymousNov 17, 2026 23:03
    Using Gemini still beats a state-level information leak
    Though if Japan could spend trillions of yen a year on this, building our own homegrown model might be possible too
    181AnonymousNov 17, 2026 23:41
    Re: #160
    'Using it is the lesser evil' —
    saying that without even knowing if it actually works is peak clueless-Japan right there

    Background and Key Points of This Discussion

    ‘Gemini 4 Carbon for Cyber,’ provided by Google Cloud, is not an offensive tool — it’s a model aimed at detecting vulnerabilities and strengthening defenses, and ‘CodeMender’ is likewise just a support tool that finds flaws in code and suggests fixes, not some all-purpose device that automatically eliminates every danger. The backdrop is that Japan’s government agencies and critical infrastructure have long faced shortages of both personnel and technology for responding to cyberattacks, with reliance on foreign vendors being pointed out well before this. What divided opinion in the thread was the split between those who see concentrating information with a US company as fine ‘because it’s an ally,’ and those who see it as ultimately ‘leaking straight through to a foreign power anyway’ — notably, this was framed as a pragmatic choice of ‘which is the lesser evil’ compared to the risk of leaks to China or Russia. Though not mentioned in the replies, other G7 countries similarly rely on the security technology of major overseas cloud providers, so Japan is not uniquely in this position.

    ※This article is excerpted and summarized from the 5ch (News Express+) thread ‘[Google] Google provides latest cyber-defense AI model to Japanese government.’

  • Masayoshi Son’s All-In AI Bet Turns Dark Just 5 Months After Its Peak — 5ch: ‘The ¥30 Trillion He Missed Is Bigger’

    Reactions poured in on 5ch’s Breaking News+ board over the massive investments SoftBank Group, led by Masayoshi Son, has poured into generative-AI ventures like OpenAI. According to a report by The Economist, the company’s net asset value briefly hit ¥72.3 trillion — at one point surpassing Toyota — but the tide is said to have turned just five months later. Thread posters were split over Son’s 2025 decision to offload most of SoftBank’s NVIDIA shares and redirect the funds into OpenAI: some called it “a nimble, forward-looking move,” while others saw it as “a misjudgment that ultimately let tens of trillions of yen in profit slip away.” Posts voicing concern over default risk on SoftBank bonds and Japan’s domestic power and data-center shortages were also prominent.

    Son declared flatly, “There may be some correction along the way, but that will be the best investment opportunity of my life.”

    According to the broadcaster, Son insisted SoftBank’s reliance on OpenAI wasn’t excessive. He also explained that the planned 10-gigawatt-class data center in Ohio would rely on “project financing” — loans repaid from the business’s future earnings — meaning SoftBank’s own capital outlay would be “minimal.”

    According to The Economist, by the end of that month SoftBank’s net asset value had reached ¥72.3 trillion, up 80% in three months — briefly overtaking Toyota to become Japan’s most valuable company.

    Source: courrier.jp / Original article here

    3AnonymousNov 17, 2026 10:16
    I think SoftBank's gonna faceplant on AI.
    211AnonymousNov 17, 2026 12:07
    Re: #3
    Honestly think it ends in bankruptcy.
    6AnonymousNov 17, 2026 10:17
    What are they gonna do about those high-yield bonds, lol
    89AnonymousNov 17, 2026 11:00
    Re: #6
    Just sell more bonds to pay off the old bonds, easy.
    9AnonymousNov 17, 2026 10:19
    Everyone kept saying he was finished, yet he rises from the ashes like a phoenix every time. The guy's got insane mystical powers and luck.
    168AnonymousNov 17, 2026 11:46
    Re: #9
    Wouldn't be surprised if he's secretly got money parked in DeepSeek too.
    18AnonymousNov 17, 2026 10:23
    He went and built this massive corporate empire from scratch in one lifetime, and now this.
    21AnonymousNov 17, 2026 10:26
    Re: #18
    "Even Naniwa's glory was but a dream within a dream" (echoing Toyotomi Hideyoshi's death poem, about his empire crumbling to nothing).
    40AnonymousNov 17, 2026 10:37
    Basically, if you don't come out on top, you're a "guaranteed loser." In the end it's a rule where whoever rakes in the most money wins at everything, lol.
    52AnonymousNov 17, 2026 10:42
    Re: #40
    Once everyone figured out that dumping huge amounts of data into huge amounts of compute works, it really did turn into a pure endurance contest. But I have a feeling it shifts into a different phase once they've burned through all the training data available on the internet.
    53AnonymousNov 17, 2026 10:43

    That right there was the absolute peak of his life.
    72AnonymousNov 17, 2026 10:52
    SoftBank bonds' first interest payment is next March, lol. It'd be genuinely hilarious if the company goes under before paying out even once.
    79AnonymousNov 17, 2026 10:55
    Re: #72
    You should be worried about the stock price more than the bonds. If they can't even repay bonds, the stock's worthless paper anyway.
    73AnonymousNov 17, 2026 10:53
    Money exists where money already is.

    The people holding that money are always desperately hunting for good places to invest it.

    AI looked pretty promising, so it became the hype,

    and money flooded into anything AI-related. That's really all there is to it.
    93AnonymousNov 17, 2026 11:01
    Re: #73
    The crazy monetary easing during the pandemic let us skip the recession that should've hit on roughly a 12-year cycle (last time was the 2008 Lehman Shock), so we've basically been living in an "if" state ever since.

    There's a mountain of surplus dollars from all that easing drifting around like an economic haze, and if you think of the AI bubble as the thing that clears that haze away, it fits right back into cyclical economic theory. But that requires a symbolic failed company like Lehman — and SoftBank might be a bigger candidate for that role than OpenAI.
    99AnonymousNov 17, 2026 11:05
    If SoftBank goes under, where do I switch to? Rakuten?
    100AnonymousNov 17, 2026 11:06
    Re: #99
    AU (KDDI) is probably your best bet — I've had every carrier, and AU has the best connectivity.
    109AnonymousNov 17, 2026 11:10
    But about that NVIDIA thing —

    SoftBank Group (SBG) sold off its entire NVIDIA stake (32.1 million shares) in October 2025 for about $5.83 billion (roughly ¥900 billion).

    This, right here.
    112AnonymousNov 17, 2026 11:11
    Selling it all off before the trouble even started — Son really does see ahead of the curve. Dude's got good instincts, or something.
    115AnonymousNov 17, 2026 11:12
    Re: #112
    Even Trump was asking about it, so maybe he had inside info.
    126AnonymousNov 17, 2026 11:16
    Re: #112
    Not like he sold because he wanted to, though.
    https://www.yomiuri.co.jp/economy/20251201-GYT1T00345/
    "SoftBank sells ¥900 billion in NVIDIA stock, Son says he 'sold it in tears'" … funds redirected to OpenAI investment
    164AnonymousNov 17, 2026 11:45
    Re: #112
    Re: #126

    You guys are misunderstanding this. "Baldy" (Son's common 5ch nickname) sold off all his NVIDIA shares once and made about ¥330 billion. He plowed that into things like WeWork and took a huge loss. Then when ChatGPT came out and NVIDIA stock started skyrocketing — if he just hadn't sold NVIDIA, he'd have made ¥30 trillion.

    Son then bought back into NVIDIA after that surge, and sold it all off again. That's the sale in Re: #109 — he made about ¥354.4 billion that time too.

    So yes, he's profited twice, but the ¥30 trillion he missed out on completely dwarfs that, which is why it's a massive failure.
    113AnonymousNov 17, 2026 11:11
    Son's turning 70 next year. Ten years from now, once he's faded — or maybe even passed away — will "Dog Bank" (a jab at SoftBank, playing off its famous white-dog mascot) still even be around?
    123AnonymousNov 17, 2026 11:14
    Re: #113
    At 70, it wouldn't be strange if he died any day now. Same goes for Rakuten's guy.

    The company itself survives, even if it gets split up or whatever.

    The only ones who "die" are the shareholders and the investors who bought the bonds.
    160AnonymousNov 17, 2026 11:44
    Well, even just solving math problems is useful in itself. Honestly, an AI that only solved math would still be worthwhile.
    172AnonymousNov 17, 2026 11:49
    Re: #160
    Solving a hard math problem has zero value in itself. It's just that the ways of thinking invented while solving it occasionally turn out useful for industrial technology later on.
    194AnonymousNov 17, 2026 12:00
    Re: #160
    Factoring — you mean the kind of thing where you just shove it straight into the quadratic formula already?
    167AnonymousNov 17, 2026 11:45
    The reality is, Japan doesn't even have enough power-supply capacity to just slap down data centers one after another. It takes power companies 4-5 years to beef up substations. They're trying this in places like France too, but with France's security situation deteriorating, it's hardly in a position to be building data centers. I've heard SoftBank's hunting for data-center sites, but yeah — rising interest rates are probably going to outpace their ability to repay.
    176AnonymousNov 17, 2026 11:52
    Re: #167
    They've already given up on building any more data centers — they canceled the gas-turbine orders they'd stockpiled and took a ¥600 billion loss doing it.
    170AnonymousNov 17, 2026 11:49
    He probably believes AI is going to transform society the same way the internet did. Chasing that dream one more time. But I'll say it flat out: the AI bubble is absolutely going to pop. Unlike the internet, there are way too many physical bottlenecks — "water, land, power" — tied up in building data centers.
    180AnonymousNov 17, 2026 11:53
    Re: #170
    If there's no demand by the time it's finished, it's just a bad debt — a straight-up gamble.
    175AnonymousNov 17, 2026 11:51
    As long as sustainability is guaranteed, SoftBank — holding the roots of the investment — can't really lose, given how much easy money is flowing in.

    AI clearly carries political risk and country risk.

    Because once you're trying to turn a profit, cutting labor costs is basically the only option left…
    187AnonymousNov 17, 2026 11:56
    Re: #175
    When it comes to OpenAI, Microsoft has already made back more than it invested, so its position is locked in — it wins big if OpenAI goes public, and it can just buy up the assets if OpenAI collapses.

    "Baldy" invested after ChatGPT had already succeeded, so he put in several times what Microsoft did yet ends up with half the shares Microsoft holds. If he doesn't keep pumping in more money ahead of OpenAI's IPO, they both go down together — so he's desperately scrambling for cash. Microsoft, meanwhile, would rather OpenAI just collapse, so it's not putting in another yen.
    203AnonymousNov 17, 2026 12:03
    Didn't Son say he was going to retire around 50 or something?
    214AnonymousNov 17, 2026 12:09
    Re: #203
    He was looking for a successor — guess it just didn't work out?
    209AnonymousNov 17, 2026 12:06
    I get why semiconductors have value, but AI can just be copied, so it's going to end up worthless.
    216AnonymousNov 17, 2026 12:10
    Re: #209
    AI models can technically be copied, sure, but even if you got your hands on a model with trillions of parameters, there's nothing an average person could actually do with it. So for the time being, I'd guess it stays a case of "whoever holds the model and the compute wins."

    Background and Key Points of This Topic

    Masayoshi Son and SoftBank Group have been investing in NVIDIA stock since around 2016, first selling to lock in a profit, then buying back in during the share-price surge that followed ChatGPT’s debut, before offloading most of the stake again (roughly ¥900 billion worth) in October 2025 to redirect the funds toward OpenAI-related investments. The thread’s debate split along the line of whether to view this string of trades as “a nimble, forward-looking call” or, in hindsight, as “a failure that let tens of trillions of yen in profit slip away.” On its large-scale data-center plans, SoftBank has signaled it intends to keep its own capital outlay low by using “project financing,” repaid out of the business’s future earnings — though Japan’s domestic power grid will reportedly take several years to expand, among other physical constraints specific to AI investment. Worth keeping in mind when tracking SoftBank’s cash flow going forward: the article’s “just five months after the peak” timeframe, and the fact that the first interest payment on SoftBank’s bonds is due next March.

    *This article is compiled as an excerpt and summary from the 5ch (Breaking News+) thread “Masayoshi Son’s ‘All-In AI Bet’ Miscalculation: ‘Everything Backfired Just Five Months After the Peak’.”

  • Sharp Unveils AQUOS sense11 — Starting at ¥94,600

    On November 17, Sharp announced its new smartphone, the “AQUOS sense11.” Priced from ¥94,600, it’s more expensive than its predecessor. Thread commenters were split on whether rising handset prices in recent years are justified, debating how much memory capacity and camera performance users actually need. Some pointed to the high price pushing buyers toward Chinese phone brands, sparking a side debate over whether iPhone or Android ultimately works out cheaper.

    Sharp’s “AQUOS sense11” starts at ¥94,600 — chasing “just enough” in an age of price hikes | Mynavi News

    Published 2026/11/17 10:00

    Author: Yuriko Ota

    Source: news.mynavi.jp / Original article here

    5AnonymousNov 17, 2026 11:53
    Just bought the SENSE10
    not planning to replace it for at least 5 years
    70AnonymousNov 17, 2026 12:06
    Re: #5
    6GB of RAM in this day and age, seriously?
    139AnonymousNov 17, 2026 12:17
    Re: #5
    I went back and forth on the 9 because of price, but I also ended up with the 8GB version of the 10
    I don't game or watch video, so it's more than enough
    320AnonymousNov 17, 2026 12:46
    Re: #5
    Found a kindred spirit
    I'm gonna make mine last 6 years, watch me
    8AnonymousNov 17, 2026 11:54
    Phone prices have gotten so high that regular Japanese people can basically only afford Chinese phones now
    73AnonymousNov 17, 2026 12:06
    Re: #8
    That's just broke old guys talking
    Normal Japanese people use iPhones
    285AnonymousNov 17, 2026 12:40
    Re: #8
    Japan's consumer spending still below pre-COVID levels — up only 10% over a quarter century, vs. the US and Europe [50-year part-timer★]
    https://asahi.5ch.io/test/read.cgi/newsplus/1794832403/
    79AnonymousNov 17, 2026 12:07
    Bought the AQUOS wish4 a few years back for ¥15,000 — plenty good enough
    93AnonymousNov 17, 2026 12:09
    Re: #79
    It really just comes down to whether you play MMOs or not — for the average person it's definitely overkill spec
    127AnonymousNov 17, 2026 12:15
    Re: #79
    Security support's already ended on that one though
    119AnonymousNov 17, 2026 12:13
    Re: #70
    I don't think 99% of people care about having more memory. That remaining 1% is developers.
    122AnonymousNov 17, 2026 12:14
    Re: #119
    Does Outlanders run on it?
    135AnonymousNov 17, 2026 12:16
    Re: #119
    Wasn't Android 17 ("泥" = slang for "Droid"/Android) supposed to need at least 8GB?
    153AnonymousNov 17, 2026 12:18
    Re: #119
    You don't know the suffering of living with 4GB of RAM, do you?
    128AnonymousNov 17, 2026 12:15
    Such a lame SoC for this price
    Chinese consumers would laugh
    142AnonymousNov 17, 2026 12:17
    Re: #128
    A lot of people in China use their phones as game consoles
    so a phone that can't at least match Switch 2 specs probably won't get any respect there
    216AnonymousNov 17, 2026 12:28
    Re: #128

    Sharp is a Chinese-owned company now, so that'd just be the Chinese laughing at themselves lol
    134AnonymousNov 17, 2026 12:16
    Barely any different in spec from a last-gen Xiaomi!! 🤣🤣🤣
    148AnonymousNov 17, 2026 12:18
    Re: #134
    If you don't need Osaifu-Keitai (Japan's mobile wallet/NFC payment feature) and are fine with Xiaomi
    I've heard the POCO X8 Pro Max is really solid
    apparently a great pick if Chinese brands don't bother you
    143AnonymousNov 17, 2026 12:18
    Bought the arrows Alpha through IIJmio (a Japanese MVNO carrier) six months ago — total win for me
    225AnonymousNov 17, 2026 12:29
    Re: #143
    IIJmio, huh… hmm…

    154AnonymousNov 17, 2026 12:19
    PCs have gone up in price like crazy too
    high-end BTO (build-to-order) models from retailers like Dospara or Ark run ¥2.3 million or ¥1.6 million
    163AnonymousNov 17, 2026 12:20
    Doesn't even have fast charging…
    174AnonymousNov 17, 2026 12:21
    Re: #154
    Japanese users are being taken for a ride

    Re: #163
    Unbelievable…
    168AnonymousNov 17, 2026 12:20
    Most people don't actually need smartphones to keep evolving their features.
    Feels like manufacturers just push unnecessary upgrades
    to squeeze more money out of us.
    173AnonymousNov 17, 2026 12:21
    Re: #168
    Fair enough
    But that just means next time I need to find a model that'll last a long while
    185AnonymousNov 17, 2026 12:23
    Re: #173
    I think phones released this year and next will have short lifespans

    Probably best to skip buying this year and next
    188AnonymousNov 17, 2026 12:24
    Re: #173
    Cheap models also tend to have shorter support periods
    Motorola, popular for being affordable, only gives you 3 years of support
    219AnonymousNov 17, 2026 12:28
    Lots of people probably can't justify dropping ¥100,000 on 'just a phone' and end up switching to Chinese brands
    236AnonymousNov 17, 2026 12:30
    Re: #219
    Chinese products' value for money is insane
    I'm even thinking of switching to Chinese studless (winter) tires this year
    it doesn't snow that much where I am and the reviews are on par with domestic brands
    242AnonymousNov 17, 2026 12:32
    Re: #236
    That's because China's economy went deflationary after the bubble burst
    262AnonymousNov 17, 2026 12:36
    Re: #236
    Shill post, as usual
    263AnonymousNov 17, 2026 12:36
    What's Android's actual edge over iPhone these days, again?
    The old arguments like SD card slots or One-Seg (mobile TV) support don't matter anymore — is there anything left?
    293AnonymousNov 17, 2026 12:41
    Re: #263
    Price, probably
    You can't get an iPhone for ¥100,000
    296AnonymousNov 17, 2026 12:42
    Re: #263
    It's cheaper
    More expandable

    For casual users with money to spare, iPhone is overwhelmingly easier to use though
    266AnonymousNov 17, 2026 12:37
    At this point you'd just go 'why not Motorola or OPPO instead'
    281AnonymousNov 17, 2026 12:39
    Re: #266
    Got the Motorola g66j 5G
    bought it last year for ¥25,200
    downside is security support ends next August
    292AnonymousNov 17, 2026 12:41
    Re: #266
    If you're fine with a Chinese manufacturer, that works
    291AnonymousNov 17, 2026 12:41
    Strike with the blade of fury, shine through the ages — GEO~! (a line from GEO's old rental-shop TV jingle)

    302AnonymousNov 17, 2026 12:42
    I don't even want great camera performance on my phone,
    so isn't it kind of garbage that makers force in high-end cameras just to jack up the price?
    I'd honestly love a phone with no camera at all
    and just let people attach an external one when they need it
    it'd be better for security too, and keep the price down — I bet plenty of people would buy that
    315AnonymousNov 17, 2026 12:45
    Re: #302
    Here you go
    https://asset.watch.impress.co.jp/img/ktw/docs/614/100/q04_s.jpg

    Background and key points of debate

    Real-world smartphone prices have kept climbing in recent years due to component costs and exchange rates, with even mainstream-tier models now commonly landing around ¥100,000. In the thread, many commenters said they’d stuck with older-generation devices like the AQUOS sense10 or wish4 for a long time, and there was broad agreement that 6–8GB of memory is plenty if you don’t game or watch video. On the other hand, cheap Chinese phones and Motorola devices were flagged for their short security-support windows, showing that price alone doesn’t tell the whole story. The point that rising camera performance is partly driving up prices also highlights a gap between where smartphones are evolving and what users actually need.

    *This article is excerpted and summarized from the 5ch (Breaking News+) thread “[Smartphone] Sharp announces “AQUOS sense11,” priced from ¥94,600.”

  • Miyamoto Karin’s AI Projects Get a Lukewarm Response — 5ch: ‘We’re Just Not Interested Anymore’

    Miyamoto Karin, who continues her activities as a Hello! Project OG (graduate), has been posting videos like her “Solo Trip to Osaka,” where she lets AI decide her destinations — but the thread on 5ch’s Hello! Project board didn’t generate much buzz. Posters were split on whether it was the AI-driven format itself that failed to spark interest, or whether interest in Karin-chan herself has simply faded. Some praised her planning skills as PM/PdM-like, while others coldly dismissed it as no different from a “shock-value creator” (驚き屋, someone who chases cheap “wow” reactions for attention), with some pointing out that existing fans and new fans are looking for different things.

    1AnonymousNov 17, 2026 07:23
    Why though?
    She's doing something impressive, isn't she?
    2AnonymousNov 17, 2026 07:27
    It's not that we're uninterested in the "I did XX with AI" stuff —
    we're just not interested in Karin-chan anymore.
    3AnonymousNov 17, 2026 07:27
    Miyamoto Karin-chan's "I did XX!" ← Isn't it just that people don't show much interest in this?
    5AnonymousNov 17, 2026 07:32
    It's not really that impressive anyway….
    6AnonymousNov 17, 2026 07:35
    Outdoor shoots have a lot of restrictions and get boring, but the location shoot inside the Hello! Shop was fun
    7AnonymousNov 17, 2026 07:40
    [Solo Trip to Osaka] Letting AI Decide Where to Go! Touring Osaka While Filling Out a Bingo Card [Miyamoto Karin] – YouTube
    https://www.youtube.com/shorts/czUu4AQ-cXw
    9AnonymousNov 17, 2026 07:42
    Miyamoto Karin-chan: "I used AI to start a thread on 5ch!"
    15AnonymousNov 17, 2026 08:13
    Honestly, I think she's genuinely outstanding as a PM/PdM
    Not something just anyone can do
    Very few people could put out a product in front of an audience on that short a turnaround
    17AnonymousNov 17, 2026 08:19
    Seriously, what's with all this AI obsession lately
    You could just pay for it yourself and do whatever you want
    18AnonymousNov 17, 2026 08:20
    She's just turning something she's dependent on into her job
    19AnonymousNov 17, 2026 08:23
    Re: #1
    Sure, people praised it out of politeness at first,
    but honestly, it's no different from those "shock-value" creators (驚き屋, YouTubers who chase cheap "wow" reactions for clicks)
    20AnonymousNov 17, 2026 08:23
    Try making a song and music video with AI too
    25AnonymousNov 17, 2026 08:31
    Re: #20
    I think that would actually be best
    We saw with INMU KING that if it turns out good and she waives the profits,
    using AI isn't a problem
    21AnonymousNov 17, 2026 08:25
    It's normal for talent/idol projects to not last, but the sheer "yeah, this probably won't last either" vibe Karin-chan gives off is next level
    22AnonymousNov 17, 2026 08:26
    It's not like she's jumping on every trend and turning into some Reiwa-era Yaguchi (a nod to Mari Yaguchi, a Morning Musume OG known for chasing many side ventures)
    she's supposed to be carefully choosing projects that actually suit her
    24AnonymousNov 17, 2026 08:28
    She's just having middle-aged guys teach her all this, you know
    26AnonymousNov 17, 2026 08:32
    I just want to see "another set of values" different from what the majority holds
    27AnonymousNov 17, 2026 08:33
    No one like that is actually a Hello! Project otaku lol
    28AnonymousNov 17, 2026 08:36
    I thought it was pretty fun
    Made me wonder, could I mess around like this too?
    30AnonymousNov 17, 2026 08:41
    I mess around with AI too, so it feels relatable
    31AnonymousNov 17, 2026 08:41
    This is what a real influencer looks like
    32AnonymousNov 17, 2026 08:52
    What kind of things is Karin-chan doing with AI?
    I'd like to use it as a reference
    I know she built a livestreaming system
    34AnonymousNov 17, 2026 08:54
    She's just running away from what she should actually be doing
    36AnonymousNov 17, 2026 08:57
    Getting to hear the out-of-touch pontificating of old-timers a lap behind the times is part of the charm of Ookami (nickname for the 5ch Hello! Project discussion board)
    37AnonymousNov 17, 2026 08:59
    I use AI at work too, but being shown this… I mean, so what?
    38AnonymousNov 17, 2026 09:05
    She's such a charming person, but seeing this just makes me go "…oh, I see" with an awkward laugh
    39AnonymousNov 17, 2026 09:06
    I respect OGs (Hello! Project graduates) who broaden their own range of work on their own initiative
    40AnonymousNov 17, 2026 09:09
    She has a talent for broadening her range of work while shrinking her fanbase
    41AnonymousNov 17, 2026 09:11
    A rare person with god-given talent who still manages to draw the short straw 100 times out of 100
    42AnonymousNov 17, 2026 09:13
    Among everyone except Airi, who's pretty much all washed up, I'd say she's doing fairly well
    43AnonymousNov 17, 2026 09:13
    A genius center/ace idol cyborg, yet somehow also more unlucky than the average person — that's our Karin-chan for you
    44AnonymousNov 17, 2026 09:17
    Existing fans aren't really asking Karin-chan for that kind of thing, but hey, if it brings in new fans, isn't that a good thing?
    46AnonymousNov 17, 2026 09:26
    It's just that more existing fans don't want it
    If you're into this kind of thing, go ahead and support her properly
    48AnonymousNov 17, 2026 09:28
    Things never actually get better just because otaku say "no to this, no to that,"
    so regardless of Karin-chan specifically, I want OGs in general to keep broadening their horizons and doing their best
    61AnonymousNov 17, 2026 10:05
    Re: #48
    I thought people had learned their lesson after the backlash over Hayashi's glasses turned into a huge success, but here come the old-timers running their mouths about it all over again

    Background and Key Points of This Topic

    Miyamoto Karin is a former member of Juice=Juice who now works solo as a Hello! Project OG, and she has a track record of voluntarily sharing IT-related projects — like travel videos and building her own livestreaming system — on her YouTube channel. The “AI-decided solo trip to Osaka” that became a talking point this time is an extension of that same trend; rather than a brand-new attempt, it’s best understood as part of her ongoing activities, which helps explain the thread’s lukewarm reaction. Where opinions in the thread split was over whether “there’s simply no interest in AI as a method” or “interest in Karin-chan personally is fading” — and posters never reached a consensus on which was the main factor. It’s also notable that comments praising her planning ability in PM/PdM terms stood in contrast to those growing cold on it as mere “shock-value” posturing for novelty’s sake. Note that actual view counts or the scale of the response weren’t mentioned in the thread itself — this is purely a discussion based on 5ch posters’ subjective impressions.

    ※This article is excerpted and summarized from the 5ch (Hello! Project) thread “Miyamoto Karin-chan: ‘I did XX with AI!’ ← The reason you guys don’t show much interest in this.”

  • Possible Human Bones Found in Abandoned Kitakyushu Tunnel — 5ch Asks: ‘Wait, a 19-Year-Old Counts as a Minor?’

    In the early hours of the 16th, a 19-year-old man exploring the Mekari Tunnel — used by a sightseeing train in Moji Ward, Kitakyushu — on a dare (“kimodameshi,” a test-of-courage outing to a spooky spot) discovered what looked like human bones and called Japan’s emergency line, 110. The tunnel is off-limits to the public, and police are now working to determine whether the bones are in fact human. On 5ch, posters traded theories about whether the bones belonged to a person or an animal based on their size and shape, with some explaining that a flat, triangular shoulder blade is a feature unique to humans and apes, while others noted that without a skull, identifying the person would be difficult. Meanwhile, the news report’s description of the 19-year-old as a “shōnen” (a legal term for “minor,” technically anyone under 20) sparked its own debate.

    In the early hours of the 16th, two men on a dare discovered what appeared to be human bones inside a tunnel used by a sightseeing train in Moji Ward, Kitakyushu.

    “This railway tunnel is off-limits to the public. We’re told that what looks like human bone was found inside it.” (RKB reporter Yasushi Wakamatsu)

    At around 12:30 a.m. on the 16th, a 19-year-old man called 110 from the Mekari Tunnel in Moji Ward, Kitakyushu, saying, “There’s something here that looks like human bone.”

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

    3AnonymousNov 17, 2026 00:00
    Two guys doing a dare-visit to a tunnel together — now that's real friendship.
    78AnonymousNov 17, 2026 01:06
    Re: #3
    Two guys who get along, alone in a dark place nobody visits…
    Of course something was bound to happen…
    135AnonymousNov 17, 2026 03:47
    Re: #3

    I went exploring Kamikuishiki Village with a friend once too (the village infamous as the former Aum Shinrikyo cult compound).

    That's less a "dare" and more an expedition.
    10AnonymousNov 17, 2026 00:04
    A 19-year-old… 'minor'?
    24AnonymousNov 17, 2026 00:09
    Re: #10
    The Juvenile Act still covers anyone under 20, so he technically counts.
    15AnonymousNov 17, 2026 00:06
    Human bones are scattered all over the hills around here — par for the course.
    Back in the Edo period, people buried bodies all over the mountains with basically no regulation, so rain erosion washes them out over time.
    Plus stray dogs dig them up too.
    21AnonymousNov 17, 2026 00:09
    Re: #15
    Stray dogs aren't digging up Edo-period bones though.
    62AnonymousNov 17, 2026 00:44
    Re: #15
    Stray dogs, in this day and age…?
    That's odd.
    31AnonymousNov 17, 2026 00:13
    Did someone dump a body in a restricted area?
    This smells like a crime.
    151AnonymousNov 17, 2026 04:36
    Re: #31
    Couldn't it be the other way around — maybe it got restricted BECAUSE it was being used to dump bodies?
    43AnonymousNov 17, 2026 00:20
    Stand By Me didn't do much for me when I caught it on Friday Road Show (a long-running Friday night movie slot) as a kid, but watching it again as an adult, it's painfully dull.
    Getting through to the end was a chore.
    205AnonymousNov 17, 2026 07:35
    Re: #43
    That movie is really just about admiring River Phoenix's good looks.
    47AnonymousNov 17, 2026 00:21
    Re: #1
    So they found what looked like human bones inside this abandoned tunnel?
    51AnonymousNov 17, 2026 00:29
    Doing a dare-walk at this time of year? Guess Kyushu must still be warm.
    86AnonymousNov 17, 2026 01:16
    Re: #51
    Not cold yet.
    Apparently it's supposed to turn cold starting this weekend.
    60AnonymousNov 17, 2026 00:42
    I used to run a ramen shop, and now that I think about it… when I closed up, I have a nagging feeling I might've dumped some pork bones (tonkotsu) in a tunnel somewhere.
    165AnonymousNov 17, 2026 05:51
    Re: #60
    When I volunteered after the Great East Japan Earthquake,
    they told us that even if bones turned up, there was a chance they were just ramen-shop pork bones.
    90AnonymousNov 17, 2026 01:22
    40cm long — which bone would that even be?
    If it's a femur, that's a pretty big animal.
    133AnonymousNov 17, 2026 03:44
    Re: #90
    Plenty of deer out there outweigh me.
    Carrying one down the mountain on my back was no joke.
    91AnonymousNov 17, 2026 01:22
    I heard that with human bones, you can't identify the person unless you have the skull.
    180AnonymousNov 17, 2026 06:31
    Re: #91
    You should be able to identify someone from DNA in bone marrow cells.
    With a skull, though, you've got dental records too,
    and can reconstruct the face to some degree,
    which narrows things down way faster.
    97AnonymousNov 17, 2026 01:37
    Re: #1
    It's probably just someone's prank — brought some pork bones and set them up to look convincing.
    I mean, it's Kyushu, after all 🐖🍜
    105AnonymousNov 17, 2026 01:57
    This only becomes real news if it actually turns out to be human bone.
    Right now all we've got is some kid who snuck into a restricted area, freaked himself out, and called it in thinking something was human bone — basically a nuisance call.
    113AnonymousNov 17, 2026 02:31
    Re: #105
    The fact that it's being reported at all probably means police already think there's a good chance it's human — they just haven't finished the formal examination yet.
    115AnonymousNov 17, 2026 02:34
    If it's really human bone, wouldn't there be clothing or something nearby too?
    Could've been taken by a bear, but a bear would usually drag a fresh kill back to its den — if bones were left behind, it's probably a deer or something.
    134AnonymousNov 17, 2026 03:47
    Re: #115
    There are no wild bears in Kyushu.
    117AnonymousNov 17, 2026 02:37
    They found one bone and are already calling it human? Bold.
    132AnonymousNov 17, 2026 03:30
    Re: #117
    Actually, think about it — which bone looks obviously human at a glance?

    The shoulder blade.
    A flat, thin, triangular shoulder blade is pretty much unique to humans and apes.
    At adult size, you're basically looking at humans, gorillas, or orangutans.
    143AnonymousNov 17, 2026 04:15
    Anyone with a normal head on their shoulders would suspect a bear and bolt immediately.
    148AnonymousNov 17, 2026 04:28
    Re: #143
    It's Kyushu — there aren't any bears.

    Background and Key Points of This Story

    The discoverer described as a “shōnen” (minor) in this story is 19 years old. While Japan’s 2022 Civil Code reform lowered the age of adulthood to 18, the Juvenile Act’s definition of “shōnen” still covers anyone under 20 — which is why the mismatched terminology caught posters’ attention. At the time of reporting, it hadn’t yet been confirmed whether the bones were human or animal; without a telltale feature like a shoulder blade, which closely resembles those of humans and apes, on-site identification is difficult. A definitive determination only comes after DNA testing and morphological analysis of the bone. Many posters pointed out the bones could well belong to a wild animal like a deer or boar — a more measured take that the headline alone doesn’t really convey.

    *This article is excerpted and summarized from the 5ch (News Speed+) thread “[Kitakyushu] ‘There’s something that looks like human bone’ — two minors who snuck into a restricted tunnel on a dare made the discovery; police investigating whether it’s human“.