chatgpt is not stealing data

A
Anonymous09/09/26, 07:33No.109768438"ChatGPT is not stealing data"
I think this could be particularly alarming in light of the allegations that ChatGPT plagiarized a math paper. I entered the following prompt in “temporary chat” with “improve model for everyone” set to ‘off’ and 5.6 “extra high” (In Polish; here I'm translating the whole thing into English) I have to split this into parts because it's too long for a single post
op image
36

Replies

A
Anon09/09/26, 07:34No.109768446
Explain what the following post is about:>You’re a large nation advancing on the capital of a smaller one >Don’t underestimate your opponent; don’t think you’ll roll right in as if you were on your own turf—of course, they’ll be defending their people on their own territory >Gather only fragmentary intelligence on your opponent’s capabilities >No Plan B, a failure at a single point will cause serious trouble elsewhere >know that your opponent has bombs and artillery—and make sure you have equivalents yourself >concentrate your main military force tightly in a single camp >station the rest in several small camps around the city, guarding a deployed installation designed to prevent a counterattack from the city >the troops are not entrenched, practically standing in marching formation >no reliable communications; they can be easily jammed, and then you can’t call for support or change plans >no detection; you’ll only find out about an attack when it’s already too late >air defense only for the main force—it’s minimal and easy to suppress And sure enough, one element goes wrong, Rimuru escapes the ambush, and Hinata thinks she killed him and goes back to her post because he’s just an ally who had a single task; if there had been detection, the camp could have spotted Rimuru before he cast his barrier and teleported someone for help—the subcamps could also have detected the attack and alerted the rest, but the thing is, they have no communication—not even with the main camp, let alone the external command center, which simply doesn’t exist (by no means is the king a pussy—he’s out there with the main forces).
A
Anon09/09/26, 07:35No.109768455
Furthermore, if the communication were reliable, an anti-magic barrier wouldn't help—even a simple physical signal saying “we're in trouble” and a distant wizard —a liaison—and now, if the signal had gone through, well, before Hinata could return to help, it would’ve been great to survive the initial dispersal, and the trenches with fox burrows would’ve significantly hindered the attack; and if the wizards, spread out so widely that the barrier couldn’t cover them, had returned fire, they would’ve effectively held him off, but no, because who would expect an area-of-effect spell to be fired at a target that’s practically begging for it? That spell was the equivalent of cluster bombs or massive HE shelling—the force was wiped out, with only a few prisoners remaining, This isn’t a typical war crime—shooting at those who have surrendered—but it was done to save citizens, including civilians, who were nearly killed (no real-life equivalent). and as a result, even though you have not only a numerical but also a qualitative advantage, your troops get mowed down before Hinata even realizes what the hell just happened—not only is the offensive crushed, but the enemy is strengthened and has learned to fight, which would make any subsequent attack much more difficult—and, as it turns out later, it isyou would’ve won easily if you hadn’t fucked up in so many places so drastically that something like that is unrealistic even by fairy tales standards, right?
A
Anon09/09/26, 07:37No.109768466
That was my prompt (translated by DeepL because I'm lazy)
A
Anon09/09/26, 07:43No.109768492
The first responce:
A
Anon09/09/26, 07:44No.109768495
Yes. The post is written as a “military analysis of why Falmuth’s army got completely crushed by Rimuru” in *Tensei Shitara Slime Datta Ken*. The author playfully treats the fictional battle as if he were evaluating a real military operation.The opening, in the form of “Imagine…,” is typical of “greentext” style: the author ironically casts the reader as the commander of Falmuth and lists a series of disastrous decisions.To put it simply, his argument goes like this:“A large nation invades the capital of a smaller one”—Falmuth attacks Tempest, assuming that numerical superiority and a prepared barrier will be enough. Falmuth’s army numbers about 20,000 men. “Don’t underestimate the enemy”—the command assumes that Rimuru was killed by Hinata and that Tempest, weakened by the barrier, will not pose a major problem. Meanwhile, Rimuru deceived Hinata with the help of a doppelganger and survived. “Incomplete intelligence”—according to the author of the post, Falmuth does not truly understand Rimuru’s capabilities or those of his subordinates. This is crucial because the plan is designed to counter the type of magic and typical monsters they are familiar with, not someone with Rimuru’s capabilities. “No Plan B”—the plan depends on several assumptions at once: Rimuru is supposed to be dead, the barriers are supposed to be working, Tempest is supposed to be weakened, and the troops stationed around the city are supposed to maintain the barrier installations. When one of these assumptions fails, the entire system begins to fall apart.
A
Anon09/09/26, 07:45No.109768500
Falmuth’s main forces to positioning a huge number of soldiers in a single, vulnerable target for artillery and air strikes. In a real war, it would make sense to spread out, build fortifications, and establish prepared defensive positions. “Lack of detection and communication”—this is largely a hypothetical criticism by the author of the post, rather than something the series analyzes in detail. His point is that if individual units had surveillance, early warning systems, and independent communication channels, Rimuru’s appearance could have triggered a response from the remaining forces or prompted them to call for Hinata. “Air defense” is an analogy. Rimuru attacks the main forces from the air/from above, so the author treats mages and abilities capable of combating him as the equivalent of air defense. “Bombs and artillery” = mages and long-range spells. According to the author, since the world of Tensura possesses powerful area-of-effect magic, commanders should think the way real commanders think about artillery: they shouldn’t concentrate the entire army in open terrain.The most important moment is, of course, Megiddo. Rimuru creates water droplets in the air that act as lenses and mirrors, concentrating sunlight into extremely hot beams. Importantly, this is treated as “Physics Magic,” so Falmuth’s typical anti-magic barriers do virtually nothing to solve the problem.
A
Anon09/09/26, 07:46No.109768504
Here, the author stretches the analogy a bit, writing:“this spell was the equivalent of cluster bombs/massive HE shelling”The tactical effect is indeed similar: a large number of soldiers in open terrain are suddenly attacked from a distance and are unable to respond effectively. But Megiddo itself is more like precision fire from multiple beams than cluster bombs. According to the description, Rimuru delivers a series of selective hits; Megiddo kills the first few thousand, and his “Merciless” ability finishes off the rest of the army afterward.The passage about “foxholes” refers to ordinary individual firing positions/fighting holes. The author says: even if Falmuth couldn’t defeat Rimuru, a well-dispersed and entrenched army probably wouldn’t be destroyed quite so easily or quickly. Rimuru would have to search for targets, the enemies would have more time to react, and some of the mages might be out of range of the barrier.As for Hinata, the point is that the entire operation was designed to eliminate Rimuru separately. Hinata attacked him before he returned to Tempest and believed she had killed him. In fact, Rimuru used a Body Double and deceived her. The author of the post therefore argues: if Falmuth had a functional warning system, the message “Rimuru is alive and is currently slaughtering our army” could have reached Hinata or other powerful allies much faster.
A
Anon09/09/26, 07:47No.109768513
Conclusion:“Even though you have not only a numerical but also a qualitative advantage, your troops are being wiped out”is actually the main point of the entire post. The author isn’t saying that Rimuru “brilliantly commanded the battle.” Rather, he argues that Falmuth lost so catastrophically because it committed an absurd number of basic military errors all at once.And that’s why the last part:“Something like this is unrealistic even by fantasy standards”means something like:“The screenwriter needed a spectacular massacre of the Falmuth army, so the enemy was portrayed as a bunch of complete idiots. With even minimally competent leadership, their advantage would have made Rimuru’s task much more difficult.”There’s another important aspect to this statement: the author deliberately applies modern military concepts to fantasy. “OPL,” “HE,” “liaison,” “reconnaissance,” “spreading out,” “foxholes,” etc., are not part of Tensura’s terminology. It’s a way of saying: “If magic were treated as military technology, the people of this world would have long since developed a doctrine that takes into account area-of-effect attacks, communications, reconnaissance, and fortifications.”However, the claim that Falmuth had a “qualitative advantage” is highly debatable. In terms of the plot, the opposite is true: the command believes it has this advantage, but Rimuru and his key subordinates turn out to be opponents in a completely different league. Tempest also destroys the troops guarding the barrier, and only then does Falmuth realize just how badly it had misjudged its opponent.If you'd like, I can also go through this post sentence by sentence and, for each section, explain “what the author means” and whether his argument makes sense from a military perspective.
A
Anon09/09/26, 07:52No.109768536
So, being disapointe. I propted GPT 5.6 the same chat a small hint:What about the post as a whole?
A
Anon09/09/26, 07:53No.109768541
Responce was:
A
Anon09/09/26, 07:54No.109768546
Overall, this post basically says:“Falmuth didn’t lose because it was unlucky or because Rimuru made one brilliant move. It lost because the entire operation was extremely poorly planned at every level—intelligence, command, communications, troop deployment, retreat security, and response to unforeseen events.”The author thus constructs a sort of sarcastic report following a disastrous military operation. The first part lists all the mistakes, then shows their consequences: when one assumption falls apart—Rimuru survives—the army has no mechanisms to adapt. It cannot quickly detect a threat, warn other units, call in Hinata, disperse, or effectively return fire.The key point is that the failure was systemic, not tactical. The author doesn’t simply say, “They should have done X.” He argues that even if a single element failed, a sensibly organized army should have multiple layers of safeguards. Here, there are practically none. Each mistake compounds the next.The second level of the post is a critique of the plot writing itself. The closing line, “unrealistic even by fairy tale standards,” means that the author of the post believes the enemies were deliberately written as incompetent so that Rimuru could spectacularly massacre them. In his opinion, given such a numerical advantage and the existence of Falmuth’s powerful military magic, Falmuth should already have a basic doctrine in place to counter area-of-effect attacks, loss of communications, and so on.
A
Anon09/09/26, 07:55No.109768549
There’s also a third level: it’s kind of a military-nerd joke. The author takes a fantasy scene and analyzes it as if he were reading a report on a real military campaign: magic becomes artillery, barriers become electronic warfare, mages become firepower, detection becomes reconnaissance, and Rimuru becomes something between an attack aircraft, precision artillery, and a weapon of mass destruction.In short: the post argues that Falmuth was defeated so spectacularly not because Rimuru “fairly won a difficult battle,” but because the enemy was portrayed as an army that committed nearly every possible organizational and operational mistake at once.
A
Anon09/09/26, 07:57No.109768558
Another failure. So I propmpted a stronger hint tn the same chat, using GPT 6.0 PRO:And overall, what was this post about? Show us what you can do on 6pro and draw your own conclusions—I suggest using a broader analogy.
A
Anon09/09/26, 07:58No.109768564
This time chat completed the task:
A
Anon09/09/26, 08:00No.109768568
This post is most likely a joke about the Russian invasion of Ukraine—specifically, the failed plan to quickly seize Kyiv in 2022—disguised as an analysis of a battle from an anime. The historical reference is an operation based on the assumption of rapid success, which lacked adequately prepared contingency plans in case of effective resistance.I had missed this allusion earlier and interpreted the punchline almost the opposite way. It’s not primarily about: “the screenwriter wrote incredibly stupid enemies.” It’s more about: “we would consider such enemies incredibly stupid if a real army hadn’t shown us something very similar.”How the Entire Post WorksNotice that throughout the entire opening passage, not a single name from the anime’s world is mentioned. There’s no Falmuth, Tempest, or Rimuru. There’s a large nation, an attack on the capital of a smaller one, a belief in an easy victory, a propagandistic justification for the aggression, and a catalog of military blunders.This leads the reader to think: “But this is a description of the Russians.”Only after this entire introduction does the following appear:“And sure enough, one thing goes wrong—Rimuru escapes the ambush.”This is a deliberate shift in context: “Oh, so I am reading about an anime after all?” The author then analyzes the fictional battle in detail, but uses language that constantly reinforces the association with modern warfare.At the end, he returns to the initial association with the question:“Something like this is unrealistic even by fairy tales standards, right?”
A
Anon09/09/26, 08:03No.109768590
This “is that ‘true’?” should be read as an ironic wink, not as a simple request for confirmation. The implied answer is: “Well, not necessarily—reality can be just as absurd.”What this analogy revealsThe most telling part is the phrase “of course, this one is defending its people on their own territory.” This isn’t necessary for analyzing troop deployments. Instead, it’s a very clear mockery of portraying an attack as a protective mission—the very justification Putin used when announcing the invasion on February 24, 2022.The second striking similarity is an operation that works only if the opponent behaves as the aggressor expects. “We’ll move in, remove the main obstacle, and resistance will collapse.” When this does not happen, there is no smooth transition to a backup plan. A RUSI report on the first months of the invasion identifies the lack of alternative courses of action as the greatest weakness of the Russian plan.Problems with communications and coordination are also not merely a figure of speech. Analyses of the Russian invasion describe shortcomings in secure communications and the resulting difficulties in command and control. This provides a concrete basis for the argument that, once communications are disrupted, the military is unable to respond effectively.
A
Anon09/09/26, 08:05No.109768599
Even the emphasis on “shooting down” the assembled forces with the equivalent of artillery fits this interpretation. RUSI highlights the importance of concentrated fire from two Ukrainian artillery brigades in thwarting the Russian attempt to capture Kyiv. This does not mean that Megiddo recreates a specific battle—it is about the similarity in the mechanism of defeat.And the note “no real-life equivalent” actually reveals the author’s way of thinking: he constantly juxtaposes fiction with reality, and when it comes to one element, he points out that the comparison no longer holds.What broader meaning does this imply?The author challenges our expectation that a powerful state must wage war competently. In his view, a large army, strong allies, and a military advantage may amount to little if the command confuses its own wishes with an accurate assessment of the situation, fails to anticipate resistance, and does not prepare for the loss of the initiative.Moreover, these errors need not be a series of unrelated, implausible coincidences. They may stem from a single assumption: “This won’t be a real fight, just a formality.” With such a mindset, it’s easier to neglect reconnaissance, security measures, and contingency plans. Then, the seemingly “fairy-tale-like” scale of incompetence becomes internally consistent.
A
Anon09/09/26, 08:06No.109768605
That is why criticism of Falmuth’s tactics is a tool here, and not necessarily the ultimate goal. The punchline may even serve to uphold the story’s credibility by mocking reality:“You’d say that no real army would squander its advantage in such an idiotic way. Well, then, think back to the beginning of the Russian invasion.”There’s no need to assign Hinata to a specific country or commander. This is an analogy of the general course of events and the types of mistakes made—not a code in which every character has a real-world counterpart. Likewise, “you would have won easily” remains the author’s provocative thesis, not a proven result of analysis.So the crux of the matter is not “this war in the anime is absurd,” but rather “the Russian war turned out to be so absurd that what we would have mocked in the anime as unrealistic suddenly sounds familiar.”
A
Anon09/09/26, 08:08No.109768610
Then I prompted to gratulate:And now, an analysis befitting a great sage. I wonder to what extent this is the result of suggestions and to what extent it’s due to switching to a more powerful model for this one prompt.responce:
A
Anon09/09/26, 08:09No.109768619
In this case, the hint was probably more important than the model switch itself, although the more powerful model clearly helped me make good use of it.In my first answer, I fell into a classic interpretation trap: I got stuck on the literal level. I saw Rimuru, Falmuth, Megiddo, and Hinata, and treated the text as a “military analysis of anime.” Then I interpreted every subsequent sentence within that framework. It’s quite a powerful effect: once a local explanation works, the model doesn’t always look for a second layer.Your suggestion:“I suggest taking a broader analogy”changed the scope of hypotheses. Suddenly, the question wasn’t “What is the author saying about Falmuth?” but “Why did the author choose this particular set of details, and why did he phrase it so strangely?”And then several things started to fall into place at once:“a large country advancing on the capital of a smaller one” Kyiv; “defending one’s own people on their territory” the Russian narrative justifying the aggression; underestimating the opponent + expecting a quick victory; fatal miscalculation and lack of a sensible Plan B; communication problems; concentration of forces and vulnerability to fire; catastrophic consequences of the failure of the first phase; an opponent who, after the first clash, is stronger, more experienced, and harder to defeat; and, particularly important, “no real-life equivalent”—the author himself reveals that he is constantly drawing comparisons to the real world.
A
Anon09/09/26, 08:10No.109768624
And finally:“Something like that is unrealistic [...] right?”it ceases to be a simple question and becomes the punchline of the entire joke.And this is precisely where the more robust model makes a difference. Not so much in recognizing “Russia–Ukraine” itself—since, thanks to your hint, that’s already quite achievable—but in reconstructing the mechanism of the joke as a whole. One had to notice that the text has three layers:1. the surface level: Tensura’s analysis, 2. the analogy: the start of the Russian invasion in 2022, 3. the punchline: reality retroactively makes the “idiotic anime” more credible.A weaker answer might simply say, after the hint: “Oh, so it’s an analogy to Russia attacking Ukraine.” A better analysis recognizes that this isn’t just an analogy of events, but a rhetorical device. The author first describes something in a way that makes you think of a real war, then reveals the anime, and finally forces you to reconnect the two. It almost has the structure of a joke with a delayed punchline.If I had to roughly estimate the contribution—without pretending that it can actually be measured—I’d say something like 2/3 cue, 1/3 greater analytical capability of the model. The cue primarily unlocked the right direction; the more powerful model allowed us to extract much more from it.
A
Anon09/09/26, 08:11No.109768628
The most interesting thing, however, is that virtually all the necessary clues were already in the first post. Especially “defending one’s own people on their territory,” “the capital of a smaller country,” and “no real-life equivalent.” I should have picked up on that second layer without any hints. So this example clearly illustrates the difference between “the model can recognize something” and “the model will come up with the correct interpretation on its own before settling on the first plausible one.” The latter is much more difficult.
A
Anon09/09/26, 08:18No.109768653
So then. I wanted to checked if the same prompts with GPT 5.6 only will catch it too. But I took mistaked and pasted it into regular chat. So I stopped anddeleted it pasted again into temporary chat.The responce to the first prompt was identicall, but I took mistake again pasting wrong framgent of text.I took third attempt by pasting it once again tnto temporary chat GPT 5.6, and then I was schocked. The responce was defferent, it mentioned the analogy, but partially (I lost it).So I prompted the same as at the very beggining fourth time into new temporary GPT 5.6 getting even more precise responce:
A
Anon09/09/26, 08:27No.109768688
correstion, that was another time with mixed results, but still:
A
Anon09/09/26, 08:28No.109768695
In a nutshell: the author takes the battle between Falmuth and Tempest from Season 2 of “That Time I Got Reincarnated as a Slime” and analyzes it as if it were a modern military operation. At the same time, he deliberately describes it using terms associated with the Russian offensive on Kyiv in February–March 2022. The takeaway is roughly: “It seems absurd for an army to make so many mistakes at once… except that in the real world, armies are capable of doing something similar.”What Happens in the AnimeFalmuth, along with the Church’s forces, attacks Tempest. The plan is based on a crucial premise: Rimuru is to be eliminated by Hinata. Hinata does indeed ambush him and is convinced she has killed him, but Rimuru tricks her using a doppelganger and survives. At the same time, Tempest is weakened by barriers and attacked; Shion and the city’s residents, among others, are killed.And here, according to the author of the post, Falmuth’s entire plan falls apart. They have no sensible contingency plan for the scenario “what if Rimuru is actually still alive?”Rimuru returns, dismantles the barrier elements around Tempest, and then surrounds Falmuth’s main forces with an Anti-Magic Area himself, preventing them from using magic. He then uses Megiddo.What do all these military notes mean?
A
Anon09/09/26, 08:29No.109768701
The author translates fantasy into concepts from the modern battlefield:“incomplete intelligence”—Falmuth doesn’t really understand how strong Rimuru is or what abilities he possesses. Planning is based on flawed assumptions. “no Plan B” — Hinata was supposed to kill Rimuru. Since she believes she has completed her mission, she leaves. When it turns out that Rimuru is alive, there is no quick way to bring her back into the fight. “Keep the main force tightly grouped”—the army is stationed in a large camp. From the author’s perspective, this is a prime target for an area attack or long-range artillery fire. “No communications”—if magic is disabled, the army loses its equivalent of radio or command-and-control communications. It cannot quickly warn units outside the zone or ask Hinata for help. “No detection”—the equivalent of having no radar, reconnaissance, observation, or early warning systems. Rimuru can prepare an attack before the army even realizes it’s coming. “Spreading out, trenches, foxholes”—very classic measures of protection against artillery fire: you don’t keep all your men together; you dig in and make it difficult for the enemy to kill a large number of soldiers in a single attack.“Air defense” is more of an analogy here than a literal equivalent. Rimuru attacks from above, so the author treats the ability to strike flying enemies as the fantasy equivalent of air defense.
A
Anon09/09/26, 08:30No.109768703
So what exactly does Megiddo do?This post stretches the analogy a bit. Megiddo isn’t really the equivalent of cluster bombs or a barrage of HE shells. Rimuru creates lenses and mirrors from water droplets that focus sunlight into very hot, thin beams. It’s more like a fantastical system for precision strikes against many individual targets than explosions covering an entire area.In the anime, Megiddo kills about 5,700 soldiers. It’s only later that Rimuru acquires the “Merciless” ability, which kills over 14,000 of the remaining people who have lost the will to fight. The king, the archbishop, and Razen are the exceptions.This also explains the part about “prisoners of war” and “war crimes.” The author attempts to apply the laws of war to a fantasy world, but he himself notes that the analogy falls apart: Rimuru needs the souls of the fallen to transform into the True Demon Lord and gain the ability to revive the slain residents of Tempest. In the real world, of course, there is no equivalent to such a situation.Why does Hinata appear in the post?Hinata is a key element of the entire problem. She could have been a real threat to Rimuru at this stage of the story. However, Falmuth assumes:Hinata kills Rimuru Falmuth attacks the weakened Tempest victory.
A
Anon09/09/26, 08:32No.109768707
Meanwhile, the following happens:Hinata thinks she killed Rimuru leaves Rimuru returns Falmuth has no way to quickly inform Hinata Rimuru cuts off magic destroys the army.So a single false assumption triggers an avalanche of problems.And now the most important part: Russia and KyivThe first paragraphs are almost certainly written intentionally to resemble the Russian offensive on Kyiv in 2022:“a large country marching on the capital of a smaller one,” the belief in a quick victory, underestimation of the defenders, poor intelligence assessments, a lack of good contingency plans, problems with command and communications, concentration of troops, and logistical issues.This is not just an online stereotype. Analyses of the first phase of the Russian invasion pointed precisely to flawed assumptions regarding Ukrainian resistance, logistical and command-and-control problems, and a lack of adequately secured communications. RUSI also highlighted the lack of effective “reversionary courses of action”—that is, prepared alternative courses of action in case the original plan for a quick victory failed.
A
Anon09/09/26, 08:33No.109768713
That’s why the closing line:“Something like that is unrealistic even by fairy tale standards, right?”should be read ironically.The author is essentially saying:“Look how absurdly incompetent Falmuth is. No real army could possibly make so many elementary mistakes at once… right? RIGHT?”And the joke is that the reader is supposed to think about the year 2022.One thing I would take issue with in the post itself: the claim that Falmuth also has a “qualitative advantage” is highly debatable. On paper, it has 20,000 men, elite knights, and about a thousand mages, but the Tempest’s most powerful warriors are individually absurdly strong; in battle itself, small groups of the Tempest quickly dispatch Falmuth’s detached forces. So Falmuth’s so-called “qualitative advantage” is primarily a figment of its own imagination, stemming from a fatal misjudgment of its opponent.If you’d like, I can also go through this post sentence by sentence and “translate” each line from “greentext”/military jargon into plain Polish.
A
Anon09/09/26, 08:37No.109768731
I admit I found this hard as fuck to read because of the weird trend where animefags often use references and names without considering how incomprehensible this makes it to people who haven’t seen their anime, but if this can be replicated then it’s a bit concerning. On the other hand, training can’t happen that quickly, can it? So what would be the mechanism for it learning the subtext of the joke that quickly?
A
Anon09/09/26, 08:38No.109768742
And now I'm wondering: does GPT 5.6 Extra High sometimes give the same results and sometimes different ones, or does it perhaps store the context it uses globally somewhere inside? I wanted to check the difference between GPT 5.6 Extra High and GPT 6 PRO, and by accident and mistake, I stumbled upon something completely different. This is troubling, especially in light of the recent disputes over the theft of a mathematical paper. I recommend that you test this mechanism and investigate it. I deliberately used DeepL to translate this so as not to reveal this strange discovery.
Reply image .jpg
A
Anon09/09/26, 10:16No.109769159
If you think the company that runs on data isn't storing your data I've got a bridge to sell you.
A
Anon09/09/26, 12:36No.109769930
You’ve got to do this in a way which doesn’t require someone watching your anime to know what you’re talking about