LukeZaz's recent activity
-
Comment on Summer Eternal presents: Red Rooster, an incendiary RPG in ~games
-
Comment on Summer Eternal presents: Red Rooster, an incendiary RPG in ~games
LukeZaz Link ParentThere's no similar.xyz tags as far as I'm aware, but you've got a point and some threads about Esoteric Ebb got the disco elysium tag, so I've added that to this topic.There's no
similar.xyztags as far as I'm aware, but you've got a point and some threads about Esoteric Ebb got thedisco elysiumtag, so I've added that to this topic. -
Comment on Summer Eternal presents: Red Rooster, an incendiary RPG in ~games
LukeZaz (edited )Link ParentMore details on that were in the Anthology, some of which is covered in this brief video by Jamrock Hobo (2:25). If you plan on going in as blind as possible, probably skip the video, but for what...and I am trying to figure out what sort of "auto-meditation device" disaster could lead to enough harm/death to merit a life sentence. (Or what an "auto-meditation device" is.)
More details on that were in the Anthology, some of which is covered in this brief video by Jamrock Hobo (2:25).
If you plan on going in as blind as possible, probably skip the video, but for what it's worth the spoilers seem to be very low-grade and I doubt the game will hide these details much if at all.
-
Comment on Summer Eternal presents: Red Rooster, an incendiary RPG in ~games
LukeZaz LinkFor context, Summer Eternal is one of the studios formed after the theft of ZA/UM following Disco Elysium's release and subsequent fallout. In my opinion, they and Red Info are the two studios...For context, Summer Eternal is one of the studios formed after the theft of ZA/UM following Disco Elysium's release and subsequent fallout. In my opinion, they and Red Info are the two studios with both the necessary talent and legacy to successfully pull off a spiritual successor, so I've been keeping my eye on them.
I'll copy some of the description from the page, but seeing as Summer Eternal dedicates a lot of time to having a strong aesthetic in everything it makes – an aesthetic markdown obviously cannot adequately replicate – I highly recommend reading the page for yourself.
A PHYSICS-DRIVEN RPG.
Red Rooster is an incendiary, physics-driven RPG combining the best of cerebral, worldbuilding-rich RPGs with the genetics of physics-driven immersive sims.
Play as Redrick Roost, a white-collar inmate thrust into the role of an arson investigator and partnered up with an opiate-dependent mathematician to investigate a mysterious string of arson that all seems to connect to his brother, the city’s fire–chief.
Explore the fractalized “border archipelago” of the tri-partite city of Arta-Tantor in a political alternate-history setting with a strong science fiction undercurrent.
Also worth noting: This reveal has been planned for some time now. A while back, Summer Eternal announced an Anthology book covering both themselves and Red Rooster in much further detail, and the book's release was done with the aim of everyone who pre-ordered it receiving theirs on this exact day regardless of where they lived, so as to do a coordinated project reveal.
-
Summer Eternal presents: Red Rooster, an incendiary RPG
28 votes -
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentI saw the BBC article via other stuff posted here, but thanks, yeah. It's good info. I'm aware and didn't say otherwise. It just remains the case that Elon is a Nazi, and so Twitter is owned by a...I saw the BBC article via other stuff posted here, but thanks, yeah. It's good info.
Whether you (and me) like it or not, it's still used by politicians and journalists worldwide, that doesn't make them Nazis nor does it mean the content isn't worth your time.
I'm aware and didn't say otherwise. It just remains the case that Elon is a Nazi, and so Twitter is owned by a Nazi, and it both platforms and supports Nazi views. Journalists choosing to remain on there is a disgrace and lends the site legitimacy it doesn't deserve, especially when such better alternatives for the same thing so readily exist.
And I don't forgive them for it for the fact that "the audience is there," because a) a not-insignificant portion is either bots or fascists, and b) if they left, the real audience would follow. It's a testament to how terrible the site is that so much of their audience left without them, frankly.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentI replied regarding reliability only because AI was brought up to me. It is not the topic I wanted to discuss. Which is why I only used it as a lead-in to my more important point in both comments...I replied regarding reliability only because AI was brought up to me. It is not the topic I wanted to discuss. Which is why I only used it as a lead-in to my more important point in both comments in which I address it.
The topic I am trying to stick with is the concept of reasonable self-doubt and avoidance of overconfidence. If I start outright focusing on AI's reliability across several comments, that changes. The reply I'd get if I'd just stuck with that argument would've been practically guaranteed to be one trying to convince me that AI is more capable than I'm asserting.
And as I said, if that were to happen, we'd have a debate that is endless and frustrating. I've had it before too many times as it is.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link Parent(Disclaimer: AI is a touchy subject for me so I'll be speaking pretty harshly here, but none of that is aimed at you. Your question is plenty respectful and I've got nothing against you!) If...(Disclaimer: AI is a touchy subject for me so I'll be speaking pretty harshly here, but none of that is aimed at you. Your question is plenty respectful and I've got nothing against you!)
If you're curious if and how much I've used AI previously: I used it rarely years back as matter of testing capabilities.
If you're wondering how much I use it now: None at all, as a matter of principal. My only AI use is against my will on the rare occasion a company forces it on me.
The reasons for this are manyfold. Hell, I've written lists, but if I linked those at this point I'd have to update them to add more. Deskilling, delusions, environmental harm, economic harm, hallucinations, hidden errors, privacy issues, training done via piracy, scraping burdens, RAM prices, it goes on. These would be enough to convince me the technology is being used terribly and should be opposed so as to pop the reckless hype-bubble we're in, such that it can be allowed to be the niche tool it is once everything recovers.
The most crucial ones for me personally though is the harm done to artistic endeavors and the rush for convenience. Silicon Valley is betting the farm on this, the U.S. economy is getting dragged with it, and as a consequence AI is being posited as the solution to every facet of life, because if it isn't then the stock market will explode. So I see people online using AI as a first-resort answer to their problems, no matter how bad of a job it does. I see executives wanting me to get excited that they'll save money firing artists so they can generate worse art instead of making it. And I see people in a rush to make a machine which is best at enabling lies and mass scams, so that my ability to believe things are real is destroyed. At least Photoshop required effort.
As one of my favorite posts put it:
Would you like me to reply to your friend so I don't have to? Would you like me to polish your email so you have no voice? Would you like me to scan this article and then tell you what you should think? Would you like me to be human for you? Would you like to do nothing, be nothing, say nothing, think nothing, feel nothing?
I've seen people who are eager for that. That's why I hate AI.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentIf I wanted to, I could argue that AI isn't this on account of it being too error-prone and hallucinatory to be viable; that you'd be just as likely, if not moreso, to introduce more errors than...If I wanted to, I could argue that AI isn't this on account of it being too error-prone and hallucinatory to be viable; that you'd be just as likely, if not moreso, to introduce more errors than you resolve. The problem with this argument is that it changes the topic to an endless "but how good is AI really?" debate, and I'm not interested in having that.
More to the point is that I question how much layering is going on. AI is often used as a matter of convenience, which is inherently discouraging of double-checking. And that's for the average user. For someone with no doubts about AI to slow them down, I'm not convinced appropriate care is being taken.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentApples-to-oranges comparison. Search engines do not hallucinate and they are less likely to deliver you garbage to begin with. Nevertheless, you still practice good hygiene anyways. By contrast,...Apples-to-oranges comparison. Search engines do not hallucinate and they are less likely to deliver you garbage to begin with.
Nevertheless, you still practice good hygiene anyways. By contrast, I'm not getting the impression that you're very thorough in checking the AI outputs you receive; you seem to be categorically separating yourself from those who've fallen prey to delusions, and much like it's folly to assume yourself immune to propaganda, it's likewise with delusions. Remember that those folks don't usually start off believing insane things — they get there over time, and "asking a factual question" might be exactly what they started off with.
-
Comment on Ukraine struck Yandex data centers, Russian bots went silent—now questions are mounting in ~society
LukeZaz Link ParentThis is unfortunately never true. It is not possible to write without subtext of some kind, because it's an inherent property of how communication happens. The only way to represent something to a...There is no implication in the title.
This is unfortunately never true. It is not possible to write without subtext of some kind, because it's an inherent property of how communication happens.
The only way to represent something to a 100% complete degree is to skip the representation and have the object at hand. Anything else is an abstraction, and fundamentally that's what language is. It's not possible to hold a concept and we can't read someone's mind directly to obtain it for ourselves unaltered; even if we could, there's no promise that differences in qualia wouldn't alter it anyway, simply from it being present in our minds.
So we talk, and when we do, details are left out. And thus we are introduced to the need for implication, so as to fill in the blanks that are left behind — something everyone does differently.
This title was a case where, while you may have had one understanding, many other people had a different one. It's being called clickbait because the consensus is that the most common understanding is a misleading one.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentThe same machine responsible for suicide spirals and delusions of grandeur, and which has a massive reputation for overconfidence and yes-manning? I'm sorry, but you're not immunizing yourself....Much like with catching security bugs, asking an AI for a quick fact check before posting or resharing might help.
The same machine responsible for suicide spirals and delusions of grandeur, and which has a massive reputation for overconfidence and yes-manning? I'm sorry, but you're not immunizing yourself. You're just changing what doors you need to guard, and then assuming they don't need to be guarded.
-
Comment on Ukraine struck Yandex data centers, Russian bots went silent—now questions are mounting in ~society
LukeZaz Link ParentTechnically not proof, since it's theoretically possible that Russia's just playing host to the bots for others and not sponsoring the bots themselves. But I don't find that terribly likely — it's...Zelenskyi is provably correct
Technically not proof, since it's theoretically possible that Russia's just playing host to the bots for others and not sponsoring the bots themselves.
But I don't find that terribly likely — it's Russian servers that were hosting bots for purposes that they've got clear motive for and documented history of pursuing, so it's probably them. I'm just clarifying that it's not dead-to-rights.
My main question right now is how reliable this data is. Most I can find from a cursory search is that it's being maintained by a group called Magnitude Media, and I didn't find much of any information on them. Do you have any further details?
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentYou're not wrong, but it's important to note that the only good way to avoid this trap is to keep your guard up, and with the energy that takes everyone has to find time to let theirs down. None...Most of the people resharing are political types who don’t care about journalistic standards of evidence when it’s something they want to believe.
You're not wrong, but it's important to note that the only good way to avoid this trap is to keep your guard up, and with the energy that takes everyone has to find time to let theirs down. None of us are immune to disinformation either.
Hell, if anything, I think "I'm safe because I'm good at catching it" would be an even more dangerous thought long-term, because it encourages lowering that very guard.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentThis is what I'm hearing from a lot of folks, yeah. I can't evaluate it for myself, because I've stripped algorithms out of my social media usage almost completely and so I effectively never see...This is what I'm hearing from a lot of folks, yeah. I can't evaluate it for myself, because I've stripped algorithms out of my social media usage almost completely and so I effectively never see this kind of thing anyway, but I can definitely say I want it to be true.
It would be a real relief to know for a fact that so much of the bigotry is fake.
-
Comment on Navier-Stokes lost in translation: Why Lean verification of AI autoformalisation does not guarantee correct natural language proofs in ~tech
LukeZaz LinkPosting the paper itself as it's more direct and substantial, but there's also a good Bluesky thread here that seems to summarize the key ideas. I recommend reading the whole thread, but here's...Posting the paper itself as it's more direct and substantial, but there's also a good Bluesky thread here that seems to summarize the key ideas. I recommend reading the whole thread, but here's the key bits:
The actual math in this paper is over the heads of non-mathematicians but the core argument is pretty easy to follow: AI theorem-prover formalizations can't actually be trusted to faithfully be the same argument as the natural-language paper they come attached to
the way openAI slopmath paper spam works is that each paper really comes as two things. one is the actual paper, which is an argument in English intended for humans (and LLMs) to read. the other is a "formalization", source code for a specialized programming language (Lean) that proves theorems.
It's generally accepted that if Lean code compiles, then the theorem that the code implements must be true. But the problem there is that there's no way to actually guarantee that a given pile of Lean code really is the same thing as the NL argument being made in a given paper.
The paper argues that this is literally undecidable: "Informally, providing semantically faithful AI autoformalisation is harder than any computational problem including the Halting problem." The only way to be sure is extensive manual checking.
Some of these Lean proofs are millions of lines long. So what OpenAI is doing is presenting a "result" that may or may not be the same as a huge git repo of Lean files. The paper demonstrates examples of incorrect proofs being "translated" into correct Lean by GPT!
-
Navier-Stokes lost in translation: Why Lean verification of AI autoformalisation does not guarantee correct natural language proofs
25 votes -
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz Link ParentFor what it's worth, I think detailed analysis encouraging critical thinking and emphasizing the importance of not jumping to conclusions is absolutely worth posting. Especially considering we're...So I'm unsure how it's "important enough to create a new topic" since it is effectively nothing itself.
For what it's worth, I think detailed analysis encouraging critical thinking and emphasizing the importance of not jumping to conclusions is absolutely worth posting. Especially considering we're in ~society and not ~news.
I just wish the poster had the sense to put their article on almost any other site whatsoever. I would have trouble taking something seriously as an "article" even before Elon turned Twitter into a Nazi hole.
-
Comment on Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact in ~society
LukeZaz (edited )LinkHere's the post for those who, like me, refuse to click on a Twitter link under any circumstances: Full post (collapsible due to length) Those viral claims about plummeting Russian ‘bot’ activity...Here's the post for those who, like me, refuse to click on a Twitter link under any circumstances:
Full post (collapsible due to length)
Those viral claims about plummeting Russian ‘bot’ activity have no basis in fact
Posted by Ilya Lozovsky on October 12th, 2026
At time of writing: 39 replies, 154 retweets, 668 likes, 56,965 "views"
For two days, X has been deluged by claims of a sharp decline in Russian “bot” activity — supposedly a result of Ukrainian strikes on three Yandex data centers this week.
These claims have been promoted by respected analysts, top politicians, and ordinary Ukraine supporters clearly eager for a win during a difficult week.
Unfortunately, they’re based on data that has been misunderstood or appears to be entirely made up:
- One chart from internet infrastructure company CloudFlare shows a decline in automated web traffic from Yandex. But this is a measure of things like search engine requests, not social media bots.
- Another chart purportedly shows a decline in activity related to Germany’s right-wing AfD party. But it contains obvious errors and is completely unsourced. My view is that it’s entirely fake.
I’ll go into more detail on both cases below. The bigger picture is that the “anti-Russian-influence” field remains vulnerable to large-scale misinformation fueled by motivated reasoning.
This isn’t the first time I’ve seen exaggerated claims about Russian influence go viral. But last night I was truly struck — really, almost radicalized — watching in real time as experts at the heights of their careers unquestioningly boosted emotionally satisfying narratives that were pretty much entirely wrong.
It’s tempting to ascribe this all to people’s ideological commitments, and I’m sure Ukraine’s enemies will waste no time pointing that out. But of course, this kind of thing permeates the right-wing information environment, too.
More illuminating, I think, is the lens of social psychology: People like to have their views confirmed, they like to cheer on their team, and they like to collect likes and retweets. And I don’t mean to descend into cynicism. Ukraine has had a particularly tragic few days, stabbed in the back by the Trump administration even as dozens of civilians were murdered by Russian drones and missiles. I understand that people are hurting and want to celebrate a win. That’s just human nature.
But it's hugely important for those of us who support Ukraine to practice good epistemic hygiene. If for no other reason than because Russian influence activity, both on- and offline, is absolutely a real, documented threat to the democratic world. I’ve done a good amount of reporting and editing on this topic: Consider the ‘Dear Compatriots’ project, this story on Russian interference in Moldova’s EU referendum, or this investigation into real Russian “cognitive strikes” against the West.
These threats are much harder to address when they’re lost in a cacophony of exaggeration. For example, the actual effects of the Ukrainian strikes on Russia’s walled-off digital infrastructure — as reported, for example, here by @thebell_io — are real and important, and have been totally drowned out.
I'll note that the responsibility to cover this issue truthfully and accurately, and not just to engage in wishcasting, is especially incumbent on people in high-level positions with large platforms.
Now, for the details on the specific claims:
CloudFlare
This chart from CloudFlare purportedly shows that Russian “bot activity” has dropped since the Ukrainian strikes. But that technically illiterate interpretation has been thoroughly debunked, initially in this thread by @antibot4navalny, a widely cited group of Russian researchers. “Bots” here refers not to “automated social media accounts” but to any kind of automated requests made to web sites, such as the crawlers that do web searches. It stands to reason that these requests fell after the Yandex centers were struck.
AfD
Most virally of all, Czech MEP Tomáš Zdechovský shared this chart, allegedly showing that activity related to the right-wing German AfD party plummeted after the strikes. The implication is that it was all inauthentic Russian influence.
But the chart has a number of issues:
- It was shared on October 11, but the second marked point appears to fall on a future date, around October 13. The “data” (the blue line) extends all the way to October 15.
- It has obvious errors, like the bottom interval on the Y-axis being noticeably smaller than the rest. No chart creation software would do that, even if you were feeding it fake data. My assumption (which I can’t prove) is that this was AI-generated.
- It’s completely unclear what it’s supposed to be measuring. The Y axis is labeled “Activity per day” — so what is that supposed to mean? 4,500 activity per day?
- Most fundamentally: Any legitimate chart like this would be more professionally labeled and titled, and we would know who made it and what data it was based on. This doesn’t even have a logo. It doesn’t even say AfD on it!
I wrote Zdechovský to ask him about these issues, and where he got this chart. His spokesperson sent a quick and polite reply, but didn’t respond to the latter question. For transparency I’ll paste the entire response below.
Now, all of this doesn’t prove that nothing happened. What it shows is that there’s no proof that anything happened. It’s certainly possible that nefarious Russian bot activity really has been reduced. Tracking this kind of thing is very difficult, and I hope someone does a detailed and rigorous study soon. Until then, I think the widely-repeated anecdotal observations — “wow, social media seems so much nicer today!” — can be dismissed as wishful thinking.
Meanwhile, I'll go back to reading X and grinding my teeth.
Here is the response to my inquiry from Mr. Zdechovský’s spokesperson, reproduced in full. In light of the above, you can decide if you find it convincing:
"In my post, I explicitly used the phrase “If confirmed” to make it clear that the suggested connection was not yet an established fact, but rather an interesting finding that deserves further examination by data analysts. I certainly did not present it as a conclusively proven fact.
At this point, I do not have any additional data available. What we do know for certain is that, according to Cloudflare data, Yandex bot traffic dropped dramatically by 99.6%. Even President Zelenskyy has pointed out that there has been less anti-Ukrainian hate on social media. Is this a coincidence? It is difficult to say. Personally, however, I believe there is a connection.
Regarding your second question, I would like to point out one important inaccuracy. The graph does not actually end on October 15, as you suggest. That date is merely the final label on the horizontal axis, while the plotted data ends before that date. It is therefore incorrect to automatically assume that the graph contains data from the future.
I have been dealing with Russian disinformation operations and foreign interference for many years, including through my work in the European Parliament. Their existence has been extensively documented in numerous reports by European intelligence services and EU institutions. I strongly reject the notion that a single unverified graph should undermine the extensive evidence of Russian influence operations or the need to take decisive action against them.
This is precisely why I am preparing an appeal to the European Commission, urging it to address this issue seriously and decisively. If we continue to allow Russian bots to operate unchecked, we risk losing our democracy."
Personally, I'm staying wary. I've seen folks jump to conclusions a bit too fast, but I'm also hearing a lot of folks tell me they've noticed a change. Hard to say for sure.
All I can say beyond that is that, for all that they appear legitimate, this person would get taken seriously a lot better if they weren't posting on the Nazi website.
-
Comment on Donald Trump strikes deal with Vladimir Putin to get diesel in sharp reversal of US policy weeks before midterms in ~society
LukeZaz Link Parentcc @adutchman In addition to this, it sounds like the diesel being bought is graded for higher temperatures than what Russia will need: So Trump's buying what Russia won't be using anyways. This...cc @adutchman
In addition to this, it sounds like the diesel being bought is graded for higher temperatures than what Russia will need:
But, Seigle said the new deal is a “great benefit to Russia and to Putin,” who are currently trying to offload their summer-grade diesel for the heavier winter and arctic grades that they will need in the coming months.
So Trump's buying what Russia won't be using anyways. This feels like it might be extra foolish since winter is so close, but I've got no idea what the limits on these grades are, so I can't say if it'll matter for U.S. temperatures.
cc @zod000, since they said similar
A lot of folks are expressing similar sentiments, which I do understand; many games that have gone for "unique" or otherwise involved control schemes usually just introduce friction and make the experience worse. In this specific case though, I feel like Summer Eternal deserves the benefit of the doubt.
I don't think it's a guarantee that they'll succeed at making it feel good, or even necessarily super likely since pulling that off is indeed quite hard. It's just that with Disco being as good as it was, I think they've got enough of a shot for it to be worth trusting, at least until we can see what they mean.