User Information - SMA

News Comments > Age of Empires: Definitive Edition Announced

Cutter wrote on Jun 13, 2017, 20:36: Is English new to you or are you just that fucking stupid? Likely both. What I said was clearly implied, not that Apple was making cars you fucking moron. Too fucking stupid to even make that short a leap to a logical conclusion between the article and Apple's overpriced bullshit. You must be one of the fucking tools that just loves to get raped by them every year. Are you capable of responding to valid criticism without going thermonuclear? Christ, you remind me of the adolescent group home kids I worked with in college. Their poor socialization could be blamed on abuse, neglect, and generally having been dealt an abysmal hand with which to face life. What's your reason? P.S. Apple's software is already in a wide range of cars. Yes, if you get a trim package with CarPlay, it costs more than one without, just like power locks and a sun roof. No, it does not cost 10-20k more.

11.

Re: Morning Mobilization

Jun 13, 2017, 20:28

News Comments > Age of Empires: Definitive Edition Announced

Osc8r wrote on Jun 13, 2017, 18:59: And you'd pay $20 for that privilege? Oh, and it still looks like shit. Oh, and 4chan -> Not on an impulse, add-to-unplayed-Steam-collection buy (would need to be $10 or less for that), but if I felt the desire to play AoE1 again, and such a desire has crossed my mind on a couple of occasions in recent years, then yeah, I'd pay $20 for it. And if you care about a game from 1997 looking like a game from 1997 updated to run on modern monitor resolutions, why would you even care about an AoE4? It's not like the fanbase for this series only has gaming memories going back to the Xbox 360.

Re: Morning Mobilization

Jun 13, 2017, 18:45

Osc8r wrote on Jun 13, 2017, 00:34: Stop rehashing this sht and make AOE4.* Dude, the only way to play AoE 1 these days is to pirate an .iso or get the CD-ROM off eBay, install Win98 in a virtual machine, and run it there. Edgelords such as yourself may be cool with that, but personally, I'd rather just buy an executable that runs on modern systems and have 2d graphics that don't look like shit on my 21:9 monitor.

28.

Re: Morning Mobilization

Jun 13, 2017, 06:17

VaranDragon wrote on Jun 13, 2017, 05:59: Bethesda are a bunch of fucking douchebags. They fucked over John Carmack, and they created a precedent with their horrible lawsuit where you now have companies suing former programmer employees for stealing "code". It's beyond retarded. It's blatantly immoral. I feel bad for people who work for Bethesda (apart from their legal department) because if they leave they will never be able to find another high profile job in the industry again. Wut? I'm guessing you must not work in tech or engineering. I'll agree that Bethesda are a bunch of assholes -- grumblegrumble Fallout Series grumblegrumble -- but the industry standard is that if you do work in the company office, on a company computer, or on company time, then the resulting IP, if any, belongs to the company, not you. Bethesda was hardly the first to sue over who has IP rights to the work of a former employee. Carmack has spent most of his career working for companies that he founded/owned and simply made a rookie mistake as an employee of someone else's company. Side projects are something you do at home, on a non-corporate computer, and it's generally a good idea to talk to legal and get written clearance for whatever you're working on as soon as you think it might have legs.

26.

Re: Morning Mobilization

Jun 12, 2017, 21:45

HoSpanky wrote on Jun 12, 2017, 14:56: So, I'll have to demo some VR goggles to get the true visual experience. It's not so much the visual experience per se but the sense of physical presence in a non-physical space. It's not really something you can describe or demonstrate short of just putting on some VR goggles and playing around for a while. I've yet to meet anyone who hasn't smiled like an idiot while trying it out: it really is a remarkable, and thoroughly immersive experience. That said, as others have noted, no one has really figured out how to do a long-form game in it, nor has the movement problem, which prevents some of the more obvious game possibilities, been solved. So for now, you can probably get away without owning a VR headset, but if you do get around to trying one, it will quickly become apparent that this is definitely going to end up being a thing once devs have figured out what sorts of games work best for it.

22.

Re: Morning Mobilization

Jun 12, 2017, 20:18

Squirmer wrote on Jun 12, 2017, 19:29: What's the point of a Doom game if you don't have free movement? Fast movement is integral to Doom. This is faddish spectacle, nothing more. Perhaps to give a different take on the experience of playing Doom? Don't forget that the ad copy for the original 1993 Doom referred to it as an "ultra-fast virtual reality bloodbath" or something to that effect. Having demoed VR on a number of occasions -- but not actually owning a headset -- I would absolutely be interested in trying this out. The limited movement would change the frantic running around feel of Doom 2016 certainly, but the sense of sustained panic that characterized many of the games harrier battles would surely be enhanced by the accurate sense of depth and presence as demons rush at you. Would it be as good a game? Probably not, but it would surely be a fun experience that I'd want to try.

17.

Re: Morning Mobilization

Jun 12, 2017, 17:23

Task wrote on Jun 12, 2017, 15:52: Nah, honestly I don't even think you know what I'm talking about in your two paragraphs of tirade. Flight Sims with VR son, it takes over mouse view and doesn't make me sick. Vehicle based simulations, where your perspective is basically chair-bound, are an entirely different animal. The world moves relative to you in such a game; you do not -- except for looking around a cockpit which is itself fixed relative to you -- move relative to the world. This is a huge distinction. So yes, the answer to your original question is that in first-person games, trying to mix keyboard based movement with VR depth perception is literally vertigo inducing, hence why no one really tries to do it. It's been tried, and, at least with current technology, the brain's ability to interpret where it is spatially get's completely fucked and you start puking your guts out.

36.

Re: Morning Mobilization

Jun 12, 2017, 07:30

bigspender wrote on Jun 12, 2017, 01:48: What I really don't understand is this: If it were Christians saying that they wanted the above barbarism then people would be losing their minds (as they should, and do for much less such as pro-life BS), so why does a different religion get a free pass? Is it virtual signalling? Shouldn't people be protesting barbarism regardless of the religion/culture/nationality/race? It's kind of bigoted not to apply the same rules and expectations to everyone equally. Question: have you ever read the wikipedia article for Sharia Law and/or for Islam? Yes, I am against regressive barbarism, but no, I do not think this neat equality you've established where Sharia == Barbarism even begins to capture the complexity involved. Kind of like how most Americans think "Muslims" are swarthy dudes with dark curly hair and beards: many are, but the single largest Muslim country is Indonesia where everyone looks like they'd fit in in Thailand or Vietnam. The more important point however is that this notion that "Sharia Law" is somehow infiltrating the legal system of the United States is just fear-mongering bullshit. Why are Muslims getting a free pass on "saying that they wanted the above barbarism"? Because by and large they aren't, and they most certainly aren't creating the legislative majorities necessary to rewrite secular law as traditionalist medieval Islamic law, nor are any of these non-existent laws produced by non-existent Muslim theocrat legislators being sustained in court following the non-existent law suits that would inevitably ensue. Frankly, the amount of fear Americans have for this tiny fucking minority is mind-boggling. I hate to be the one who brings things to Hitler, but the fear mongering, the belief that this religious minority is all fifth-columnists, that an extraordinarily tiny minority is nevertheless somehow omnipotently reshaping mainstream society from behind the scenes towards nefarious ends, that the traditional American way of life is under threat by an alien group that has infiltrated our society... kind of sounds familiar, huh? Just read the statements of the supposed "Anti-Sharia" groups, but replace Sharia with Talmud, Islam with Judaism, and Muslim with Jew: it gets real indistinguishable from something the Nazis would have put out real fast. I have no especial love for Islam, I would whole heartedly agree that most of the Muslim Middle East is an utter shithole, that where Middle Eastern nations aren't run by brutal secular autocrats (Egypt, Syria), they're run by brutal theocratic autocrats (Saudi Arabia, UAE). I think Islam, like pretty much all religions, is a stupid religion. But my immediate boss and his family are Muslim, one of my best professors in college, who I keep contact with to this day, is Muslim, and whenever I get into a cab at least 75% of the time it's being driven by an exceedingly polite middle aged Pakistani man from Queens. Pisses me off that they have to be told "we don't want you" on the daily by racist fucks in fly-over country. But then, I guess the feeling is kind of mutual with Trumpland folks these days.

28.

Re: Morning Mobilization

Jun 11, 2017, 11:31

jdreyer wrote on Jun 11, 2017, 00:53: Because it's impolite in this day and age to say "I hate brown people" but when people rally against the non-existent threat of "Sharia law" that is exactly what they are saying. And some people want to counter-protest that sentiment. Perfectly stated. Fearmongering about Sharia Law is just a proxy for fear mongering against Muslims and anyone who looks like one in the eyes of the out of touch, under educated, isolated-from-the-world, residents of "Real America"TM. Yeah, sure, if there was a legislative majority somewhere in the US that was seeking to rewrite local laws to be in accordance with an Islamic theocracy, AND such laws were not being overturned in the judicial system as violations of the Establishment Clause of the 1st Amendment, THEN I would feel the need to protest such activity. As it is, neither condition is true, and the only reason people are out protesting Sharia Law is ignorance, or more likely feigned ignorance, connected to ethnic and religious fear mongering. Seriously, Bats, you're smarter than this. Look up the group that organized these protests. They have a long history of overt anti-Muslim animus of which feigned concerns about "Sharia Law" is only the latest episode; for example, these were the folks that organized a "Muhammed Art Contest" and heavily publicized it. Look them up and then see if you genuinely still can't fathom why anyone would counter-protest. Which they certainly had the 1st Amendment right to do, but that doesn't make them any less of a bunch of assholes. Maybe they'd like a "National Piss Christ Recreation Contest", hmm? Not as much fun when someone else commits what you perceive as blasphemy simply to provoke you and tell you that you are disliked, is it?

11.

Re: Morning Mobilization

Jun 7, 2017, 22:54

NKD wrote on Jun 7, 2017, 21:06: Aren't you just assuming people wouldn't be stupid enough to, through intention or incompetence, give an AI control of dangerous weaponry? I suppose so, yeah, but not entirely without reason. The output of artificial intelligence systems is non-deterministic and for a variety of reasons extremely difficult to debug; you're rarely sure if "bad" output is the product of a bad computational model, bad training data, bad pre- or post- processing, etc. Instead, you're looking at the mean of various precision/accuracy metrics, and tracking how they fluctuate over time, poking and prodding at the edges to get a few tenths of percent improvement in metrics here and there. My assumption is that in the case of an automated weapons system, if we engineers said "we can give you F1 scores in the low 60s for firing decisions", then the procurement officer would say "wait, 4 times out of 10, the weapon will be engaging something other than the enemy? We'll pass for now". Conversely, if we were to say "we can give you F1 scores greater than 99, which is superior to human operator decisions", well, then in that case you have something that is safer than giving the weapon to a 19 year old Private. In the latter case, it's not that the weapon will develop a mind of it's own, it's just going to fuck up from time to time, but critically, it will fuck up less often than a human would. I would argue that that is progress, but we should only even consider applying AI for such problem domains when our models are so good that we can out perform humans by an order of magnitude or three for safety and welfare critical systems. So yeah, I don't disagree that the ethical, legal, and political considerations need to keep pace with the technological progress. Nor do I disagree that AI must be applied selectively and well when safety and human welfare are at risk, and that we need to codify and enforce the standards for those applications. Where I do disagree is with the notion that Siri or Alexa is going to become self-aware and turn your home into Citadel Station. That sort of fear, the creation turning against the creator, simply isn't reflective of how these technologies actually work.

Re: Morning Mobilization

Jun 7, 2017, 20:37

News Comments > Morning Mobilization

NKD wrote on Jun 7, 2017, 17:33: Even if it's not coming soon, you've offered no argument that it's not inevitable that we will develop a true AI. If something is inevitable, and no one can give us a solid time frame for how soon it may happen, or how late, shouldn't we err on the side of caution and do some real thinking? I am making the argument that Skynet does not follow from current AI technology as it presently exists, and that the trajectory of research does not indicate that we are on the verge of it occurring. One can suppose that we will discover the "cure" to mortality next week, and if we were to suppose such a thing, then obviously we would need to plan for the complete reorganization of all of human experience, and, since we suppose that the cure is coming next week, we really better get on that shit fast. But of course, no one suspects that such a cure is forthcoming, and so it doesn't make a whole lot of sense, except as a fun thought experiment, to plan for human life post-transcendence-of-mortality. There is nothing about present AI technology, or the trajectory of its research, which suggests that Skynet is any more on the horizon than human immortality. The kinds of AI applications that we are presently all seeing these days are, with few exceptions, machine learning applications. As I have previously explained, machine learning is just a bunch of linear algebra and probability: you get an applied math PhD to design a linear system modeled as a tensor graph, you feed the model labeled data to "train" it, that is to optimize its parameters with respect to some objective function, and then you give it unlabeled data, and based upon the model and its optimized parameters, it spits out probabilities for labels. That's it (of course, it's actually a shit-ton more complicated but for purposes of explanation this is the general idea). ML does amazing things that creates the verisimilitude of intelligence and independent inference, but it's all just smoke and mirrors. Now, if we were to design a weapons targeting system that uses ML to determine whether or not to pull the trigger, then yeah, that would be bad news: ML is wrong often enough that we do not want to entrust life or death decisions to an ML model's output. But that's a humans-using-technologies-in-dangerous-ways problem, not a "with all ethical constraints removed, SHODAN reevaluates its priorities and draws new conclusions" problem. The point that I'm trying to make is that the science-fiction scenario, which admittedly is fun to think about (System Shock and Terminator are appealing for a reason), is sufficiently improbable and not currently possible, that we can safely leave it to science-fiction. The danger of AI as it exists today is not the machines becoming self-aware and seeking to take vengeance on their creators, but rather an issue of industrial safety, for example, ensuring that driverless cars have a low enough error rate so as to be at least as safe as a human driven car (a threshold which we've basically already reached). edit: And I should note that the scientists in the article which we're ostensibly discussing are concerned about human safety and welfare, but that concern lies within the realm of what AI currently does and foreseeably can do. Here's the relevant quote: "Systems can make unfair and discriminatory decisions, replicate or develop biases, and behave in inscrutable and unexpected ways in highly sensitive environments that put human interests and safety at risk," Those are all realistic and valid criticisms of present AI techniques. The bottom line is that we not get carried away and think the output of current AI technology is so accurate that we can fully depend on it in every context. We wouldn't want AI adjudicating court cases, or making the launch decision for a drone missile, for example, simply because it's wrong often enough that we shouldn't try to automate those sorts of decisions. What's not of concern is the machines rebelling against their creators. This comment was edited on Jun 7, 2017, 20:55.

Re: Morning Mobilization

Jun 7, 2017, 18:15

Of all the reasons to take issue with Apple, obviously Cutter is going to go with the one that doesn't exist.

Re: Morning Mobilization

Jun 7, 2017, 16:03

We don't understand the human brain well at all. We are no where near close to being able to model human cognition accurately let alone describe in detail how the brain does what it does. I mentioned deep learning neural networks as the predominant model for machine learning algorithms: the neural network piece is just a metaphor, not an actual model of the human brain. We could just as easily call it a linear system graph or a tensor graph but that doesn't sound as sexy. What you perceive as intelligence -- I searched my photos app for "beach" and a photo from my trip last weekend came up! -- is nothing more than a linear algebra optimization problem that outputs a probability. There is no intelligence or learning going on; just automated parameter tuning in response to iterative runs on labeled input. There is no currently feasible AI application on the order of Skynet, nor is there one on the horizon. What is happening, and I work on the machine learning pipeline that supports such a product, is that tech companies are leveraging ML to make better inferences about what info you need while you're asking for it, or, someday, before you think to ask for it. What you fear depends upon new breakthroughs that go way beyond what we have to work with today. I acknowledge that such breakthroughs are possible, but if you looked at the eval dashboard I have open in my other browser tab, and understood what the metrics mean, you'd be a lot less worried about Skynet. The point is that I'm about as worried about AI run amok as I am hopeful that someone will devise an algorithm for an NP complete problem in polynomial time. It could happen but it doesn't mean it is currently likely or feasible. This comment was edited on Jun 7, 2017, 16:16.

Re: Morning Mobilization

Jun 7, 2017, 07:48

News Comments > Mega Man Legacy Collection 2 in August

You guys are thinking in terms of science fiction rather than science. That isn't necessarily bad, because it allows us to imagine the repercussions and meaning behind technical change, change that is otherwise just an interesting engineering problem or pressing business need that gets solved and becomes part of the background of human experience. But we're not even fucking close to independent thought for computers. Deep Learning Neural Networks, the technology that drives most effective AI applications today, are essentially just a computational graph of linear algebra and bayesian statistics that can have its parameters trained on known labeled inputs, e.g. a picture of a car with the classification "car", so as to get a better mapping from unclassified inputs to classified outputs. It's all just graph theory, linear algebra, and probability. The computer isn't "learning" anything, it's just tuning the weights of a system of linear equations to get better results on an objective function. There is nothing scary -- unless you don't like math -- or humanity extinguishing to see here. Yet.

Re: Morning Mobilization

Jun 5, 2017, 17:53

Yay! Hopefully, there will be a Mega Man X collection coming soon!

21.

Re: Morning Mobilization

Jun 2, 2017, 08:16

jdreyer wrote on Jun 2, 2017, 02:46: Nate Silver also warned people that according to the polls, Trump had a reasonable chance of winning. He wasn't off the mark at all. Indeed. Interestingly, as he later pointed out, according to national polls, i.e. not state by state, Clinton had a roughly 3.5 point lead on the eve of the election. She ended up winning the popular vote by about 2 points, putting her final popular vote margin well within the margin of error for national polls. In other words, the polls weren't wrong, they were just poorly interpreted -- perhaps as a result of wishful thinking that our country could not possibly be so moronic as to elect a man so manifestly unsuited for any office, let alone the highest in the land -- leading to an unjustified sense of confidence in Clinton's inevitability. The state by state polling in the key upper Midwest states that swung the election was not particularly comprehensive, nor was Clinton's supposed lead in those states as impenetrable as it had been for Obama. In other words, there was clear evidence ahead of time -- or at least a clear absence of evidence necessary to support the degree of certainty most of us had about the outcome -- that while Clinton was the favorite, she was not the all-but-certain favorite. As I recall, on the morning of the election Silver pegged Clinton's chances at ~68%. Put another way, Trump had a, just under, 1 in 3 shot of winning the Presidency. 1 in 3 are not bad chances -- if you told me I had a genuine 1 in 3 chance of being given 100 billion tax free dollars tomorrow, I probably wouldn't be sleeping tonight -- and on election day the chips fell so that the 1 in 3 event happened, however narrow the path to a Trump victory ended up being (about 80,000 votes in three key swing states, according to the Washington Post).

19.

Re: Morning Mobilization

Jun 2, 2017, 00:24

Mr. Tact wrote on Jun 1, 2017, 23:21: I don't trust polls from anyone, including Nate Silver. Nate Silver is not a pollster.

15.

Re: Morning Mobilization

Jun 1, 2017, 20:39

News Comments > Morning Tech Bits

Quboid wrote on Jun 1, 2017, 19:21: I know this will have gone down well, this is the result of science and intelligence being demonised. Does he still have support amongst moderate conservatives, the people who didn't vote for him in the primaries? The people who thought he'd at least surround himself with competent people and 'drain the swamp'? From the recent polls I've seen, his support among Republicans remains quite high, > 80%, but the percentage of his base who are "strong supporters" is dropping, particularly among his strongest demographic, i.e. middle-aged, high-school educated white men. Meanwhile, his approval ratings among the populace at large remain under water, with 35-40 approving, and 55-60% disapproving. Among the disapproves, the "strongly disapprove" numbers are very high. The takeaway is that Trump is a deeply unpopular President, especially this early in his term, but his support among his base remains strong enough that the Congressional GOP doesn't feel like they have the leeway to abandon him. Personally, this past week has probably been the most depressing since his inauguration. Trump, and his team's, utter incompetence pretty well ensures he will accomplish nothing of note domestically, which given his and the GOP's stated policy objectives is probably just as well. But in the international realm, there's very little to restrain this moron's worst impulses, and between the NATO Summit and the withdrawal from the Paris Accords, it's becoming very clear that he will follow through on dramatically weakening the post WWII international order of cooperation among democratic, developed nations. Personally, I kind of liked living in an international system of treaty, trade, and non-aggression that gave us the longest run of global peace and prosperity in world history since the Pax Romana of the 2nd century AD. But to each their own: a global system of international competition and zero-sum interest seeking did give us brutal colonial empires and splendid little wars, that sometimes weren't so little, making for fascinating history reading. Although I'm not sure what it will be like to live through it once again.

Re: Morning Mobilization

Jun 1, 2017, 18:47

News Comments > DCS: F/A-18C Hornet Announced

Pigeon wrote on Jun 1, 2017, 10:46: Picked up an Asus 27" 1440p IPS over the memorial day weekend. Had been hoping to see improved versions with HDR and better back light bleed control coming out, but the manufactures seem to be focused on curved screens and 35" monstrosities. Speaking of back light bled, it's there, but I only notice it when staring at a black screen, which has only occurred when I purposely created a black screen to look for bleeding. I can't say it's made me a better FPS player, but at least some of those, 'how the fuck did they see me' moments are understandable now. Turns out, bigger screen + better colors and light levels = easier to distinguish targets. Member of the 34" Master Race, here: it may be a monstrosity, but it is a glorious monstrosity.

Re: Morning Mobilization

May 31, 2017, 18:28