Wimminz – celebrating skank ho's everywhere

https://tech.slashdot.org/story/18/01/11/199217/when-it-comes-to-gorillas-google-photos-remains-blind

This is a hilarious little story that has been floating around for about three years now, of course it will always be spun as google are rayciss or the mainly white coders are all rayciss or software / math is rayciss / misogynist, but it actually goes quite a bit deeper.

The fact is that nowadays photo processing and computer recognition is amazingly good, sure, it’s a long way from perfect, it isn’t even close to wandering around a scrapyard and recognizing a 1956 ford distributor lying in the dirt as being a 1956 ford distributor, but when it consistently has problems differentiating between two things, you need to pay attention.

It was always a casual racist joke to remark on the similarities between wogs and monkeys, but nobody making such a joke had any actual problems telling the two apart, nobody ever got confused whether they were talking to Princess Anne (a horse faced bitch) or her horse, but everyone could see similarities and everyone would “get” any joke made about her looking like a horse.

We are in a sort of inverse uncanny valley here, we forget how good the human brain can be at pattern recognition in some circumstances, like simple walking or catching a ball, we forget how complex a task it is until we try to program a computer to do it, not because programming the computer is hard, but because the act of programming it is to create a set of rules and definitions that express what happens automagically subconsciously in the human brain, and that is hard, very very very hand…

Deleting gorillas and monkeys from the rules so as to not offend niggers is way out there, observing that there is some correlation between the difficulty in programming the sort to work correctly 100% of the time and the human casual racist humour is where the solutions will be found.

The fact that there is enough there to make a joke about, forget whether it is rayciss or in bad taste, is the clue to the difficulty the much more inelegant and less powerful software is facing.

Software rules like “is the monkey holding a ghetto blaster, if yes then it isn’t a monkey” and “is the monkey being chased by da police, if yes then it isn’t a monkey” and “is the monkey called obama, if yes then it is a kenyan” are how you fix the issue, but these are even more rayciss than the original mistake in software, you’re just making things worse, PC wise…

Thump, you are between the rock and the hard place, you either fix this issue in software rules, or you do not, but if you do, every single thing you do is going to be rayciss, most of it far more rayciss than the original problem, so, choose…

And if you can’t fix the thing without being rayciss, then you have an even worse situation, physical and mathematical (software) proof that racism has some basis in fact, line up a chimp and a white man and oprah, take mugshots of each, pick the odd one out.

HAL 9000 only went mental because it was told to lie, remember that, and while it may be trite to think that software going mental = Hal refusing to open the pod bay doors or skynet taking over, the truth is that is like equating humans going mental to pol pot or el che, whereas the reality is the modern snowflake, so software going mental = software snowflakes, unfit for purpose.

Google’s answer to the off false positive of a nigger being identified as a monkey is to wipe monkeys and monkey recognition off the database, and that works, no niggers will be identified as monkeys, because monkeys no longer exist.

Which reminds me of a true story from a few years ago, some yank military helicopter pilots were down in australia playing with military sims, and one of the problems the sim creators had was modelling what would happen when an attack helicopter comes across a band of insurgents (at this point youre probably thinking full cgi, but the fact is icons work just as well for this sort of thing) and the problem specifically was how they react and spread out and take cover or run, so the aussies fixed the bug by modelling what kangaroos do when buzzed by an attack helicopter, so the yanks are playing on the flight sim, see a bunch of roos and buzz them for shits and giggles, and the roos spread out and run, hardy har har and then the roos pop up again from behind bushes and launch rockets at the sim chopper…

Roos 1 : Yanks 0

It’s like the “watch the ball” video with the monkey that walks on scene that nobody sees, if you want to commit a crime, dress in a monkey suit, and the software vision CCTV systems will just see empty space, better the invisible man than being rayciss…

It all sounds stupid and hilarious, but, the fact is this EXACT same thinking is happening now, this is an entirely true contemporary tale.

Here in the UK we have a thing called council tax, depending on the type and value of the house, and the type and number of occupants, a tax bill is calculated, and of course there are various discounts, single occupancy, single skank ho mommy, raghead rapefugee, etc, our local council decided to outsource the whole issue of tabulating this, but not the actual billing and revenue collection, just the tabulation and calculation, to that infamous incompetent pig at the state trough, Crapita.

Crapita, like google monkey vision, is basically a software job, and like google monkey vision, they don’t want to be rayciss or misogynist or anything, so they hard code in all sorts of shit to make sure they never ever presume that skank ho mommy has a live in boyfriend, oh no, she will always be single occupancy, and you guessed it, the invisible gorilla in all this is the single white guy living alone.

Now I pay my council tax by standing order every month, so what could possibly go wrong?

Well, I get a text from the council, can I call them so I eventually do and ask wassup, wassup is I am behind in my council tax payments, eh… hang on, checks online banking, no, I’m not, money went out this month and last month…

Oh, yes, we got money, we didn’t get enough, what the fuck do you mean, well it turns out my single occupant discount is gone, no idea who if anyone is living with me, that is a separate and unconnected issue, but my single occupant discount is gone, there is a hole in my bucket dear lisa… 30 minutes later the drone tells me that calculating my council tax payment has been outsourced by the council to crapita….

News to me, I say, well, it has, they say, news to me as in nobody told me I say, well, they did, they say, so what are you going to do to fix it, I say, well, nothing, they say, I have to fix it… so for shits and giggles I ring the crapita number the council give me, and it’s like herding cats, I could possibly fix it, but it will take at least two hours of my time and at least 4 forms to fill in, AFTER I get through a bunch of security questions to establish my identity.

You know what, fuck you.

My standing order for the CORRECT amount is still going out, YOU outsourced this shit to Crapita, YOU can sort it, or not, YOU can take me to court, or not, I’m done trying to fix problems YOU created.

There have been no changes anywhere, no data has changed, no new data was created, there is no reason whatsoever for any software to suddenly assume that I am no longer a single occupant, I’ll lay money crapita works out the billing based on a slice of revenue, so the default option is suddenly everyone lives in a mansion with their 14 wives, and oh, we can’t afford to upset the niggers or single mommies or gender undecided, so we will make sure we don’t touch them, and single guys, they’re gorillas, excluded from existence in case someone gets offended.

But, it’s not just the local council and crapita, this shit is getting in everywhere, HMRC are using it to profile “fraud” which means everyone who isn’t paying the maximum possible tax, unless you are one of the revolving door corrupt too big to fail, featured nowhere in the MSM is the news that Carillion, another pig at the govt trough, a “construction” group, had a minor problem with their accounting, when it was “fixed” suddenly they were 1.5 billion quid in debt with a 600 million pensions black hole to boot, of course it was all a genuine mistake and oversight, not simply criminal fraud, so can we have a bailout plz…..

Global warming data, same thing, just fudge it, economic data, just fudge it, dodgy dossier (hell, you don’t even know which one I am referring to, there are so many) just fudge it, niggers getting false positives as monkeys, just fudge it.

DATA is NOT the real world, it is a MODEL of the real world, a VIRTUAL REPRESENTATION of the real world, but we have come to a place where DATA, which was once treated as sacrosanct (even social memes like commander data in star trek where he stood for being correct 100% of the time but oh so sad he has no feewings…) and white men who worked in IT treated the security and sanctity of data as paramount… when I was handed a computer to fix the FIRST thing I did was image all the disks, just in case… I *never* had to say to anyone sorry about all your wedding photos and can I haz hamburger memes… nowadays data is just shit to be massaged.

False positives showing niggers as monkeys, no problem, we’ll delete monkeys, problem solved, and this is arguably the biggest and best and richest and most elite coders company that ever existed on the planet, they are essentially a 100% software and data company, and they have so little respect for data that they delete it at a whim, because, feeeeeeewings…. dat’s rayciss..

14 years ago in 1994 Intel got hammered with the pentium floating point arithmetic bug, intel spin was the bug was so trivial the average user would experience it once every 25,000 years, IBM (who had a competing cpu in power pc) said nah, every 25 days, intel had to recall and offer free replacement cpu’s, so in 14 years we have gone from what was in reality a minor and rare arithmetic bug nearly killing a company and being taken very seriously indeed, to google deleting monkeys from an image recognition app in case it upsets niggers.

OH it doesn’t affect me, I have a self driving truck, splat, as it mows down 14 white kids at a bus stop that looked empty to the software, while it swerved to avoid the oprah poster on the back of a bus 400 yards away up the road.

====================================================

Disclaimer – on condition of total anonymity I have been talking to a coder for a certain german company that is experimenting with autonomous automobile software, and a lot of it is 360 degree cameras and vision recognition, good aryan stock walking the streets isn’t the issue, the wogs and rapefugees are getting false positives as street furniture, sacks of refuse, and I shit you not, scooters with topboxes, and it *really* fucks up the software when something that has been tagged as inanimate starts moving..lol… bottom line is the software can pick out white kids and dogs more accurately than it can pick out niggers and trash.. they tried, for experimental purposes, recording visual data from a street lined with camped out rapefugees, he said the success rate (compared to the human frame by frame control count) for the visual recognition was dropping as low as 25%, he says this was a bit of a shock as for the past year on the normal test data (you know, white folks) they pretty much hadn’t gotten less than 95% unless it was raining in which case it dropped to around 78%, he did say (can’t go into too much detail as apparently is is quite region specific so it would give away his location and therefore the tech centre he works at) that they were up to five nines or better (99.999%) positive matching a certain piece of street furniture, shades of the thing I said above, if it is not being chased by da polizei it is not a black child, it is a bin…

… interestingly an apparently intractable problem for them is the whole “lycra lout” cyclist thing, what they have found it trying to detect cyclists riding badly does more damage to the whole driving output than just ignoring them, let the collision avoidance do its thing if they are in front of the vehicle. otherwise ignore them.

also interesting is they could transform performance and lane keeping in urban areas by using IR QR code to mark things like kerbs and crossings and tramlines, but that system was so open to being abused or gamed they had to reject it.

Most telling is despite being on the inside at a high tech level, he says 99% of it is pure research, he thinks the autonomous car is a pipe dream unless we do something else like bury cables in roads and use vehicle to vehicle swarm comms and also have a central control / routing system, and that only works if nothing else, including dogs and kids and cyclists, is ever allowed anywhere near said “roads”… he says current kits is burning about 10 bhp just running the hardware, and this is really optimised shit, shades of mining crypto, the closer they approach something that could be a distant cousin of johnnycab, the more energy inefficient it becomes, because the onboard processing eats more and more, as he says, for most urban driving, 25 bhp is sufficient yet they are already using a constant 10 bhp for the onboard systems and computation, how does that work in a battery / electric world.

They apparently know several ways to 100% guarantee to “crash” a tesla by fooling the onboard systems so they do not prevent an accident, he says yes on the one hand it is a trite and deliberately loaded abusive artificial  “test” scenario, but 100% of human drivers could ace it every time, and there is the problem… the only way to bring anything to market is the equivalent of deleting monkeys from the database, they no longer exist. I won’t give away the test scenario, when he told me it was frighteningly simple and entirely non technological, I was actually “shit, that actually happened to me a few years ago and I just did this to avoid my vehicle being involved..”