This is a tale as old as time. Not love, it is about keeping something you made safe from those who would destroy something beautiful. In this case, the thing of beauty is a talking banana who reads Twitch and Youtube comments. The ne’er-do-wells are trolls seeking to ban-anana the account by forcing it to recite restricted words.
The problems stem from a visit from [Greekgodx], whose followers tend toward the dark side. When they set their sights on [Mike Nichols]’ yellow automaton, things slipped into a bleak place, and a twenty-four-hour ban falls on the fruit. A bunch of filtering is done, but it isn’t enough to stop the trolls, and the tally-man adds a second permanent strike against the account. An arms race of slurs and filtering ensued until the robot was able to reject all attempts at racism.
The banana has since been peeled from the feeds, but if silly robots are your cup of tea, check out [Simone Giertz] turning a car into a computer mouse.
Thank you [Itay] for another fruitful tip.
Well that banana story was a little bent and was quite curly at the end.
I hope you liked how things shaped up.
That what she…
This guy’s deadpan humor is a-peel-ing. He didn’t mention if he’d blocked pig latin though. Or swear words in foreign languages. There’s almost always a way around any language filter.
Yeah, and people adapt – once you block the old words, new ones appear.
And then he blocked the phonetics, give it a chance it’s a good article
A guy once called me a socialist as an insult.
Took me many hours to realize it was meant as an insult because I am not from the same circles as he. When I finally realized I had to laugh.
This has been a good place for tech entertainment. Now however it’s going to be a place for politics
I disagree – it doesn’t seem to be political in the slightest, even though the headline might have you think otherwise. The situation is pretty well-defined – Twitch has prohibited a set of words from being used by streamers/chat users, making it a .bannable offense. There are now certain people going around looking for automated donation-driven viewer engagement systems (mainly text-to-speech systems, very popular with streamers). Once found, they’re trying to make the system inject the prohibited content into the stream, banning the streamer. The banana creator, however, decided to combat that to make sure his stream and channel stay alive – and succeeded in doing so.
To sum up – I went to read the article expecting politics, but was pleasantly surprised by the neutral tone of the article.
it was so sanitized i had to read your post to figure out what the hell the article was about.
Censorship is what you just saw in action. It will become political.
The article says it’s about one specific word though, not ‘words’ plural.
Which is odd since of course there are more than one word and one group you can use racist epitaphs against, but the challenge for the crowd seems to be that one word. And as such it’s probably less about ‘toxic racism’ than suggested?
Hmm, which article are you talking about? If this one, I just re-read it and it doesn’t talk about one specific word – I’m quite sure there’s more than one in the Twitch banlist.
The linked one
What race is this?
https://upload.wikimedia.org/wikipedia/commons/thumb/6/6c/Atlas_from_boston_dynamics.jpg/400px-Atlas_from_boston_dynamics.jpg
Seems to be white and likely from an affluent country.
:)
HaD censors like a deranged maniac on anything that is said that is vaguely political. And yet they keep posting subjects designed to evoke a political comment.
I wonder if it’s part of some clever campaign. The whole thing reminds me of China come to think of it, they pretend they want to ‘openly discuss’ everything and are really into ‘political dialog’, but that is only if it fits certain parameters and else it’s radically avoided and blocked and destroyed.
And when I talk about HaD censorship it’s not about avoiding extremist or anything, that is to say not about the normal editing you would do with a site on the internet.
First of all, I have to disagree on “designed to evoke” – instead, it’s just that HaD covers a wide range of topics and some of them overlap with politics. For example, I’m interested in this article for non-political reasons – I actually tried to make a couple of Twitch-integrated devices, including one or two designed for viewer engagement. Then, I can assure you – it’s not a clever campaign, it’s just a recent trend among tech companies to add this kind of censorship. Lastly, I only saw, what, 2 comments deleted that I’ve read before? One of them could be construed as coming from a racist viewpoint (not a well-defined one, though), another was the child reply to that comment (IIRC yours). I’ll check my crawler’s logs if you’re interested.
I’ve read the logs – there were only 3 comments deleted, out of 38.
Dr Lee tuning Nick’s filter again?
The phonetic filter is a very clever idea.
Yes it is, but it does not stop racism perse though, you can make extremely racist comments using all very proper words.
“This is a tale as old as time. Not love, it is about keeping something you made safe from those who would destroy something beautiful. In this case, the thing of beauty is a talking banana who reads Twitch and Youtube comments. The ne’er-do-wells are trolls seeking to ban-anana the account by forcing it to recite restricted words.”
Like Microsoft’s chatbot. People just seem to have a destructive side.
Nothing can ever make youtube comments ‘beautiful’, nor can anybody reading them ever approach anything resembling a ‘beautiful’ thing.
That really really depends on the channel, though. For example, when you find a discussion channel with reasonable comments, you know it’s a good channel – I’m coming more and more to believe that, while YouTubers don’t directly incite their audience, tone of the video sets the tone of the discussion in the “comments down below” (c).
Any news video, no matter how proper and balanced, will get a hundred ‘the evil jews are controlling the world’ and that type of stuff.
You can’t blame the video. If there is none of that it’s because of extreme obscurity of the video or heavy and constant editing and pre-approving comments.
In fact most all news orgs just gave up and disable comments on youtube, and even then at some point just give up posting videos on youtube altogether.
Oh and I found out from some TV documentary that they actually do allow internet for people in mental institutions as part of ‘social interaction’ therapy.
Draw your own conclusions.
I’m not talking about video orgs. I see what kind of comments you’re talking about – I’d call them fringe comments =) They’re going to be there in case of news orgs, sure, but they also don’t tend to get a lot of upvotes and rise up to the top, compared to a more reasonable opinion. I never see them – but then, I also don’t consume traditional media on YT, I don’t see any reason to watch on YouTube what people don’t tend to watch on TV either.
What I’m talking about are discussion/commentary channels, they tend to embrace all the features of YouTube (including comments) and interact with their audience. For example, let’s take a commentary channel, and the commentator releases a video reviewing a thinkpiece video from another creator. What I noticed is – if a commentator comes from the point of “I’m right, you’re wrong”, they attract an audience with the same mindset, reluctant to actually give the video the benefit of the doubt, and the comments will usually be shitposting and jokes, funny if you agree with the viewpoint presented but tricky to laugh at otherwise. If they come from the point of “let’s go through this, try to understand the viewpoint together and see if anything is wrong”, they attract an audience willing to discuss the viewpoint from many different angles, and that’s what happens in the comment section. In general, the more videos a channel has going from the “I’m right, you’re wrong” viewpoint, the more likely it is that each and every video of theirs will have one-sided comments – and vice-versa. It’s probably something obvious I’m describing.
Talking about your “mental institutions” comment – I think you’re off the mark here. First of all, I’d say the overwhelming majority of fringe comments has nothing to do with people in mental institutions, but instead with members of our society that might or might not have any mental disorders, but actually hold the viewpoints in question. I know, because I’ve met some of the “Jews control the world 9/11 was a hoax” people in person, if they had any kind of mental illness, you wouldn’t know => it wouldn’t be something worth putting people in mental institutions (there are certain criteria for that, they usually include violence or self-harm). Second of all, mental illnesses don’t imply having delusions of some kind as a certainty (and if they did, I’m highly sceptical they would use social therapy for treating that, I’m not a psychologist but I do know a thing or two) – that’s actually a quite wrong viewpoint, as there’s a wide spectrum of mental disorders and only a small number of them deal with the delusions of the kind you might be hinting at. For example, there’s a wide spectrum of anxiety disorders – plenty of people have those, they’re just undiagnosed and/or untreated. That’s something you could use social therapy for, but it’s not something that’d make you spout nonsense on the Internet.
Quote:[RFP-A]
“Oh and I found out from some TV documentary that they actually do allow internet for people in mental institutions as part of ‘social interaction’ therapy.
Draw your own conclusions.”
My conclusion is that in your view, all people who have a mental health issues also have diminished intelligence.
This view is, of course, delusional. Perhaps you may wish to visit said institution?
My takeaway:
If you get unicode, dump it. If you get unicode emails, trash them. If you acidently hit a unicode domain, null route them. If you do user comments, delete words with unicode.
Why am I hostile to unicode? The set is a obfuscated hacker’s wet dream. There’s more crap in the 2^16 charset than you can ever possibly plan for. Proof’s right here, and apple.com , and all sorts of other failure modes.
Same issue with IPv6 really, good luck blocking unpleasant IP’s with IPv6 in operation.
But instead of running away we just have to find a solution instead.
no guthub repository?
Kind of ironic how his “Useless Duck Company” will sound like “Useless F*** Company” when said fast enough or by a crappy enough TTS engine.
I’m sorry but I somehow doubt your statement that a TTS engine would ever pronounce duck as fuck.
Plus it’s not a racist statement is it.
What is questionable is how a TTS repeating something like snicker, which has to be there in written form as evidence it’s an innocent word, can lead to a ban.
I mean there is proof that the word was OK and if you hear something else it’s not really an offense of the bot saying it.
So they should just have allowed it until people got tired of the joke. Which I would expect to be quite quickly if there wasn’t such a reaction to provoke people into trying it.
And for many people censorship, especially when overdone, is more offensive than many of the things censored.
Absolutely agree on the “snicker” post – guess that’s the heavy-handed way Twitch uses to prevent certain groups of people of engaging, but, with inevitable complaint troll groups arriving, it ends up hurting the community in a different way.
Wait, so at 7:12, he’s saying that because the servo had a dead short inside of it, and was connected to a 5V 3A power supply (which was powered from mains power), that the dead short inside the servo is the same as holding the two ends of an extension cord together?! How can someone who knows anything about electronics say this? Or have I missed something here…
I agree with you. That’s a couple steps removed from shorting the extension cord.
Ya I goofed there, I was just trying to explain it in simple terms. The entire event was a surprise to me I have never experience such a fault with a servo.
Bananas were the reason that littering and sanitation laws were created a century ago to stop what is an often shown gag in the oldest cartoons. Before bananas, apples were comparatively safe unless drunk.
Yes, drinking apples can be very dangerous. It is to be avoided.
That’s what the internet does, it tries everything possible. If it’s a webapp or a chatbot. Anything which receives traffic will be raped. Great story about the filtering!
https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist