There’s this people called stenographers who are paid quite well, they can write hundreds of words per minute and essentially transcribe a conversation in real time. They are hired by courts to create records of the sessions, by journalists, parliaments and to transcribe subtitles for audiovisual media. They use this cool typewriter like machine called a stenotype that was invented in 1880. The thing is, they tried to replace them with speech recognition computers. They discovered they needed a human to sanitize input for the computer, essentially a person who can speak really fast and really mechanically, repeating what others said in the room, or what was said in the movie or whatever, into an oxygen-mask-like sound proof microphone. So, they still had to pay someone to be there. Many places decided they could just pay the stenographer and receive higher quality products despite the slightly higher costs. Then YouTube tried to use machine learning to auto-create closed captions. Before that they used a community contribution approach that depended on volunteers to take some time to transcribe the subs. That change to automation was such a fiasco that some big YouTube channels now advertise that they pay an actual company with humans to do the closed captions for their videos in the name of proper quality accessibility. Because automated closed caption tends to do interesting stuff and it’s even worse when they try to throw auto-translation into the mix.
The point is, people tend to not understand technology and how it relates to humans, specially techbros and techies who have the most skewed biases towards tech and little sociological understanding. Nothing can be accurately predicted in that realm, and most relations that result from the appearance of new technology are usually paradoxical to common sense.
There’s this people called stenographers who are paid quite well, they can write hundreds of words per minute and essentially transcribe a conversation in real time. They are hired by courts to create records of the sessions, by journalists, parliaments and to transcribe subtitles for audiovisual media. They use this cool typewriter like machine called a stenotype that was invented in 1880. The thing is, they tried to replace them with speech recognition computers. They discovered they needed a human to sanitize input for the computer, essentially a person who can speak really fast and really mechanically, repeating what others said in the room, or what was said in the movie or whatever, into an oxygen-mask-like sound proof microphone. So, they still had to pay someone to be there. Many places decided they could just pay the stenographer and receive higher quality products despite the slightly higher costs. Then YouTube tried to use machine learning to auto-create closed captions. Before that they used a community contribution approach that depended on volunteers to take some time to transcribe the subs. That change to automation was such a fiasco that some big YouTube channels now advertise that they pay an actual company with humans to do the closed captions for their videos in the name of proper quality accessibility. Because automated closed caption tends to do interesting stuff and it’s even worse when they try to throw auto-translation into the mix.
The point is, people tend to not understand technology and how it relates to humans, specially techbros and techies who have the most skewed biases towards tech and little sociological understanding. Nothing can be accurately predicted in that realm, and most relations that result from the appearance of new technology are usually paradoxical to common sense.
Here is an alternative Piped link(s): https://piped.video/watch?v=1C7leljxnG4
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source, check me out at GitHub.