Slashdot Mirror
Replacement of Writers Leads Gartner's Predictions (computerworld.com)
dcblogs writes: Gartner's near-future predictions include: Writers will be replaced. By 2018, 20% of all business content, one in five of the documents you read, will be authored by a machine. By 2018, 2 million employees will be required to wear health and fitness tracking devices as a condition of employment. This may seem Orwellian, but certain jobs require people to be fit, such as public safety workers. By 2020, smart agents will facilitate 40% of mobile interactions. This is based on the belief that the world is moving to a post-app era, where assistants such as Apple's Siri act as a type of universal interface.
The kind of document Gartner's talking about isn't the kind that's written, it's the kind that's transcribed from facts with some formatting applied. As the article says, it's sports scores and budget reports and such. It's the kind of stuff I call "boilerplate" and write scripts to handle, eg. to take a small input file with the information defining a C++ class ("This is the class name, these are the data members and their types.") and spit out a properly-formatted C++ class definition complete with all the constructors, assignment operator and standard methods needed (which is oftentimes 2 orders of magnitude bigger than the input file). Actual creative writing, the kind that requires coming up with the information to put into the document, is in no danger of being replaced any time soon.
They are so hilariously wrong so often you could build a successful career out of assuming they will be wrong about everything. A selection of their idiocies:
Bogtha Bogtha Bogtha
Because voice recognition - just for starters - hasn't come on much in the last twenty years.
Last time I used Siri (which was only a few months ago), I asked it a simple question and it just sat there baffled. I spent twenty minutes trying all kinds of simplification, better pronunciations, and rewording but still it wasn't able to fathom anything useful from it. No, I don't have a strong accent (but what the fuck should that matter anyway?) and no I wasn't in a room full of noise (but - again - are we going to have to go outside and find a quiet spot to get these things to work in the future).
Apart from where there are obvious detectable keywords that they can make up the rest of the query around, these things are SHIT, and always have been.
I work in schools, I've dealt with a number of teachers and "learning support specialists" who hear that there is a voice recognition software, who then insist we need to use it for those children unable to write properly, and then trial it and discover just how useless it is - especially if the child already has even the most minor of communications problems too - and then realise what a waste of time it is.
One teacher I know wanted to write all their school reports using voice recognition because they were sold how wonderful it was by some guy paid to train them. Yeah, in a silent hall, using his exact phrasing, it seemed to work. Ten times slower than typing, but the demo was nice. However, you've not saved time or effort, you still have to double-check everything before it goes out (and inevitably on a computer because the devices aren't even close to being able to be controlled by voice - "Oh, no, change that word elephant to giraffe, please") and the accuracy in any real-world environment or using anything other than very basic phrasing SUCKED. I laughed when they told me that's how they wanted to write their reports - hundreds of them each per member of staff within a one-week window. The technology is honestly that bad.
And the rest is just bollocks of the highest order.