We will still need human translators for serious translation work by 2040

Created by InquilineKea on 2015-08-04; known on 2041-01-01

• InquilineKea estimated 45% on 2015-08-04
• JoshuaZ estimated 55% on 2015-08-04
• HonoreDB estimated 40% on 2015-08-04
• RoryS estimated 70% on 2015-08-04
• timujin estimated 70% on 2015-08-05
• btrettel estimated 65% and said “I’m interpreting this as human translators will still exist, because their existence implies their usefulness.” on 2015-08-07
• Jayson Virissimo said ““Need” and “serious” shouldn’t be in here. They seem to be making the claim less falsifiable. A better approach would be to make a claim about how many people are working as translators, fraction of the workforce, etc.” on 2015-08-07
• btrettel said “I agree with Jayson. I’d recommend withdrawing this prediction and creating a new one that’s more specific.” on 2015-08-07
• PseudonymousUser estimated 70% and said “I imagine we would still use them for “serious” diplomatic talks, if only because we’ll believe they can pick up on more nuance.” on 2015-09-19
• PseudonymousUser estimated 66% on 2015-09-26
• luxpir estimated 75% and said “The last 25 years saw convenience brought to the working lives of translators, through IT, the next 25 years will bring us unnervingly close to replacement, yet the humans will still retain ultimate trust in ‘serious’ translation work.” on 2015-09-29
• themusicgod1 estimated 6% on 2016-07-10
• JoshuaZ estimated 42% on 2020-06-12
• Baeboo estimated 36% on 2020-06-13
• wizzwizz4 estimated 45% and said “We’ll need new algorithms; there are classes of errors that humans just don’t make (with understanding) that existing algorithms can only reduce (with improved predictions from increased training). Plus, NLP is NP-hard (e.g. garden path sentences).” on 2020-06-13