Forecast 2018:  What's Ahead for the Language Industry?

Forecast 2018:  What's Ahead for the Language Industry?
November 22, 2017
With Michael and Renato
In 2017, we saw significant advancements in translation technology, most notably in the growing use of neural machine translation. But technology isn't the only thing that is having a major impact on our industry these days. Hear what our hosts believe are the major trends we'll all be facing in the coming year.
Download episode
Show transcript

Transcript

Speaker Transcript
Renato I’m Renato Beninatto.
Michael And I’m Michael Stevens.
Renato Today we’re doing our annual show about trends for 2018.
Michael It is that time of year again isn’t it Renato?
Renato Amazing that this year just went by.
Michael We sit here and we pull out our crystal ball…
Renato Well, one of the things that I like to present as a caveat in the beginning of this conversation is that we’re talking about very short-term trends. We’re not talking about revolutionary trends. These are topics that we are probably already talking about but they’re going to increase in scale next year.
Michael Yes, so for instance last year we talked about neural machine translation, this was after the New York Times article on it. We released three podcasts giving an intro/summary to our listeners about what neural translation means for the industry. It was already in the ether.
Renato Yes.
Michael It hadn’t started trending on Twitter I don’t think yet, but it did get to that after, of course, we mentioned it as the big trend.
Renato Yes, and because we are connected and we’re talking to many people, we participate in events and associations and so on. We’re pretty much taking the pulse of what is going on in the industry, and we have a sense of the things that are ready to pop up, to be the next big thing. This doesn’t mean that they are new, we don’t have such a good crystal ball yet.
Michael No.
Renato Actually we should mention here, and I want to make an interesting comment, the founder of TAUS Jaap van der Meer is clearly the most visionary person that we have in our industry. Do you know why?
Michael Why?
Renato Well, this is a guy that back in the ‘90s, he was one of the fathers of the localization industry; we should bring him to our podcast. He had one of the first localization companies, but he was always talking about the future, how translation was going to be, how translation technology was going to affect us and so on. But the very interesting thing is that we’re living today a reality that he saw 1520 years ago, and everybody thought he was crazy. So, that guy has the crystal ball. We can talk about trivial stuff.
Michael Trends and things we’re seeing on the marketplace. Well, let’s do this. I think the format we’ll do is we’ll each share one with another, we have each of us brought three things that we think will affect the localization industry…
Renato I don’t know what are yours.
Michael I don’t know yours as well, and so why don’t you go first. We’ll go down that route.
Renato Okay, age before beauty, is that the rule?
Michael I was going to say that and then I thought better, so you said it for me.
Renato That makes me think of that concept of the Cochrane collaboration in healthcare where this guy’s revolutionized medicine by bringing evidence and statistics, and one of the things that I love about it, and you’re just proving that that works, is that you’re trying to replace eminence with evidence. So, I am the eminence, and you give me the first pass so that you can prove me wrong.
Michael Yes, exactly.
Renato Okay, so His Eminence Renato Beninatto believes that we’re going to have… we chose three topics. My first topic for next year I think is an increase and a tightening for global talent acquisition. Let me put a parenthesis on that word talent because the word talent in itself is overused, not everybody is talented. Talent should be an exception.
Michael So, you’re saying we’re going to see more talented people available?
Renato No, there is a huge demand for people with specific skills, and those skills are constantly changing, and they’re changing at a much faster pace.

You have an old guard, people like me, who have been in the industry for a long time that have their set ways and beliefs and a certain resistance to change. Some people are more open than other people to that kind of challenge to change and adapt. The majority of the old guard, people who started working on typewriters, like I did, will prefer to complain and say that the changes that are coming are not going to take hold, because we need to resist, we need to unite, we need to combat machine translation and innovations that make productivity better.
Michael Yes, and if you’ve been around that long you have full right to complain because you have been through so much change. You’ve seen…
Renato Well, absolutely but you see what happens is that… and I’ll refer to the Cochrane initiative again, what happened in medicine, for medicine to change from eminence-based to evidence-based was that this old guard died. So, the old guard, the older people, the more experienced people, let’s put it this way, the more experienced people need to learn how to work with the new guard.

Innovation is going to come from the young generation, they’re going to come from the millennials that are coming into the market that have different skills from the Gen-Xs and the people that are in the market right now, but they need coaching, they need to learn some basic protocols, even basic behavior, to be honest.
Michael Yes.
Renato But they’re more collaborative, they’re more creative, they’re more flexible and those are the skills that buyers and vendors of translation and localization services are looking for. Because the job that we’re doing today is going to be different tomorrow, and it’s going to be completely different the day after tomorrow, and I have no idea in this fictitious timeline what we will be doing next month.
Michael This idea, you’re pushing the idea of talent and these skills further. In the podcast we did on the Netflix effect we talked about the translator shortage and how there’s just a need for that, but you’re saying there are even more jobs in our industry available.
Renato Well, we have project managers, we have engineers, we have accountants. We have a bunch of people that need to have… embrace the new skills and the new mentality instead of resisting change.
Michael Alright, well for my first prediction I believe we’re going to see more consumer goods companies go direct to consumer. Consumer packaged goods.
Renato What do you mean by that?
Michael So, you’re looking at a company who used to distribute through say a Walmart or online, some place like Amazon, and you’re finding, Nike being the gold standard, Nike did approximately $9 billion in direct to consumer selling.
Renato Apple is another classic example.
Michael Apple is another. They just cut out the middleman who…
Renato They still use it but it’s a small part.
Michael It’s a smaller part of their channel. So, you’re going to see companies that we are used to going into the grocery store and finding their products are going into a larger big box store, and these larger stores are coming up with their private label items more and more, and promoting them more and more. So, it’s to these brands’ advantage to get their message directly to buyers.
Renato This speaks very closely to what I was talking about, it’s a change in behavior, right? I remember when online shopping came in, I hate to sound like an old man, but…
Michael It’s the time of year.
Renato Yes, and people said that they would never buy online, and yet these same people are buying on their cell phones, and this behavior has changed completely, and it’s funny because the main trend that I see that you’re talking about is this intermediation. Because they’re going to sell more online, but also through their own channels, they’re eliminating the distribution. And what’s the implication for our industry?
Michael Well, you see you these brands are going to be trying to be more effective with the customer, they’re going to be trying more campaigns. Again it’s this proliferation of content. They’re going to be experimenting, sending things out, you think of like Unilever a brand that’s very traditional, we all probably have Unilever products in our bathroom. Now they own Dollar Shave Club, right? So, it’s an entirely different line of marketing that’s coming which means more words, which means more work for our industry.
Renato Fantastic, and from more channels so that creates the opportunity for smaller players, smaller LSPs, to be dealing with local brands that expand internationally, these barriers to delivery tend to fall.
Michael Absolutely.
Renato Excellent, I love that.
Michael Yes, alright, what is number two for you?
Renato Well, my second trend it’s actually the same that we had last year, I’m going to mention AI and machine learning, but I think that in 2018 we’re going to go to this super hype curve going all the way up. Everything will be AI and machine learning.

Not only in the language industry, everywhere, but if you are small LSP in Wuhan in China, or in Peru or in Argentina, where we have listeners, if you don’t have AI and machine translation, machine learning in your website, you’re missing an opportunity to charge more money.

Everybody will be talking about it. It’s going to be overused, and I think that this happened with Crowdsourcing maybe 1015 years ago. This happened with translation memory, longer ago, but now everything is going to be neural machine translation, AI, machine learning. These three things everybody’s going to be talking about.

I think I remember a similar period we had in the localization industry was when markup languages started to become popular and the first one was SGML. And I loved a description that I heard at a conference, because everybody was talking about SGML, and I had no idea what SGML was, and a guy from a client says, you know, SGML you either do it or for you it’s something that sounds good maybe later.

So, I remember that joke, and I repeated that every time, and everybody thought I was an expert because I could make a joke about the topic. But it took me years to understand what SGML was, and by the time I understood it, it was obsolete.
Michael Yes, XML had replaced it, or whatever the next line was.
Renato Exactly XML and even HTML had replaced that before that.
Michael And this is really part of that idea of we over-anticipate the short-term impacts of these technologies and under-anticipate the long-term. So, it’s going to take this period, and this relates to my second point which I’ll just go into.

It takes this period of many people getting involved in these technologies, vetting them. I got to hear about a company who was using Google’s neural machine translation for UI currently. This is something that Google themselves haven’t tested the effectiveness of, and it’s because there’s… you have to prioritize. So, they’re looking at what’s most effective for them as a company whereas other players can come in and start using this.

So, my second point was that these open source products, while pretty much every large company in the world, Airbnb, Airbus, eBay, Intel, they’re all using like the Apache Spark, these large open source projects to run a big part of their IT infrastructure.

Now there are these areas of machine learning that are available, so TensorFlow, it’s Google’s open source library for machine learning, is available to anyone. So, companies of all sizes can get involved and start working on this.
Renato So, Google, Amazon, Facebook, all of them have these send boxes where you can just bring your code, and test it, and play with it and see what results are coming out of that. It’s a completely… it’s a much more collaborative approach, and this goes back to my first point of the new talent.

This new generation is more collaborative than my generation. They have less concerns about proprietary information by sharing talent. We go back… it’s funny because now that I think of it, it’s like waves, right? I mean the biggest promoter of sharing information was actually Benjamin Franklin who created the patent system, and the patent system is nothing else than protecting information in order… so that it can be shared, and more people can learn from it, and use and improve on the inventions.

And this is pretty much the approach that we’re taking now. I think maybe artificial intelligence is replacing lawyers, and that’s why we have more openness for this kind of stuff.
Michael Well, I’m excited because I think it’s a place our industry has the opportunity to have big impact. We have many people who are already familiar with the fundamentals of machine translation and areas where this is incorporated, and now we can start looking at other business functions to incorporate this.

Well, I’m going to jump into my number three since we dovetailed on number two, and then you can close with your last one because it is somewhat related. I think this year we’re going to see the rise of the bots. Do you know how old bots are Renato? How many years have bots been around?
Renato I would say that from the beginning of coding, people were developing.
Michael Some type of automation to do it.
Renato Yes, or like “Hello World!” was the first bot ever.
Michael Well, I saw today that bots celebrated around 50 years at this point they’ve been around. And they’ve been incorporated in many of our daily uses. This relates right back to AI and machine learning, so if you use a product like Slack, the Slack bot is going to come and offer its assistance to you when it sees that you’re doing something that it understands.

Facebook has this clearly tied into their messenger service for companies and they’re able to take advantage of that. Going to conferences where they have people standing around and directing you where to go within the conference. There was a presentation of someone from Oracle who made a bot that could serve so you have the interaction of, hey I need to get to my session, where’s the map, how do I get there, what’s the quickest route? Opens your map to lead you right there.

So, I just see that this is becoming easier and easier, people are seeing ways to incorporate it in really, really useful moments, and I think we’re going to have more of our lives impacted by these automated helpers.
Renato And this has impact in customer… multilingual customer service, it has an impact in chat bots, and self-serving shopping carts, and support and things like that. All the activities that used to be available from 7am Pacific to 9pm Eastern, or whatever, now you have these automated response agents that can help you, and the funny thing, I’m sure you saw that, is that they put bots to talk to each other and they…
Michael That was the big news of this year, was the bots at Facebook who began creating their own language. They just sort of got into abbreviations and it was great. The news about it was all, the bots are going to take over, we can’t understand them and then when you read into the details of what was really happening, they were just using shortcuts like any people.

Any slang that you would have with friends these two bots started communicating in ways that were more efficient, and the programmers looked at it and said, oh we messed up, we didn’t tell them to continue to speak so we could understand so of course they’re going to go do that.
Renato Well, this is the beginning. I mean I love a quote by William Gibson who says that the future is already here, it’s just not evenly distributed. So, we have parts of this science fiction already around us.

One of my favorite movies is The Minority Report, not so much for the story but also for the technology that appears in that and when it came out, maybe 1015 years ago, all those things seemed so crazy and so futuristic and now…
Michael Right, and now police departments are able to estimate what crime’s going to happen on what corner of what block.
Renato Yes, and iPhone 10 has this face recognition that Tom Cruise was being recognized every time he went to a store and so on.
Michael That’s right, yes.
Renato So, all these things, I mean we need to learn how to live with them. We are probably part of a generation that is a transitional generation, our children and grandchildren will be living and sharing a lot more of the machine experience than we have, and it will be totally natural.

There was resistance when the first steam engines came out, and they thought that in England that the cows would stop giving milk because of the noise of the trains going through the pastures. So, they didn’t, we still drink a lot of milk.
Michael Yes, that’s true. Alright, so for your final one to take us out, it better be good.
Renato Actually it’s not good. I will say that I don’t like this very much because I thought we were going out of this. I’m not going to talk about evolution, I’m actually talking about an involution. We’re going back to a period of fear, and it has to do… I think that a big topic that is going to resurface in our conversation is going to be security, privacy and confidentiality.
Michael Yes.
Renato And I can pinpoint this moment. It’s like 911 was the end of freedom to travel. I think that this, the story that happened in 2017 with translate.com that had some proprietary content be part of the machine translation algorithm that was made public. And also the leak with Equifax here in the United States where the credit history of tens of millions of people were affected.
Michael Yes, the first story you’re talking about it was translate.com and what was interesting is it was the lack of reading user agreements, because it’s not like they didn’t publish what they were doing.
Renato Yes, but the problem is that nobody reads this.
Michael No one reads these things.
Renato When was the last time you read…
Michael I never do.
Renato Okay, just click agree.
Michael Just click agree, it’s the easiest way.
Renato Yes, you sold your mother-in-law, agree, right?
Michael Yes, but whereas the Equifax breach are hostile actions being taken against a company and succeeding.
Renato Yes, but, you know, it doesn’t matter because the company that had this issue was translate.com. You have no idea of how many people think it was translation.com or that it was translated.net, that it was other companies…
Michael Similar domain names.
Renato That in reality have nothing to do with that. So, unfortunately that story is taken by the people that work on fear, uncertainty and doubt, the famous FUD approach, to avoid or to prevent the advance of machine translation technologies and things like that. So, if you put this with the talent challenge, this buys some more time for the people who resist change, this kind of fear. But this is something that is not affecting only our industry, it affects every industry.
Michael Does this mean language service providers should review things like their NDAs with translators? What does it mean specifically for people in our industry?
Renato We need to make a show about this because this is a long conversation and my opinion is essentially we live in make-believe environment where we make believe that NDAs are worth something. We make-believe that people understand what confidentiality and privacy means. We make-believe that these things affect quality in any way and so on. So, I’d love to have a debate at some point with somebody that is a fierce defender of those kind of things.
Michael I’m sure we’ll get some notes from those people after that comment. They’ll reach out to you in Twitter and say I would be glad to debate you.
Renato Exactly, so the last point that I would like to make Michael is that we need to keep in mind that all the things that we talk about here will affect parts of our industry in different ways. Our industry, the language business, affects every human activity that you can imagine. It’s one of those all-encompassing professions. We have so many new answers and every industry requires some kind of specialty. But the speed of change, the speed of adoption of machine translation, the speed of adoption of technology is going to be very different in different segments.

So, if you look at regulated industry, adoption is slower because you have controls and regulations that maintain a certain way of doing things. You have industry standards and things like that, but you have other areas like software and like e-commerce, and things like that that will have a lot more automation and a lot more reuse than other spaces. So, keep in mind that for some people the things that we’re talking about in this episode are already the past, for some other people it’s tomorrow…
Michael It’s pushing them.
Renato And for many, many people it’s still the day after tomorrow.
Michael Yes, absolutely but the one thing that everybody can take is flexibility and openness that’s going to be required to move forward with all of these changes. With every change there comes opportunity, with every change there comes loss, with every change you have to decide what it means for you personally. So, I think… I’m hopeful about what 2018 has in store for us and we’ll check back in and see how we did next year.

End of conversation

Stay Tuned

Subscribe to receive notifications about new episodes

Play episode
0:00
0:00