Josh Goldsmith on How Language Experts Become Successful

Josh Goldsmith, Geek-In-Chief of techforword, joins SlatorPod to talk about his multiple roles as a linguist, conference interpreter, and educator.

Josh shares his multifaceted background in linguistics and language teaching, highlighting his passion for language and education, which led him to co-found techforword.

Josh explains the current landscape of conference interpreting, noting a shift back to on-site interpreting post-COVID, which he prefers due to better sound quality, easier access to documents, and the human connection it offers.

He touches on the importance of personal branding for translators and interpreters, suggesting practical steps like professional email addresses, good headshots, and a strong LinkedIn presence.

Subscribe on Youtube, Apple Podcasts, Spotify, and elsewhere

Josh delves into the practical applications of AI in interpreting, with AI tools for glossary management, terminology extraction, and preparation for interpreting assignments. He also stresses the ethical concerns of AI and the need for human interpreters to navigate cultural differences and ensure accurate communication.

Josh discusses his various projects, including the AI Language Club, which offers tutorials on using AI for language learning. Looking ahead, Josh plans to focus on the AI in Translation Summit and the AI in Translation Bootcamp, which offer hands-on experiences with the latest AI tools.

Transcript

Florian: Today on the podcast, we welcome Josh Goldsmith. So Josh is a linguist, conference interpreter, educator, and many more things. He’s also the Geek-In-Chief of techforword, which provides on-demand training for translators and interpreters, and he’s also a regular contributor to our Slator Tool Box, a monthly newsletter for linguists. All right, look, complicated multifaceted background you have when we researched you, and of course, we’ve known you for a long time. So just maybe give us your intro yourself. You’ve done so many things from conference interpreter, linguistics, everything. Just give us the background.

Josh: I think it all boils down to the fact that I’ve always loved languages and I’ve always loved teaching. And so I basically have made my own path between the two. I studied linguistics in undergrad in the US. I went on to do a master’s degree in issues around linguistics and language teaching and multilingualism in Barcelona. Then I worked for a while as a high school teacher, a college professor, translator. I retrained as a conference interpreter in Geneva, and then I think throughout it, I’ve kept up my teaching. And about four years ago now, I founded techforword, and that’s where many people know me from, from teaching all sorts of geeky fun tech things for translators and interpreters.

Florian: How long were you an active conference interpreter?

Josh: I still am an active conference interpreter.

Florian: Wow, because I really admire conference interpreters. I’ve said it on the podcast before. I tried it in my translator education, and I think I could hold it for 20 seconds. And then my mind started wondering, and I just couldn’t keep it up. Very interesting. How do you manage all those multiple roles? Because when you told me what you are doing, I was just thinking, okay, how do you still have time for conference interpreting? How do you manage all these roles? Conference interpreter, educator. On the tech side, you’re very, very active reviewing products and things like that.

Josh: Good time management, I guess. I mean, conference interpreters, most conference interpreters don’t work every day. And on the days when I’m interpreting, I don’t do much in the way of teaching. And on the days when I’m teaching, I don’t do much in the way of interpreting. And I make it all fit in to the puzzle. I mean, I very carefully plan out my calendar to make sure all of the pieces fit together. We could record today because today I’m not interpreting, but tomorrow I am. So tomorrow, obviously, I wouldn’t have sat down for a recording session with you.

Florian: And now with conference interpreting, now obviously, COVID is two years past, et cetera. Is it now mostly back to in-person in the booth, or are you doing a lot of remote now as well, or what’s the proportions there?

Josh: I think that certainly there was a period when I was doing a lot of remote work, and now I’m mostly back to doing on-site work. Most of the time I work for international organizations. The European institutions are my biggest clients, and so that’s either on-site, on missions, and so on. I also work some here in Barcelona and in Spain, and there is some remote work that I do. But obviously, that varies from interpreter to interpreter. Some people work much more in remote settings, and some people work much more on-site, and that depends on you.

Florian: Do you have a preference?

Josh: Yes, I prefer working on-site. There are a lot of reasons. Sound quality is often still quite poor in remote settings, despite the fact that I mean, I and many other people have been educating clients about this for a long time. I think most colleagues now have good microphones and setups, but the same can’t be said for many of the people we’re interpreting for. Also, sometimes people are in places where the internet connection isn’t all that great. And of course, there’s all of the issues with visual cues and so on that you have on-site. Getting documents, it’s a lot easier to get documents on-site than it is when you’re working online. I think we all love the human aspect, too, of being able to see your colleagues and see the people you’re working for. And sometimes I’ll come up to you after a meeting and thank you for your work and that’s just really lovely.

Florian: I didn’t ask you what languages, so Spanish into English or do you have another one?

Josh: I have five passive languages. Spanish, Catalan, French, Italian, and Portuguese into English.

Florian: Oh, wow. Okay. That’s impressive. You built this big brand in the language industry. Was this a conscious decision at some point, okay, I’m going to build up that brand, or was it more incremental and it just moved?

Josh: I think techforword took off because of the pandemic. There was a period when all of a sudden, so many colleagues were home and struggling for answers and looking to learn. And with Alexander Drechsel, who co-founded techforword with me, he’s moved onto other projects now. But the two of us started a weekly webinar series back in February 2020, 2021, whatever it was, where we were just sharing our knowledge every week with colleagues, and that grew and spiraled. And we realized there were places to turn that into online on demand courses. And then one thing led to another. And then there were summits, the Innovation and Interpreting Summit, the Innovation and Translation Summit. They’ve now been rebranded as AI in Translation and AI in Interpreting Summits. A membership, and then another one sprung up. But basically, it all just developed over time as I was trying to help colleagues keep learning about the most cutting edge developments and stay up with the changing times and have fun while doing so.

Florian: Okay, so the branding and the marketing aspect of it was more a result of the content and the thing you wanted to do. So it wasn’t like, okay, let me go out there, build a big personal brand, and then start selling something. It was more like I had the thing first, and then I needed people to join the conferences, and it just grew like that.

Josh: We did some work around the brand, coming up with a name, coming up with a logo, coming up with colors before then, because we were always planning on having some courses. I think the original idea was quite far from where we are now, but I didn’t go out to build a big brand and then try to attract people to it. I tried to meet colleagues where they were to answer the questions they had. And throughout the process, I’ve been building the brand or brands, I guess you could say, because now I have quite a few, all under the techforword umbrella.

Florian: Would you agree, let me pause at something, would you agree that many translators, language experts, interpreters still sell themselves a little bit short in terms of personal branding and marketing? I think they do. What do you recommend will be maybe some easy steps, like very low barriers to entry steps to start building a bit of a personal brand?

Josh: I do think actually most language professionals don’t have much in the way of a brand. It’s obvious when they’re writing to you from a Gmail or a Hotmail account, we need to have a professional email address, and that’s not very difficult to set up. There are a lot of easy steps you can take. You can get a decent headshot that doesn’t have to be shot by a professional photographer, but it does need to show just you looking like you’re doing your work, smiling without lots of other people in the background. There’s lots of tools out there. One that I really like is called Free Profile Picture Maker, which can give you a fun background in your brand colors. You can do a little bit of brainstorming about brand colors. There are fun tools out there for that tool, like Coolers.co, which basically you just get to play and see which colors might match your brand. I do think it’s a good idea to do some thinking and write out a little mission statement, basically how you can help whoever your target audience is achieve whatever their goals are. So for techforword, it’s simple. It’s ‘at techforword, we teach translators and interpreters, that’s the target audience, how to become more productive and stay relevant in a constantly changing market.’ So they see some takeaway that they can get. You don’t need to build a website first things first. I do have a whole challenge about building a website, but just figure out who you’re serving, your mission statement, and put it on your LinkedIn profile. Put up a good photo, write a short bio so people can find you there, and use that other tool I mentioned to put up a decent banner. That’s a strong starting point. That and a professional email address, I think, are the prerequisites for personal branding, and then you can take it further. But one thing I say all the time, whether we’re talking about branding or tech or whatever it might be, is just one step at a time, lay the groundwork, circle back afterwards. If you try to do it all at once, you will never get anything done. Actually, I tell this story in my website challenge when I first went out to design a website, I wrote out all the content, I hired colleagues to translate it into all of my languages, I hired a graphic designer, and I didn’t understand how the site worked. So I was unable to update it myself afterwards. That is not the way to do it. Just put up a one-page site and slowly add to it over time. That’s how my brands have grown. Letting that trickle down to your social media presence, if you have one, you want consistency in your branding, of course. But yeah, it’s very possible, and it’s not even as hard as it seems. It’s just scary because most translators and interpreters and other language professionals don’t know much about branding or graphic design. And we want to work with words, not the rest of the aspects of business. Unfortunately, we also have to look at everything related to business.

Florian: Yeah, I 100% agree. I think baby steps is definitely the way to go if you’re intimidated by all of the things that you could do. But yeah, just set up an initial website. As you just said, headshot, get a presence on LinkedIn. LinkedIn is surprisingly efficient, actually, even though I prefer to be on X or Twitter for just info and scrolling, but just for branding and reach, I think LinkedIn is easier. I find it really hard to get any traction on Twitter, frankly.

Josh: Yeah, and LinkedIn is a good starting point for potential clients to find you as well. Just a few quick LinkedIn tips. If you’re doing the LinkedIn thing, don’t use abbreviations for your languages like EN or ES because your clients won’t look for that. Put the words your clients are going to be looking for into your LinkedIn mission statement, your tagline, and into your bio. Sure, you can write all of your professional associations and certifications, but no client is going to understand what ATA means, and they’re not going to go looking for it. So use the words they understand. Many, many colleagues get lost in this very easy step.

Florian: You’re so on topic. I taught a short course at my university here for bachelor students, like multilingual content writers. I used a lot of abbreviations, starting with TMS and MT. Then one feedback after the second session was like, hey, you’re using so many abbreviations. I don’t know what any of this means. I’m like, oh, yeah, okay. Your bachelor level, you have other stuff on your plate. I should tone it down. This industry is so full of these abbreviations and acronyms and things that outsiders probably don’t understand. Amazingly, you have time to also write for our Slator Tool Box journal, like our monthly linguist tech update. What are some of the topics you recently covered or maybe are planning to cover?

Josh: The last couple have been about a really incredible AI empowered writing and editing tool called Lex.page. The best AI transcription tool I have yet to find, which is called turboscribe.ai. I wrote a post about Zoom hacks for teaching online for people who teach translating and interpreting online, because I do a lot of online teaching, and I realized I knew all these little shortcuts and clever things, and I wanted to share those with people. I’ve written extensively about using AI for a wide range of things, from upgrading your vocabulary to looking for collocations, to preparation, to terminology extraction, and even getting AI to answer your emails for you. So it’s quite a range of subjects. I’m very lucky that you pretty much give me carte blanche to share whatever I think would be useful for colleagues.

Florian: Yeah, really great content. So let’s talk a bit about techforword. You mentioned it before, how you started it, but maybe currently, what’s the mission? How do you determine the topics, tools to teach? How do you build this out? Because it’s something we’ve at Slator looked at as well, but it’s incredibly complicated to come up with. It’s a very involved process. It needs a lot of dedication to do something like this.

Josh: I think it really helps for me to have a finger on the pulse of what’s going on by actually being a practicing conference interpreter. I translate a little bit for a few clients whose work I especially enjoy. I have a board game translation coming up later this month because I like board games, and I have a friend who is a board game designer, and I translate his games. Anyways, I think really just hearing from people about the challenges they face. I mean, I run this community in techforword called the Insiders, where people are constantly asking me questions, and those questions are great starting points, jumping off points for the content that I develop. So yesterday I ran a webinar. This is mid-June right now called Scan Like a Pro and that was because somebody asked about scanning an OCR, and so I was like, oh, I can teach people about that. That Zoom hacks post that I mentioned was because I was teaching on-site a few months ago. I started the session by asking, what pieces do you need to flesh out your teaching? It was a training of trainer session, and they said, we need to know how to work on Zoom. I was like, oh, I can teach that. And so the ideas come up organically. I also run quarterly office hours with the members of my community, and some of my blog posts have come out of the answers I gave people there. I wrote a whole about how to prepare for the EU accreditation exams, basically because somebody asked me. He said, I’m about to take these exams. What tips do you have? And that’s how it all develops, I think.

Florian: Yeah, it’s fascinating. There’s so much feedback that you’re getting from colleagues in the network. And I think that’s what makes it very relevant, that it’s really practical and not theoretical or a little bit 30,000 foot. So let’s talk about practical applications of AI in interpreting, like things you and other interpreters are actually using in real day to day, if anything. I mean, you said before that you like the on-site, maybe in the booth. But what tools or AI-powered tools are you using, if any?

Josh: Let me back up and say that there are plenty of tools out there that are not AI tools, that are very useful, that every interpreter should have. Every interpreter should have some way of managing their glossaries, and you don’t need AI for that. I mean, even a database tool will work. I don’t recommend Word or Excel. I’ve got a whole post on my website called Stop Using Word and Excel for Interpreting Glossaries. But you need a glossary management tool so that you can quickly look things up in the booth. And some of them do have an AI integration so that they’ll pull a running transcript of what’s being said in the meeting. And if one of the terms is in your glossary, they’ll show you the term alongside the equivalent from your glossary as you’re working. That can be a huge assistance. But there are plenty of other ways that I think we can use various tools. I love terminology extraction. I often get very long documents at the last minute and if that’s the case, terminology extraction is a huge lifesaver. You don’t actually need to use AI for that. I think the market leading tool for me is Sketch Engine, which doesn’t use AI, but does excellent terminology extraction. You can also get more fine-tuned terminology extraction for a specific field with AI tools. I’ve got tons of prompts about that in various blog posts scattered across my website. So how you can narrow it down to say, okay, but now just give me the legal terms, now give me the multi-word terms from this document or these documents or this website, or summarize this website for me so I can figure out if I need to read the whole thing or what the key points are or what the sentiment is, right? Whether it’s positive or negative, that can be a huge assistance when preparing for an interpreting assignment. I do lots of vocabulary research, terminology research, research about subjects as well using AI tools. You can do some fun stuff, too, like get it to process an image, right? Upload an image of all of the, I don’t know, muscles in the torso because you’re going to have a medical conference coming up and you need to know the names of them. And it can actually give you a list of the names of the muscles in the torso, even in multiple languages. This is something that I actually do and I also will actually take speeches from the people who I’m going to be interpreting and mine them with that tool I was just telling you about to get a transcription, run it through AI, find out what the key terms might be, even listen back to get a sense of the speaker’s accent and delivery. You can also do some stuff, a lot of colleagues are using machine translation as well. There’s a real trend in interpreting towards it being much faster and much more read out, so speeches that are written in advance. And a machine translation tool, if it protects the confidentiality of the information, which is a very big if, can be a good starting point. If you get it not that far in advance, you can at least get a speech machine translated, then read through it, correct it a little bit, and be able to deliver it and keep up with the pace. And then, of course, we’ve got everything related to speech recognition, which I think takes three different shapes right now. So one is just a running transcript, running machine transcription. Another is a running machine transcription that’s translated into another language, and then we’ve got the idea of machine interpreting. But I think we’ll talk a little bit more about that later. Those are not necessarily tools that colleagues are using right now. I think some are using a running machine transcription, which is good if you happen to be in a meeting that is public, where you can take the content and run it through a tool and use it as a starting point. And of course, it takes practice to be able to work with materials like this.

Florian: But a lot of it from what I’m hearing now is in preparation, so you’re going into the interpreting session fully prepared, knowing the speaker style, having a list of terms that you’re familiar with, et cetera. You mentioned before, some prefer to have it maybe live next to them, but is that because the distraction is too much or there’s nothing out there that’s actually useful for this for you personally?

Josh: So a huge part of interpreting is and always has been preparation. Way before we started talking about AI, interpreters went out and they prepared and they read the documents and they annotated them and they highlighted them and they prepared glossaries and they looked up terms so that you could take some of the cognitive load. You preload the cognitive load. That’s technical jargon for saying, get the heavy lifting out of the way in advance so that when you go into your assignment, this term pops up and you can say it. I mean, people would quiz on flashcards so that they knew the terms going into a meeting, and that stuff is just as important. And tech tools can help to speed it up or to quickly search for a term that you know is in your glossary when you’re working. A completely different field is around what happens when we have these tools feeding us numbers, feeding us names, feeding us terms from our glossary, or potentially giving us a running transcription during meeting. And there hasn’t been all that much research on this. Really, the leaders, in terms of the people who are doing the most interesting research on this, are Bart Defrancq and Claudio Fantinuoli. Some of the research has shown that a running transcription could lead to over-reliance, so people are relying too much on it. And they did some work in the ergonomics of the Artificial Booth Mate project to see what information should be displayed on the screen. Should it just be the numbers? Should it be the numbers with units? Should it be the names? Should we see the name and the equivalent from the glossary? Or should we see the full running transcript? What sizes should these things be on the screen as we’re working? The jury is still out on this. We don’t know the answer. This is really brand new stuff for interpreting, but I think it’s interesting, it’s helpful. The tools available on the market haven’t really caught up to what the research is saying. I think there’s one tool that’s out of beta that does some pretty interesting work. It’s called Cymo Note . There are two others that are still in beta. So InterpretBank is in beta and SmartTerp is in beta. And all of these are what are called computer-assisted interpreting tools, CAI tools, that could potentially help interpreters with their work with proper feeding of the glossaries in advance, right, preparation and with practice learning how to use these tools.

Florian: Yeah, I would imagine it’s incredibly difficult to add another layer because you’re there, you’re interpreting, you’re fully focused, and then you have this other thing going on next to that’s supposed to be helpful. As I just admitted before, I couldn’t do it. I couldn’t hold two things in my head at the same time. If you have a third thing going on on the side and you need to focus on that, then it has to be really well thought through to be helpful, I guess.

Josh: And that’s why we preload the difficult stuff in advance. That’s why we prepare so heavily, and that’s why it takes practice. I don’t look at glossary all the time while I’m working, and I don’t look at a running transcript all the time when I’m working. But as soon as a number comes up, I’m certainly going to check that number. I mean, interpreters have asked colleagues, their boothmate, for the number, for help with numbers for decades. One of the things I learned in interpreting school was just write it down if you’re not working, and your colleague will look or they won’t. But if they need that help, they can look over at the numbers that you’ve written down for them. It’s the same thing. And the tools that are doing this well are putting the numbers in colors, perhaps highlighting the units, maybe even doing things like giving us the equivalence, for example, if it comes up in miles, but you need to render it in kilometers, it would be helpful to have that right there. A machine can do that very easily to help us do our work better. Notice that I keep saying to help us do our work better, not to replace the work that we’re doing.

Florian: Let’s talk about the replacement thing. Without full AGI, things we can’t really understand yet, do you think it’s possible at all? I mean, I still think it’s kind of not possible, like the jumps and the leaps and the creative leaps you have to take in translation, but even more so in interpreting. I don’t see a machine being able to make all these omissions or micro-decisions on the fly just yet without, again, having some AGI megabot thing.

Josh: I think it’s instructive to look at what’s happened in the world of translation over the last couple of decades. People have claimed that machine translation was going to take over and there would be no role for human translators, and that hasn’t happened, but our work has changed. Tools like Google Translate can be helpful when you’re, I don’t know, traveling to a foreign country and you need to quickly get a translation of the menu to more or less figure out what you might be eating, but it doesn’t replace a professional translation. And I think what’s happening is we’re seeing these new tools entering the market of interpreting, machine interpreting. They’re often called speech translators rather than machine interpreting tools because they know it’s not really interpreting. And look, if the aim is to help you figure out, I don’t know, make a booking at a hotel in a foreign country while you’re on the phone or a Skype call or a Zoom call or something like that, and if it’s not quite right, you can ask for clarification, ask for repetition, those types of settings. I think a tool like this can be helpful, but it’s definitely a huge leap to go from that to professional quality machine interpreting. So there are also huge issues around access, justice, equity. Rolling out tools like this in medical settings, in legal settings, in educational settings is highly questionable from an ethical point of view, and I certainly wouldn’t recommend it. If you want to try it for your conference, go ahead, but do ask your listeners who don’t speak the main language of the conference how much they actually understood and what their experience was like, because they probably miss important things and their experience may be quite difficult. Right now, when we talk about speech synthesis, it’s changed quite dramatically in the last six months to a year, but it’s still not all that easy to listen to speech synthesis for long periods of time. And speech synthesis, which is one of the components, right, of machine interpreting, after it gets translated into another language, it needs to be rendered in the language that the listener is going to be listening in. It works decently well if you have the entire context and the machine has time to process it all. But interpreting is a really different beast. I mean, I have to change my intonation mid-sentence sometimes based on what my speaker has just said and the rhetorical devices they may have been using. And those types of shifts aren’t currently taking place in speech synthesis for certain. So there are a lot of pieces that aren’t there. And then you were mentioning things like gaps and omissions and background knowledge that people should have. And I think Jonathan Downey has spoken quite elegantly differently about the fact that interpreters do a lot more than just the words, right? We also are ensuring that a meeting or an encounter can run smoothly. We’re helping to navigate cultural differences that might need explanation and so many more things that humans do and that currently machines aren’t doing.

Florian: Do you think the fact that the speech, and that word always trips me up, synthesis, that it’s gotten so much better? I mean, sure, it’s not perfect, and probably after listening to it for 30 minutes, it’s hard. But do you think that may lead to people going more towards choosing automatic speech translation or speech interpreting or thinking it’s better? Then there’s also the problem that they hear something that doesn’t sound like a robot, at least for maybe a minute or two, and then giving it too much credibility or too much weight and thinking it’s all correct just because it sounds convincing, as opposed to maybe two years ago where it was just like, obviously a robot voice, right?

Josh: I mean, look, neural machine translation seemed really convincing, too. And there were still mistakes, things that were getting left out, things that were not quite right. The same is the case. We listened to the speech synthesis, and it sounds pretty good. If people want to test this out, just take your phone, open the ChatGPT app, and start having a conversation with ChatGPT in another language. It can handle quite a few. It still sounds like a synthesized voice, although strides are being made. And yes, people, potential clients, are going to think that this is good enough. And now, more than ever, it falls to language professionals to educate our clients. And I know it’s not easy, but to educate our clients about the gaps in these tools and the ways that a human can help to ensure better communication.

Florian: Some of the biggest clients/users for this are the United Nations and the European Union, and you’re accredited with both those organizations. How do you see them? What do they think about this? Are they running a bunch of pilots? Are they involving the interpreters in this? Or how do they think about it?

Josh: I don’t think they’re planning to replace interpreters with these types of tools anytime soon. There are some pilot initiatives testing this out, especially at the EU, mostly for accessibility. There was a huge project at the European Parliament run by Marcin Feder where they spent three years, it was called the Speech-to-Text, I think, Unit, where what they were doing is they were training tools to provide relatively accurate machine transcription of parliamentary proceedings, and then to do so in all of the EU languages. Interesting experiment, decent for promoting on the fly understanding, but there are big gaps. I mean, if you look at the demos from them, you can see that they weren’t perfect. And this is with a relatively contained data set, right? We know what people are going to talk about at the European Parliament. That’s a lot much smaller than the entire world of what people might talk about elsewhere. So I mean, the same thing, look, if you take a company like Booking that’s using machine translation, one of the reasons it works well is because the writers who write the descriptions of these hotels have to narrow down the speech that they use to say, this hotel has 43 rooms, 52 of them have double beds, there is a pool. Very simple language is much easier to machine translate or potentially machine interpret. So there are certainly projects underway at these institutions to look into these tools and see where they might potentially be able to help. There are also major concerns about using tools like this. Confidentiality, for starters, is huge when you’re talking about international institutions, political settings, diplomatic settings. That’s one piece we need to bear in mind and I think they value accuracy and quality. That’s why I don’t think interpreting is going anywhere in these settings in the near future.

Florian: Now, you run something called the AI Language Club. Tell us more about that.

Josh: Sure, AI Language Club. I’ve been talking about all these professional applications, translation, interpreting. AI Language Club is a project that cropped up around September of last year, September 2023, because a friend of mine, Kerstin Cable, who is a language teacher who teaches everything related to language learning online, and I were talking about ways that AI could be helpful for learning languages. At first we thought it was going to just be a one-off course, and then we realized that we had tons and tons of ideas for teaching language learners at all levels, from absolute beginners to professional levels, about how to have fun creating language learning materials that were adapted to their level and their interests. And AI can do a really, really good job of that. So the way the club works is that every week our members get a new 5 to 10 minute video tutorial with accompanying exercises and prompts to test out some fun way to use AI to improve their cultural knowledge, their speaking, listening, everything related to language learning. Would you like to some of the tutorials we’ve created? Because I can wax poetic about this for a long time. We’ve put out almost 40 tutorials at this point, one a week from September. It’s things like, all right, YouTube videos are great for language learning, for hearing authentic language in use, but wouldn’t it be neat if we could get those multiple choice questions or open answer questions generated with AI, and it could even tell us, oh, that’s not quite right, go back and listen at minute 1:32 in the transcript to see if it’s right. Well, it can. You can use AI to help you prepare for a conversation, such as one you’re going to have with your language partner or a tutor or a teacher, or you can even have a conversation with ChatGPT, like I was saying before. You can use it to listen to speech synthesis to get some ideas for how language is used. You can get it to correct your work, to quiz you, to explain grammar points. We’ve got quite a few tutorials that we’ve been putting out about these things. Even AI can be a fun monolingual dictionary because things that haven’t made their way into a dictionary yet, well, you’ll probably be able to get AI to give you a definition of those. And you can also get it to, as I said before, adjust it to your level. Basically, it can be a learner’s dictionary. There’s a good one in English, but not for every language, so AI can be fun for that. You can play games with AI. You can write stories collaboratively. You can dig into culture. We’ve got a tutorial about exploring etiquette scenarios, another one about generating recipes. And there’s a lot of fun stuff you can do. You can get AI to write songs for you in other languages, to create videos, animated dialogs in other languages, generate images. We’ve got a tutorial about fortune telling, and another one about digging into sports results. So there’s a lot of fun stuff you can do to learn more about the stuff you’re interested in at your level and we wanted to make this really accessible, so it’s only $9 a month.

Florian: So are you beta testing it yourself or testing it yourself in language number seven?

Josh: I’m testing it myself in all my languages all the time. I have six working languages, but I’ve got another six or so that I’ve dabbled in over the years, so yes.

Florian: What’s in the bucket list for fluency?

Josh: I don’t know. I’m pretty happy with the languages I speak right now and then dabbling a little bit more in others. But I was on vacation in Scandinavia last summer, and I said, wow, I love it. I’m loving Swedish. So who knows? There are good odds that because I’ve moved, I used to live in an institutional market in Geneva, and because I’ve moved to a market where there’s a lot more bi-directional interpreting. I think if I’m really going to double down on something, it will be on adding an active retour, working back into Spanish and probably back into Catalan as well, because they’re relevant here in Barcelona.

Florian: All right. Difficult question. Would you still choose interpreting and translation as a young graduate in 2024?

Josh: Yes. Easy answer. If you love languages, this is a great profession to be in, but you do need to specialize, and you do need to be willing to put in the legwork to start building a portfolio of clients, to constantly be upgrading your skills, potentially to add additional languages depending on the market you’re targeting for interpreting, especially for some institutional markets that can be hugely helpful. And you have to be willing to and committed to a life of lifelong learning. But I think a lot of translators and interpreters are. I mean, that’s how we got into this in the first place. We always loved learning about languages and cultures and keeping up with what’s going on in current affairs and digging into whatever that specialization might be. I mean, board games for geeky Josh, but maybe it’s fashion or cosmetics or whatever you’re into for you. There is work out there, but it takes quite a bit of legwork now to break into the market. I think it always has.

Florian: Okay, so I was browsing on your LinkedIn, and then I saw that you asked people to guess how many website domains you own. And since I’m also a little bit of a website domain collector, I’m like, oh my God, so how many? And then I scrolled through, and by the time I looked at it, you hadn’t given it away, so now I think you have.

Josh: The reason I asked this question is because I have a couple of challenges to help colleagues take these steps to learn about branding and web design on LinkedIn and so on. Right now I’m running the Design your Website challenge, and I realized I have a lot of domains. So drum roll, please, the answer, I had to look this up, is I have 19 domains right now.

Florian: Okay, same ballpark, because I think after 20, you start really questioning yourself and you’re like, why am I paying all these GoDaddy or whatever, with fees? And it gets actually a little expensive if you keep renewing them.

Josh: Well, yes and no. I mean, if they’re business expenses for specific projects or brands, then it’s a very small investment. So I have the domains for Innovation in Translation and AI in Translation because I run a big summit every year, and that helps to cover the costs. Same thing for Innovation in Interpreting, AI in Interpreting, for techforword, for AI Language Club. They do add up, but these are all active projects. So I try not to hang on to too many for inactive projects. And I do always recommend, if possible, to grab your full name, right? Firstandlastname.com, if you can get it, because even if you’re not using that, I would hang on to that one because someday you might want it if you change paths, if you’re a translator now, but later you move into interpreting, but you’ve picked a brand that has the word translation in your domain, that can be a little bit iffy. So anyways, I have 19. How many do you have, Florian?

Florian: I don’t know, but same ballpark, but not as many active ones. I have to count, probably 15 or 20 or something. All right, so you do have a lot of projects and things going on, but anything particularly you want to focus on in 2024? Anything new coming up? Anything you want to double down on?

Josh: Definitely doubling down on the AI in Translation Summit. We ran the first one last year in November, and the next one is scheduled for this November, November 12th to 14th, where we will be diving into the newest innovations in AI and bringing you lots of speakers to teach you about that. And Nora Díaz, with whom I run the summit and the AI in Translation and AI in Interpreting boot camps, and I are both committed to also running regular boot camps. So the next edition of the AI in Translation boot camp, which is a four-day, really hands-on experience where every day for 90 minutes a day, we dive into different angles of how you can use AI for translation or interpreting. The next one is coming up for the translation side of things in July from, I wrote this down, the 15th to 18th. And I will, of course, keep coming out with a new monthly webinar every month for techforword Insiders and for the broader community because I love exploring whatever new stuff is out there. If you told me what this year, the last twelve months of webinars would have been 12 months ago, I would never have in a million years, guessed that I would have been teaching about a good number of the topics that I’ve taught about and the same is probably the case for what the next 12 months will hold. We’ll see.

Florian: Cool. So where do people go? Top three destinations to get your content, connect with you.

Josh: techforword.com is the easiest place to find me. There are no A’s in that domain. Something I learned is be careful with plays on words. I thought it was clever to put the word ‘word’ into my brand name. The good thing is that nobody else had it, and it was easy to get all the handles and the domains and so on, but some people misspell it. So techforword.com is the number one place to find me. You can find me on LinkedIn as Josh Goldsmith. You can find AI Language Club as ailanguageclub.com and the same for those summits and projects I was mentioning.