SlatorPod

In this week’s SlatorPod, Florian and Esther discuss the quirky use of social media ads by the Royal Australian Navy in their quest to recruit 40 crypto linguists (which also happens to sound like a pretty cool job). We also talk about Salesforce’s new open-source XML-tagged dataset for NMT training, based on thousands of XML files per language pair to help improve translation of online help documentation.From Silicon Valley tech to the UK public sector, Esther shares her takeaways from a presentation given by the UK’s central government procurement arm, CCS, to nearly 80 would-be suppliers on a huge upcoming language services tender. Finally, as earnings season kicks off the two give a run down of the latest quarterly results from TransPerfect, Australia-listed Straker Translations and Teleperformance, the owners of interpreting giant LanguageLine.

Florian brings you this week’s language news from his hideaway in the mountains of Switzerland, sharing news on the world’s largest language model, GPT-3, from OpenAI. He and Esther (who is still hunkered down in London) discuss its applications in language generation, and whether or not it’s any good at the task of machine translation yet (spoiler: not really). Esther talks about two funding stories from the week: a USD 1.1m raise from Sweden-based localization and tech provider LocalizeDirect, which has built a headless CMS to support multilingual game development; and multilingual web conferencing platform Kudo, which raised USD 6m from VC based on their remote simultaneous interpreting tech. Leaping from distance interpreting to a machine interpreting, Florian discusses the launch of M.INTerpreting, an open-source project to help further research into machine interpreting, spearheaded by Claudio Fantinuoli from the University of Mainz. Also in the land of open-source, the two talk about Amazon Translate, which has just announced added support for powerpoint files.

In this week’s episode, Florian and Esther delve into the numbers and rationale behind the Memsource transaction, as the translation technology (TMS) provider sold a majority stake to private equity giant Carlyle in a landmark deal for the language industry.The two talk about highlights from the game localization panel at SlatorCon Remote, moderated by Daniel Finck with machine translation and localization experts from Frontier Developments and Electronic Arts, Tamara Tirják and Cristina Anselmi.From gaming to media entertainment, Esther talks about media localizer ZOO Digital’s financial results for the 12 months to March 2020. ZOO said the new year is off to a good start, and that the coronavirus lockdowns had proven to be a catalyst for buyers to try remote dubbing. Finally, Florian shares news from India, where a new language industry association, CITLoB, has formed, attracting 50 members within a month of launching.Finally, Florian unveils that SlatorPod now has its very own website here (the perks that come with new tools!).Links to the stories discussedNew! SlatorPod websiteMemsource Sells to Carlyle: The Inside StoryGame Localization: Why Nuance Is Key in Discussing AutomationZOO Digital Says FY 2021 Off to a Strong Start

Florian talks about research from Silicon Valley start-up Stripe, which surveyed hundreds of e-commerce websites for issues at checkout, the most common of which was a lack of local language translations. The two discuss SlatorCon Remote, where the 300-strong audience heard about e-commerce localization courtesy of multilingual content provider Quill. Also present at Slator’s inaugural remote conference was EasyLing, whose CEO described their web proxy solution for translated website. We had a private equity perspective on the industry, heard from CRM SaaS company Pipedrive about their localization operations, and featured experts across media and game localization, and moreEsther shares news of media post-production giant Deluxe’s partial sale to private equity (PE) and discusses a bit about the winning papers from ACL, the organizer of the world’s largest NLP conference, while Florian talks more about Memsource selling a majority stake to a major PE firm.Links to the stories discussed:Stripe Says Lack of Translation Most Common Error in European E-commerceDeluxe Sells Its Media Localization Arm to PE in Carve-UpJonckers Promotes Silke Zschweigert to CEOAnd the Winner Is…ACL 2020 Announces Best Paper Awards

In this week’s episode, Florian shares breaking news from the world of M&A and funding, as translation technology provider Memsource sold a majority stake to private equity giant Carlyle and machine translation matchmaker Intento raised USD 3m. Esther discusses a nationwide mega-tender for language services due to launch in August 2020 in the UK: the pan-government RFP is worth GBP 110m over three years and is designed to replace an existing contract that has been extended due to coronavirus. Keen to get the procurement process right, the buyer is hosting a webinar at the end of July for suppliers to give input on how the RFP should be structured.The two talk about Slator’s newly launched pro guide on How to Run a Translation and Localization RFP. If you’re a buyer keen to get your translation and localization procurement right, this guide is an indispensable resource for your organization and is packed full of practical insights on running an RFP.Florian provides an update on the Slator Language Industry Job Index (LIJI). Since plummeting between March and May 2020, the downward trend has continued into July 2020, but shows signs of slowing as a restart in investments points to the beginnings of a recovery for the industry. Links to the stories discussedThe Carlyle Group Invests in Translation Software Provider MemsourceIntento Raises $3 Million to Serve Increasing Demand of Global Companies for AI TransformationUK Plans Translation and Interpreting Mega-TenderPro Guide: How to Run a Translation and Localization RFPDecline in Slator Job Index Slows in July 2020

Florian and Esther discuss Apple’s announcement that its new operating system, iOS 14, will come pre-loaded with a Translate app when it’s released in September 2020. The Translate app will run on NMT, be one of just a few dozen built-in iPhone apps, and allow users to translate without an internet connection. What will this mean for other translation apps?Esther shares news of e-commerce company Localised’s USD 6.5m funding round with investment from a UK growth fund and celebrity entrepreneur Peter Jones of Dragon’s Den fame. Localised provides pre-built localized websites to help brands connect with prospective customers internationally. The second of three M&A and funding features this week, Esther talks about US government contractor SOSi’s investment in machine translation and speech recognition provider AppTek. Florian discusses the acquisition by Japanese financial printer Takara of local LSP Simul International from the owner of Berlitz Language Schools.And the two review a blog post published on Medium by a Localization Manager at Deliveroo. Links to the stories discussedNow We Know Why Apple Stepped Up Machine Translation Researchhttps://slator.com/machine-translation/now-we-know-why-apple-stepped-up-machine-translation-research/Localised.com Raises USD 6.5m in Series Ahttps://slator.com/ma-and-funding/localised-com-raises-usd-6-5m-in-series-a/SOSi Buys Minority Stake in Language Tech Firm AppTekhttps://slator.com/ma-and-funding/sosi-buys-minority-stake-in-language-tech-firm-apptek/Takara Pays Berlitz Owner USD 46m for Japanese LSP Simulhttps://slator.com/ma-and-funding/takara-pays-berlitz-owner-usd-46m-for-japanese-lsp-simul/

In a half-year anniversary episode, Florian and Esther unpack a wide-scale European Union survey of small and medium enterprises. Nearly 2,900 SMEs responded to questions about their language needs, MT use and attitudes of MT, as the EU rolls out an expansion program for their free automated translation platform, eTranslation.Florian ponders why a public body like the EU would be concerned with making MT available for use among private sector companies, and discusses a recent blog post from Google on the latest tweaks and techniques they’re applying to MT. The two talk about media localizer IYUNO’s acquisition of German dubbing and post-production studio Scalamedia, and highlight a possible language industry IPO reportedly being considered by Ai-Media in Australia. Florian shares his take on the diverging fortunes of language industry stakeholders amid the pandemic, as the gap widens between those who are suffering financially (e.g., the EU’s contract interpreters) and those who are seeing their businesses at all-time highs. Links to the stories discussedHuman or Machine Translation? Survey Reveals EU SME Preferences by Use Casehttps://slator.com/demand-drivers/human-or-machine-translation-survey-reveals-eu-sme-preferences-by-use-case/IYUNO Acquires German Dubbing Studio Scalamediahttps://slator.com/ma-and-funding/iyuno-acquires-german-dubbing-studio-scalamedia/Media Localizer Ai-Media Reported to Consider Sydney IPOhttps://slator.com/ma-and-funding/media-localizer-ai-media-reported-to-consider-sydney-ipo/Contract Interpreters Rally Over Lack of Support From EUhttps://slator.com/industry-news/contract-interpreters-rally-over-lack-of-support-from-the-eu/

Florian and Esther talk Google snafu, delving into the story of how Google “inadvertently sent” users’ translation memory data to someone else. Users of the now retired Google Toolkit, a free cloud-based translation productivity, may have had their translation memory contents shared with another user as a result of a technical issue affecting download requests.Florian discusses UK-based RWS’ acquisition of machine translation provider Iconic Translation Machines for a whopping USD 20m maximum, or 16x revenues. Why would the world’s most valuable LSP buy an MT startup rather than build the tech internally? An existing partnership, as well as Iconic’s focus on critical content and their customer base likely helped move the needle in favor of an acquisition.Esther outlines salary data coming out of the US Bureau of Labor Statistics, which indicates that US translators and interpreters earned a median wage of USD 51,830 annually in 2019. We also look at which states offer the best pay and employ the most linguists. Finally, Florian shares an update from the California Assembly, which passed AB1850 on June 11, 2020, exempting only “certified translators” from the proposed bill to ban gig work.Links to the stories discussedYour Google Translator Toolkit Data ‘Was Inadvertently Sent to Another User’https://slator.com/industry-news/google-translator-toolkit-data-was-inadvertently-sent-to-another-user/RWS Acquires MT Provider Iconic and Indian LSP Webduniahttps://slator.com/ma-and-funding/rws-acquires-mt-provider-iconic-and-indian-lsp-webdunia/*SlatorPro* Why the World’s Most Valuable LSP Bought a Machine Translation Startuphttps://slator.com/ma-and-funding/why-the-worlds-most-valuable-lsp-bought-a-machine-translation-startup/US Translator and Interpreter Salaries Rose 3.8% in 2019https://slator.com/industry-news/us-translator-and-interpreter-salaries-rose-3-8-in-2019/

Florian and Esther discuss the plight of the EU’s contract interpreters. While conference interpreting work has dried up, the ca. 3,000 contract interpreters, many of whom generated 100% of their income from the institutions, have been given a one-off payment to see them through the coronavirus pandemic. In a show of solidarity, however, staff interpreters created an emergency fund to help support their contract colleagues.Florian draws attention to the Slator Language Industry Job Index figures, showing a further drop from May to June, and highlights anecdotal evidence of localization job losses, including a LinkedIn-publicized cull in WeWork’s Globalization team. Esther shares insights from Verizon Media’s localization teams on their tech stack, including a proprietary loc platform, called Dragonfly, which helps to process 500,000 translated words monthly across platforms including Yahoo Mail, HuffPost, TechCrunch and more, in as many as 79 languages. Picking up on a news item in Jost Zetzsche’s Tool Box Journal, the two talk about free usage of SDL, Google and DeepL NMT. Use of SDL and Google’s free NMT has ramped up since March, with use of SDL’s free NMT around double that of Google and DeepL, according to SDL data. In other NMT news, Esther summarizes a research paper on the surprising and volatile behavior of MT in dealing with almost identical source sentences.Links to the stories discussedContract Interpreters ‘Puzzled’ and ‘Hurt’ by EU Offerhttps://slator.com/industry-news/contract-interpreters-puzzled-and-hurt-by-eu-offer/Slator Language Industry Job Index Falls Again in June 2020https://slator.com/industry-news/slator-language-industry-job-index-falls-again-in-june-2020/This Is How Verizon Media Runs a Centralized Localization Teamhttps://slator.com/features/this-is-how-verizon-media-runs-a-centralized-localization-team/Researchers Explore Surprising Behavior of Machine Translationhttps://slator.com/machine-translation/researchers-explore-surprising-behavior-of-machine-translation/

In this week’s episode, Florian and Esther discuss Netflix’s recent paper on sentence simplification and machine translation, destined to help improve the quality of machine-translated subtitles. But how will audiences react if simpler source sentences mean a lack of idioms and colloquialisms in foreign-language subtitles?The two talk about hyper-localized markets in the language industry and point out that a lot of demand is still local to an LSP’s home turf, limiting widespread industry consolidation. They discuss a few examples, from onsite dubbing and interpreting requirements in countries in Europe, to specific types of work that only exist in a particular geography, as with IPO translation in Hong Kong. They also revisit remote simultaneous interpreting (RSI) in Zoom, having received new user feedback from a listener. The conference interpreter, who had been using Zoom for a year, praised the platform: all they need to do is tweak a couple of things and it should be a solid interpreting option, she said. Florian schools Esther on some new terminology requirements out of France, where the official government gazette, the Journal officiel, has published a series of ‘correct’ French terms for English neologisms like, fittingly, podcast (audio à la demande).Links to the stories discussedHow Netflix Researchers Simplify Subtitles for Translationhttps://slator.com/machine-translation/how-netflix-researchers-simplify-subtitles-for-translation/Why You Need a Local Office to Compete in These 9 Translation Marketshttps://slator.com/features/why-you-need-a-local-office-to-compete-in-these-9-translation-markets/‘Fake News’ Should Be ‘Infox,’ French Government Sayshttps://slator.com/industry-news/fake-news-should-be-infox-french-government-says/