Written about my last startup, vivatranslate.com
A Complete Abstraction of Language in the Workplace
In today’s world, talent is equally distributed but opportunity is not.
English is the current global language of business, yet 80% of the world does not speak English at all and 90% of the world does not speak English fluently.
From the Tower of Babel to Star Trek’s universal translator, the world has dreamed of what life would look like if you could speak with anyone else effortlessly.
With the advent of large language models such as GPT-3 and BERT, AI translation output has in many applications become indistinguishable from that of a human translator. Machine translation has even been used to communicate in high-stakes situations such as medical cases. For daily situations around the world, AI translation is ubiquitous - in 2020, over 1 billion people used a translation application, or roughly 15% of the world’s population.
Despite these advancements, the most crucial use cases for MT are still yet to be unlocked. How do we enable someone to connect with any other person around the world intellectually, emotionally, and professionally?
Here are some cases that are possible with true removal of language barriers:
- Workers are hired completely independently of their language abilities.
- Businesses can work with employees abroad, bringing hiring costs down while working with great talent.
- Workers abroad get remote jobs regardless of their language abilities, allowing them to make more money and access to professional growth opportunities.
- People in developing countries can work remotely without leaving their homeland, family, or friends in search of better opportunities. There will be no need to sacrifice community or culture for career.
- Students can learn from teachers across languages.
- Researchers can collaborate with any other research independent of language.
- Diplomats would have lower friction to communicate with one another.
- Loved ones can communicate across generations where language was previously a barrier.
- Adults aren’t required to learn another language in order to work internationally.
- Adults are better able to learn a language through language immersion at work.
- Engineers and designers abroad get remote jobs regardless of their language abilities - therefore, they make more money and have access to professional growth opportunities.
- Companies can hire freelancers completely independent of their language abilities - and save money.
- People in developing countries can work remotely without leaving their country, family & friends in search of better opportunities.
- Adults aren’t required to learn another language in order to work across borders.
This is not solely due to language model quality.
This is due to lack of: (1) accuracy with rich, context aware translations, (2) design for live, multi-modal cases, and (3) ease of use across software and devices. Viva solves this through integrations, domain-specific data, and speech translation.
Viva is a translation ecosystem that encompasses all day-to-day pain points around collaborating with someone who speaks a different language. Translations are domain-specific and context-aware, so they are more accurate than general-use tools. The first application of Viva’s ecosystem is to support remote work between the U.S. and Latin America.
There is an economic, social, and educational imperative to accelerate towards cross-language work. This outlines the reasons that ecosystem players of educators, learners, professionals, and companies should contribute to a cross-language world.
Today’s Language Arbitrage
Here is the current state of the world:
- 80% of the world is not English-literate but is tech-literate
- There is a large and growing shortage of 40 million skilled technology workers globally
- AI, including machine translation, can now translate real-time conversations
80% of the world is not English-literate but is tech-literate
80% of the global workforce does not speak English.
Tech literacy is growing at a much, much faster rate than English literacy. Tech literacy, defined as knowing how to use a smartphone, is at 84% globally. This number increased by almost 50% from 2017 to 2022.
Tech work has becoming the cornerstone of our global economy. There are approximately 250 million knowledge workers globally and while knowledge workers only make up 10% of the world’s workforce, they account for over 50% of GDP.
Technology & remote worker shortage
There is a massive growing gap between English-speaking technology workers and technology jobs. By 2030, there will be an estimated global shortage of over 85 million tech workers, representing $8.5 trillion in lost annual revenue.
What’s more, tech workers in developing countries are eager to get remote jobs abroad. Tech workers in developing countries often get paid twice, sometimes 3-5 times as much, working for a company in the United States. Freelancer earning potential for specialized roles is much higher than a general role - software developers can earn up to $1,000 USD / hr. As a result of this, technology is one of the two top industries for new freelancers.
Remote work is increasing at a rapid rate the world over, with 70% of the workforce predicted to work remotely a minimum of five days per month by 2025.
Working remotely for a foreign company is also not only for monetary gain. Working for a company in the States is also a tremendous learning opportunity.
AI for cross-language communication
The origins of NLP actually originated from machine translation devising ways to decode German intelligence during WWII. At that time, they projected full cross-language fluency.
It’s been several decades more, but now the world sits on another advancement in both AI infrastructure and modeling - large language models such as GPT-3 and BERT. Language translation and NLP has drastically improved.
There is still massive room for improvement in MT. Recent focus has been on training multilingual models to cover low-resource languages like Cherokee. Viva takes a fundamentally different approach by focusing on high-accuracy language pairs and a specific domain.
There are many other context issues that can be aided by using large language models applied in chained steps alongside MT.
Viva: A language tool to fast-track to the next decade of communication barrier removal
Viva provides a translation ecosystem across bi-directional communication, long-form documents, and real-time speech. This increases comprehension for any professional conversations across linguistic and cultural literacy.
- Accuracy: Rich, context-aware translations
- Design: Live multi-modal translations
- Integrations: Anywhere all at once
Players that Make Cross-Language Work a Possibility
There are multiple types of players in the global ecosystem that require effective remote communication together. For each, there are economic reasons to work together across language.
| Player Type | Example |
|---|---|
| Companies | Coca Cola, IBM, Mercado Libre |
| Remote Service Providers | Teleperformance, Globant |
| Educators | Stanford, Carnegie Mellon, Platzi |
| Learners | Computer Science Undergraduate |
| Technologists | Viva |
| Government | California Office of the Small Business Advocate |
Applications of Translation
Just like Google Translate, Viva will be able to be used ubiquitously. Here is a non-extensive list of possibilities for what we intend to tackle in the next decade:
- Freelance Developer works with U.S. Client
- Researcher collaborates with Researcher
- Technology Company hires Remote Worker
- Remote Worker takes online English course with Educational Institution to upskill
- Student learns from any Teacher in the world
- Call Center hires Agen
Industries
Industries that will be impacted will include but are not limited to:
- Software & IT
- Manufacturing
- Agriculture
- Ecommerce & Trade
- Freelance Work
- Business Process Outsourcing
- Mining
- Education
Conclusion
Let’s make it such that access to opportunities is decoupled from the language you speak.
Imagine this world.