A falling tree makes more noise than a growing forest

This week, I read a series of articles, posts, and news that made me think about where we are and where we are going with AI in the language services industry.

The first half of 2023 presented a seemingly slow period for the localization industry, aligning with a general perception of economic malaise. Major players reported flat or decreasing revenue, contributing to a narrative of industry-wide stagnation.?

Despite this, some segments within the localization industry, notably in infrastructure, manufacturing, and healthcare, experienced growth. And services like interpretation and media localization (not related to entertainment, which was affected by the Hollywood strikes) experienced surges in demand.

However, this negative perception, heavily influenced by tech sector events like layoffs and the shutting of startups, especially in Silicon Valley, was shaped by negative news overshadowing positive economic indicators like low unemployment and controlled inflation. As Bill Gates says, "Headlines, in a way, are what mislead you because bad news is a headline, and gradual improvement is not." Or, as the African proverb goes, "A falling tree makes more noise than a growing forest."?

According to the Bruegel Newsletter, despite the perceived challenges, the European Union and the United States have shown resilience, with the EU particularly demonstrating notable growth in output per capita and productivity.

In a conversation with a client this week, the company reported that they had lost a significant customer to AI earlier in the year, only to regain them a few months later because the promises of LLMs and genAI didn't materialize. This indicates and exemplifies the industry's dynamic nature and capacity for recovery.

Two other readings, an Interview with Brad Lightcap, the COO of OpenAI, and a post by my friend John Tinsley on LinkedIn, got me thinking more about what I wrote about adopting AI in the language industry.

Brad Lightcap from OpenAI has tempered expectations around AI, highlighting its limitations in current business applications. Despite the initial excitement around technologies like ChatGPT, their real-world impact has been more moderate than anticipated. Lightcap’s insights shed light on the evolutionary nature of AI integration in business settings. For more on Lightcap's views, see The Verge's article.

Following Brad Lightcap's insights, John Tinsley from Translated adds a perspective on applying Large Language Models (LLMs) in machine translation. Tinsley emphasizes that LLMs, including GPT-4, are not yet suitable as a direct replacement for enterprise-level machine translation. While LLMs show promise in some areas, they lack critical factors like speed, cost, and information security. Tinsley points out that current machine translation tools are already utilizing advanced AI, and while LLMs may become standard in the future, they are not there yet.

These readings reinforce my belief that the language industry sells not words but the ability to manage those words in complex environments with a mix of technologies and formats. AI is here to help our industry, not to replace it.

At Nimdzi Insights, our analysis of interim data makes us think of the initial fear and pessimism of 2020, driven by COVID-19, that contrasted sharply with 2021's record performance in the language services sector.?

Our surveys note caution, particularly in the IT-focused LSPs, with its cascading effect from MLVs to SLVs. But 2023 also saw AI technologies overpromising and underdelivering, adding to the industry's cautious outlook. Yet, based on our industry's history of overcoming challenges, we maintain an optimistic view, expect modest growth in 2023, and expect to return to normal trends in 2024.