Content Operations

What Is LanguageOps? A Practical Guide to Running Multilingual Content

Translating one short document can look straightforward: prepare the text, translate it, review it and deliver it.

The difficulty appears when the content keeps changing, several languages are involved, terminology matters, and the same material is reused across a website, product, help centre, documents or campaigns. Then the challenge extends beyond translation to keeping the workflow organised, consistent, reviewable and maintainable.

That is the problem the term LanguageOps is useful for describing.

Correct grammar does not mean your translation is production-ready: how to unit-test your text

This post is aimed at developers and people new to localisation who might be tempted to think an LLM’s output is perfect. At LanguageOps we have automated processes that check for:

• Variables, $0, {name} are correctly formatted.
• Product names and terms use the approved spelling.
• Number and date formats correct for target and match source.
• Links or markup tags still operational and correct.
• Forbidden terms nowhere in the text.
• Source capitalisation matches target.
• …And dozens of other checks we run (nearly 40 in total).

These are usually mechanical checks which require no LLM use and can be run quickly with a fixed set of rules, to save your reviewers falling into these repeated traps across languages and projects. Whole QA software products exist to prevent these leaking into translations - we’ve built them in to the editor for linguists or PMs to quickly run, filter and resolve.

AI safety beyond the frontier labs: uncensored local models

A recent Guardian report showed researchers asking frontier models to help plan violence and crime. Frontier labs (OpenAI, Google, Anthropic) responded that they’d already employed better safeguards since the research concluded. Ignoring the fact that you can still work your way around safeguards in many cases, they make no mention of local models.

I’ve experimented with highly capable local models - both from Google and from Alibaba - Gemini and Qwen - and found them to have similar safeguarding to hosted models. But then I checked the “Uncensored” versions of the same models, where researchers and enthusiasts take the base model and surgically isolate and remove the weights relating to guardrails and alignment. They report that the models answer 100% or 95+% of the unsafe questions asked of it, vs 1-5% before the further work.

Developer Agency in the Age of AI

As a bit of background, I have written about performance and security in posts that hit the front page of Hacker News, long before LLMs were in use. I’ve been setting up and securing Linux machines and servers for decades. I was writing in QBasic, a menu-driven “OS”, then simple games in C++ from around 12 (i.e. in the 90s). I am a big proponent of open source (FOSS)¹ and software anyone can run and work on, enabling me to have enterprise capability at no cost or low cost because I learned how to set it all up securely and to run at scale. I’ve brought this knowledge, of course, into the LanguageOps project. This is one area often missing from AI developed projects. People tend to underestimate the effort to operationalize a project.

AGI v1.0 Will Be Harness + Skills

Consider the trajectory of LLMs so far. Starting in Google’s translation research departments, following on from their groundbreaking neural network models, transformers were first and foremost a “language” model. They have grown to be able to convince people of their utility, much like the market stall owner would do pre-2000s, holding groups of up to 50 people in awe because of their fast talking, confidence and skill demoing whatever they were selling. The same works on shopping channels and more recently Youtube. Speak fast and confidently, cut all the dead air, and you got yourself the makings of a popular channel.

Council Translation: How Multi-Model Consensus Beats Single-Engine Output

Every machine translation engine has blind spots. DeepL struggles with certain idioms. Google Translate sometimes loses register. LLMs occasionally hallucinate. When your translation workflow relies on a single engine, you inherit all of its weaknesses.

The localization industry has accepted this as an unavoidable cost of automation. Post-editors spend hours fixing the same predictable errors, project after project. Quality teams develop mental checklists of “things to watch for” with each engine. It’s become so normalized that most platforms don’t even question it.