Home Blog Two stories about ChatGPT

our blog

Two stories about ChatGPT

Two stories about ChatGPT - 1
03 June 2026

Interesting facts about AI

In early April 2024, artificial intelligence expert Jeremy Nguyen shared a thing he noticed: for some reason, the unremarkable word delve was used unexpectedly often in the texts generated by ChatGPT. So much often that Paul Graham, co-founder of Y Combinator and noted IT evangelist, considers the presence of this word a sign that the text is generated by AI.

Delve is a usual word in the “common” Internet, but it is very popular in its African segment, particularly in Nigeria. And it looks like Africans were involved in the creation of the dataset on which ChatGPT was trained. So, you get out what you put in: ChatGPT speaks the language from the dataset, no reason to be surprised.

In addition to delve, there are some more common words used by language models too often: explore, captivate, tapestry, leverage, embrace, resonate, dynamic, testament, elevate.

Recommended content

Two stories about ChatGPT - 2

Artificial Intelligence is not Intelligence

A few thoughts on artificial intelligence For an AI chatbot like ChatGPT to work, it first needs to train the Large Language Model (LLM) it relies on using a huge amount of text data. Modern LLMs are trained on almost everything that can be found on the internet. The texts generated by chatbots are often […]
Knowledge Hub
Two stories about ChatGPT - 3

Where will experienced translators come from if there are fewer inexperienced ones on the market?

Abstract semi-philosophical considerations Machine translation is gradually pushing out of the market those translators who cannot keep up with it in terms of quality, i.e., those who are not very experienced. As a result, the ratio of experienced to inexperienced translators should theoretically gradually change in favor of the former. This seems to be a […]
Knowledge Hub
Two stories about ChatGPT - 4

How to fix «Failed to save target content: Cannot find central directory» error in Trados Studio

About the ways to fix the issue in Trados Studio When trying to generate a translated document in Trados Studio, this error sometimes occurs: Failed to save target content: Cannot find central directory. It happens because of corrupted data in the source file used to create and store the SDLXLIFF file you translated in Trados […]
Knowledge Hub
Two stories about ChatGPT - 5

How to fix «Failed to open return package» error in Trados Studio

About another issue in Trados Studio When importing a package, Trados Studio might say that it Failed to open return package, because it contains a project that does not exist in your setup. That it’s not a Trados Studio issue, but yours: you tried to import a return package for a project Trados Studio doesn’t […]
Knowledge Hub
Two stories about ChatGPT - 6

Lorem ipsum

About the dummy or placeholder text Lorem ipsum is a meaningless Latin text, so-called dummy or placeholder text, which is used to estimate how a layout will look with text presented in a particular font. This is usually done by designers, but it is also needed for pseudo-localization. This text became widely popular due to […]
Knowledge Hub