Ego querying

One way to measure the accuracy of large language models applied to internet search is to use your own name as a query.

Since OpenAI released their own search engine combined with GPT, I couldn't resist doing some "ego querying."

What I found is that GPT indexed information on websites where I posted blogs or contributed indirectly.

When I realized where GPT got its data, I asked myself if there is still a "cut-over," since there was nothing about the blog released a few months ago, or even the contributions I am making on a weekly basis on LinkedIn and Twitter.

Google does have some legitimate bias when displaying the search results for a query, but at least you can make your own choice on which one you would like to use to be informed about the query you asked.

Here, you have a persuasive and efficient digest based on a subset of data found on the Internet. This can be super relevant since it might be enough, or it can actually be misleading since the way the information is gathered creates mistakes (I didn't move to Japan in 2001!).

This leads me to think of a motto I heard a few years ago: at the end of the day, you have to do your own research.