So, just for interest, I thought you might like to hear about my recent experience, as an AI Ranking Assistant.

I was invited to join a new project to train a Large Language Model, which I later discovered was Mistral AI, which I had never heard of, but the company is in fourth place in the global AI race and in first place outside the US!

Being a huge fan of AI, I was intrigued and decided to take the assessment tests. 

The first one was easy, just an English spelling and grammar test.

The second and third tests increased in difficulty and required writing prompts to encourage the AI model to engage in conversation. 

There were a lot of strict rules to follow and the final test took me two hours.

However, I was delighted to learn I had been accepted and added to the project.

Full training followed, consisting of reading an 80-page document and a 90-minute Zoom session with 9 others who had also passed the tests!

Finally, we were given access and could begin. 

And man, was it difficult! 

Essentially, you had to think of tons of questions on any subject, and then the AI model would give 2 answers. 

You then had to evaluate the answers, and rank them on a scale, and decide which was best, and then justify your answer. 

And then... ask another question relating to the best answer and repeat the process up to 8 times, to create a long, ongoing conversation!

With no guarantee your conversation would be deemed acceptable, and therefore be paid...

The compensation rate was purported to be $19 per hour, but to achieve that, you would need to write over 40 questions and evaluate all the answers!

The maximum I achieved was 8 questions in one hour!

After a few hours, I gave up, it was too time-consuming, given the low rate of pay.

I was paid $128 for passing the tests, attending the training, and writing a couple of hours of prompts.

It was an interesting experience, to see how the AI models are trained, but I wouldn't do it again!