Chapter 1: Introduction to AI and Medium Earnings

Have you ever wondered how much your favorite Medium writers earn? With the aid of artificial intelligence, that mystery can be unraveled! If you're not yet a member, you can access this information for free.

In this continuation of our previous discussion, we will enhance our approach to predicting earnings by incorporating a more precise method. Our focus will be on answering a crucial question: Is it feasible to estimate earnings based on metrics like claps, comments, and read time?

Photo by Michał Parzuchowski on Unsplash

Artificial Intelligence (AI)

Recently, I developed an AI tool capable of estimating Medium earnings based on claps and comments. Now, I'm back with an upgraded version, thanks to the insightful suggestions from Bin Jiang and Theo Rose, who recommended adding read time as a key factor in our predictions. Kudos to you both! Additionally, this model can now offer an earnings range for predictions.

Key Terms

• Predicted: The estimated earnings the AI expects for a story.
• Actual: The real earnings recorded for the story.
• Confidence Interval (CI): The range within which the actual earnings are likely to fall, with a 95% confidence level, presented as [lower earnings, upper earnings].

Chapter 2: Predictions Based on Personal Stories

To illustrate the predictions, I will first share forecasts for my own articles before expanding to include estimates for some of my favorite Medium authors.

My Stories

• 0 Reads? How I Get 100+ Reads per Story:
  • Views: 875 | Claps: 105 | Comments: 3 | Read Time: 2 minutes
  • Predicted: $0.88
  • Actual: $0.80
  • CI: [$0.17, $1.59]
• Medium Killed My Page:
  • Claps: 720 | Comments: 17 | Read Time: 2 minutes
  • Predicted: $6.93
  • Actual: $4.55
  • CI: [$6.21, $7.66]

As we can see, the predictions for my articles appear to be more accurate than in previous assessments. But how will the AI perform when evaluating other authors' works? I selected five writers from my top 100 favorites to analyze.

Notable Authors and Predictions

• Danny Wolf:
  • Claps: 536 | Comments: 11 | Read Time: 4 minutes
  • Predicted: $4.72 | CI: [$1.77, $10.81]
• Martin Kessler:
  • Claps: 431 | Comments: 8 | Read Time: 3 minutes
  • Predicted: $3.39 | CI: [$1.15, $7.97]

You can explore more about your top writers through Medium’s new hidden feature!

Chapter 3: Assessing the Model's Performance

While the predictions for my own stories seem more reliable, the model still exhibits some discrepancies. For instance, even though Johnson's story garnered more claps and comments than Ryan's, the predicted earnings for Johnson are lower. This inconsistency arises because the model is primarily trained on my data. Most of my longer articles are coding tutorials, which typically do not perform well monetarily. Additionally, when a post receives over 1000 claps, readers are less inclined to clap multiple times.

Summary of Issues:

• The model appears to be overly fitted to my writing style; it requires a more diverse dataset.
• It does not consider the unique number of clappers or commenters.

If you're interested in contributing data to enhance this project, please feel free to get in touch!

Chapter 4: Conclusion

In this piece, we've designed an AI capable of predicting Medium earnings by analyzing claps, comments, and read time. Compared to the initial version, this model shows improvement thanks to the integration of read time and the ability to calculate prediction confidence. However, it is clear that further refinement is necessary.

The first video titled "Here's How Much Passive Income I Earned From Writing AI Medium Articles" provides insights into the earnings generated through AI-assisted writing.

The second video, "I Tried Writing on Medium with AI For 30 Days. Here is How Much I Made," shares a personal experiment on earnings through AI writing over a month.

In conclusion, while the AI model shows promise, it remains a work in progress, eager for improvement and accuracy.