# ðŸ¦‰ Birds of night are a hoot. They give it owl they've got!

Late to the Party ðŸŽ‰ is about insights into real-world AI without the hype.

Hello internet,

How are you doing today? Weâ€™ve got some fresh, hot machine learning right from the source today!

## The Latest Fashion

- CNet ran a bunch of AI-generated articles and found errors in more than 50% in hindsight
- Pyvista looks really nice for 3D visualization in Python
- You can write GPT from scratch in 60 lines

*Got this from a friend? Subscribe here!*

## My Current Obsession

I am getting sick, so I havenâ€™t been up to much, unfortunately.

I have been polishing up ml.recipes. In my usual manners, I have been terrible at marketing this resource in my usual manner, so if you shared it with some colleagues: Thank you!

Someone also wrote a fantastic long review of my AI Art course, which is lovely. I will probably share that one more next week.

## Machine Learning Insights

Last week I asked, ** What is the p-value and how is it relevant in statistics?**, and hereâ€™s the gist of it:

Let’s say you want to test the hypothesis that earthquakes are more likely to occur in California than New York.

You collect data on earthquake occurrences in both states over a certain period and run some statistical tests. If your p-value is low, this will provide evidence to support your hypothesis - in other words, earthquakes are more likely to occur in California than New York. However, if your p-value is high, this would suggest insufficient evidence to support your hypothesis.

In statistics, a p-value measures how likely it is that a particular result occurred by chance.

It’s kind of like a way of testing the likelihood of a hypothesis being true.

Let’s say you’re a geophysicist studying earthquakes and want to test a hypothesis that a specific area is more likely to experience tremors than another. You collect a bunch of data, and you run some statistical tests. The p-value tells you how likely it is that the data you collected supports your hypothesis.

Let’s say your p-value is very low, less than 0.05. This means that there’s a tiny chance that your data supports your hypothesis by chance. In other words, it’s implausible that your hypothesis is wrong. This is usually taken as evidence to support your hypothesis.

On the other hand, if your p-value is higher than 0.05, this means that there’s a high chance that your data supports your hypothesis by chance. In other words, this is often taken as evidence that the data does not support your hypothesis.

## Data Stories

These owl drawings are 130 years old and simply incredible.

There are many more in the source linked below.

[Source]

## Question of the Week

**How can you select the most important features in a data set?**

*Post them on Twitter and Tag me. I’d love to see what you come up with. Then I can include them in the next issue!*

## Tidbits from the Web

- Revisiting a tweet with unused emoji slot is a fun idea!
- Do you actually need a fancy knowledge management tool? Unbundling tools for thought
- I found these 20 things they learned in 20 years of career interesting

Jesper Dramsch is the creator of PythonDeadlin.es, ML.recipes, data-science-gui.de and the Latent Space Community.

I laid out my ethics including my stance on sponsorships, in case you're interested!