In this episode we tackle a rising topic related to mental health that is directly linked with prolonged Chatbot use.

In this episode we tackle AI psychosis, what it is, and where it comes from.

We talk about the involvement of LLMs and chatbots and what the reported cases of AI psychosis reveal about the users.

All AI models sound confident, but should we really believe it?

In this episode we tackle AI uncertainty, what it means, and why it matters.

The episode explains the different types of uncertainty in AI and whether or not we can overcome them.

We also go over methods of estimating when AI models are uncertain and why its getting harder to do as AI models get larger and larger.

In this episode we tackle the basis of AI uncertainty and overconfidence.

This episode serves as an introduction to AI uncertainty explaining why AI models in today's world (including LLMs) are overconfident, poorly calibrated, and unreliable.

with the rapid advancements of AI technology and people wondering if AI will take everyone's jobs I think its important to talk about one of AI's flaws that seems to be neglected in most talks. AI models of today really struggle with explaining their answers.

But what does it mean for an answer to be explainable? and what does it mean for an AI model to be interpretable?

And why don't models of today have either of those

What does it actually mean to reason? And do AI models do it like humans do?

In this episode we tackle the difference between AI and human reasoning. The episode highlights key research regarding the topic. We define the different types of reasoning and figure out which type of Reasoning can AI models really do, and if they're as smart as everyone says they are.

In this episode we delve into the role of AI in creative fields. The goal is to find an answer to the emerging question "Are AI models creative?"

The episode goes through the official definition of creativity and puts AI to the test going through how AI creates new things. We also tackle the biggest legal and ethical dilemma in AI generative content with ownership.

A lot of people have been wondering whether AI systems have gotten to the point where they have their own thoughts and feelings. In this episode, we explore how AI systems detect and respond to human emotions through data, patterns, and learned behavior. Then, we turn the lens around and look at when machines appear to show empathy, anger, or sadness, is anything really going on inside? We break down the difference between emotion recognition, expression, and true emotional experience, and discuss what researchers and philosophers say about whether machines could ever feel anything at all.

In this episode we tackle AI training data.

More specifically the thought that more data will always improve your models. From overfitting and noisy datasets to issues of scale, labeling, and ethics you'll understand that with AI models Quality is almost always better than quantity.We tackle problems where lack of good data leads to real issues and the ethicality of using widely available data on the internet without consent.