This is the audio version of my Substack newsletter. My name is Alperen Zeki Gokmen.

In this episode, I talk about what makes reasoning models like o1, DeepSeek R1, and Claude 3.7 Sonnet different from classic language models, and how these new models can help you solve more complex problems step by step.

I also explain the new update from OpenAI that lets you use any model with custom GPTs, making it much easier to create specialized assistants for your work.

You’ll hear a simple breakdown of temperature and top p settings, so you can get better results from AI just by tweaking a few options.

Finally, I take a look at Apple’s latest “Illusion of Thinking” paper and what it really means for AI progress and reliability today.

Enjoy the episode!

Audio Disclaimer:

This episode is made with AI voices. The AI hosts talk about the main ideas from my newsletter. Some small details may be different from the written text. For the most complete information, please read the full issue.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit alperenzekigokmen.substack.com